SF.35.24.B10212: Cyber Security for Radio Frequency Receiving and Processing Systems
The ability to detect and mitigate cyber vulnerabilities is critical to providing mission assurance. While there has been significant research and development to determine cyber vulnerabilities in Enterprise systems (e.g., desktop computers), much less focus has been placed on mitigating vulnerabilities in sensor and other embedded systems including radio and radar systems. This effort will focus on developing dynamic software vulnerability detection capabilities for radio frequency (RF) receivers and associated signal processing software including Software Defined Radios (SDRs). This research has three major objectives:
(1) Determining the threat accessibility to signal processing software and associated susceptibilities. This includes a study of signal/waveform propagation and the logical and physical reachability of software susceptibilities by the input signal from the receiving antenna to the backend software containing the targeted susceptibilities. (i.e., perform a data flow analysis to determine accessibility to the susceptibility)
(2) Study the potential impact of the exploited susceptibility on the system. Potential impacts include “crashing” or disabling the receiver, causing misclassification of targets, and other anomalies that intentionally mislead or misguide the receiving system. Results of this study will allow for prioritization of the RF software for further analysis.
(3) Demonstrate the ability to discover and mitigate RF signal processing software susceptibilities by fuzz-testing software within the RF system using simulated encoded pulses/waveforms.
The end-goal of this project is to develop a proof-of-concept capability that can extend existing software fuzz-testing tools to perform “RF pulse fuzzing” to discover and mitigate software susceptibilities in RF-based embedded systems.
SF.35.24.B10153: Automation of Knowledge Discovery
Sense-Making is understood by the Intelligence Community as a process that meaningfully translates relevant data into usable information. This includes processing, correlation, and fusion of data from various sources followed by reasoning to infer new knowledge. Automation of this process requires increased interoperability between various information sources and automated reasoning that incorporates expert and common sense knowledge. The Department of Defense has made significant progress towards solving the interoperability problem by adapting Object Based Production and introducing common information representation standards. Automated reasoning remains a challenge. The objective of this research is to develop and demonstrate the capability of automated reasoning with structured data that utilizes the Department of Defense recommended knowledge representation standards and incorporates uncertainty in reasoning. Particular challenges include automation of knowledge elicitation, representation of common sense knowledge, and representation and reasoning with uncertain information. Particular interest is in utilizing logic based techniques combined with probabilistic and machine learning methods.
Key Words: Knowledge Representation, Probabilistic Logic Programming, Common Sense Knowledge, Ontology Learning, Ontological Reasoning, Information Fusion
SF.35.23.B10135: Electronic-grade dielectric integration for high-power, high frequency electronic devices
Successful integration of dielectrics into a transistor process flow with negligible defect density has historically been the key for wide scale application of electronic devices. Dielectrics are needed not only as gate insulators for operation of metal oxide semiconductor field-effect transistors (MOSFETs), they are also needed for passivation of metal semiconductor FET (MESFET) and high-electron mobility transistors (HEMT; which is a different form of MESFET). The presence of defects either in the bulk or in the interface of these dielectrics critically affects the performance of transistors. Transistors for RF operation use all the above transistor configurations. The semiconducting channel in these transistors are generally made with III-V (like GaAs, GaN, AlGaN) or III-O (like Ga2O3, AlGaO) materials. These materials do not have a native dielectric as Si has in the form of SiO2; and therefore, have an unoptimized dielectric/semiconductor interface even 40 years after their introduction into RF electronics. In addition, formation of novel dielectrics on these materials poses additional challenges in terms of bulk and interface defects, and carrier injection into dielectric, which leads to instability in device operation. Significant research opportunities therefore exist in integrating classical and novel dielectrics in III-V and III-O based semiconductors. These are especially important for high power RF applications that require use of wide bandgap (WBG) materials like III-N and III-O and require high voltage application across the dielectric.
This research targets successful integration of dielectrics in high-power GaN-based RF transistors. This will require optimization of a wide range of process parameters during device fabrication in AFRL/RY’s class 100 (ISO-5) cleanroom. Resultant devices will go through extensive electrical (C-V, I-V, transient, noise), optical (different forms of spectroscopy and microscopy) and materials characterization for confirming the effect of different process parameters on device performance. AFRL has excellent characterization capability that will be useful for such characterization. The goal of this project is to generate critical and novel knowledge that will enable application of WBG materials that will satisfy unique requirements of the United States Air Force and Space Force.
SF.35.22.B10110: Riemannian geometry and locally most powerful invariant tests of covariance equality
In statistical machine learning and autonomy research, a fundamental challenge is determining when two samples come from the same population. As an example, one may wish to determine whether adversarial activity is present in a given sample by comparing it to another sample. One approach to this problem, which is of limited utility, is to assume that the samples differ only in their means, in which case the Hotelling T^2 test is standard. A more realistic approach is to assume differences can occur in both means and covariances. However, standard techniques for comparing the samples' covariance matrices are ad hoc and have not been shown to be optimal in any sense in high dimensions. The main objective of this project is to use tools from Riemannian geometry and random matrix theory to develop more powerful invariant tests, at least in the regime of low signal-to-noise ratios. Additionally, to enable utility in control-loop systems, a mathematical performance analysis of any developed test should be conducted, including provable significance and confidence levels. Likely approaches will include high-dimensional asymptotics, linear spectral statistics, local shrinkage eigenvalue laws, and geodesic distances on the statistical manifold of multivariate Gaussians.
SF.35.22.B10097: Design and Verification of Automatic, Autonomous, and Intelligent Aerospace Control Systems
Autonomous systems that utilize novel control techniques, including those based on machine learning, promise to improve the speed with which systems can react in real time to changing mission needs and reduce the number of humans required to operate large numbers of unmanned aircraft or a constellation of satellites. However, hard to detect design flaws in advanced controllers could have catastrophic consequences, especially in safety or mission critical applications. The controller designs could address issues such as simultaneous satisfaction of multiple design constraints, real time task assignment and prioritization, effects of uncertainty with information and game theory considerations, operation on processing and memory constrained computing hardware, interactions with human operators, and scalability of the design. Verification of these controller designs could include provably correct algorithms and architectures, formal verification, reachability analysis, fuzzing, novel test case generation approaches, and/or offline verification techniques such as monitoring and bounding of system behavior. Development of verification evidence could be conducted in concert with hazard analysis approaches that consider software failures and human interaction as well as novel approaches to certification based on structured assurance case arguments are of interest. Verification approaches could complement traditional approaches such as failure modes and effects testing/analysis, Monte Carlo simulation, hardware in the loop simulation, flight test, etc. The objective of this research is to develop planning, guidance, and control methodologies and accompanying verification approaches for aerospace systems such as fixed wing aircraft or spacecraft operating collaboratively. Verification approaches can include offline verification such as formal methods and mathematical analysis or novel test case generation techniques, as well as online verification approaches such as run time assurance.
SF.35.22.B10074: Semiconductor Laser Source Phenomenology and Development
Applications in active sensors and infrared countermeasures may often be significantly improved by unique sources. We theoretically and experimentally investigate high-brightness semiconductor lasers with unique properties such as broad tunability, mid-wave and long-wave operation, fast modulation, and multi-wavelength operation. Devices such as vertical external cavity surface emitting lasers combine the inordinate semiconductor optical gain with a classic macroscopic laser cavity, which results in a device that can achieve signal spatial-mode operation under many watts of optical output powers approaching in any wavelength achievable by semiconductor lasers. We are interested in longer wavelength semiconductor lasers in uncommon operation regions, such as the 2-4-micron band through interband Type-I and Type-II lasers, large conduction-band offset intersubband lasers. It is essential to understand material systems (such as GaAs- and GaSb-related compounds) to appropriately design a laser cavity. For example, Auger recombination and carrier confinement both become extremely influential at high power and temperatures. Mitigation of these deleterious effects is essential for efficient operation. Interest areas include material design and development, fabrication, coupled resonator designs, passive mode-locked lasers, spatial temporal waveform generation schemes for semiconductor lasers, as well as nonlinear frequency conversion. Fabrication facilities are available including semiconductor growth, typical microfabrication techniques, electron-beam lithography, as well as numerous metal and dielectric film deposition techniques. A full testing capability suite is available for cryogenic through high-temperature measurements for device operation. Keywords: Lasers; Semiconductor; Infrared; High-brightness; Resonator; Unstable beam; Nonlinear;
SF.35.22.B10073: Polymeric Materials for Sensors and Electronics
Our interests encompass characterization and proof-of-concept devices leveraging polymers that absorb in the infrared and radiofrequency regions, such as focal plane arrays and microwave absorbers. They additionally have properties that can be applied to polymer-based, direct current electronics such as electrodes. Our current work includes fabrication and testing of photoconductive detectors from conjugated polymers, modeling of antennas and RF absorbers made from highly conductive organic polymers, and electro-optical characterization of conductive, aliphatic polymers. The work involves theoretical modeling as well as experimental characterization. Modeling capabilities include Lumerical, Comsol, Matlab, and T-Spice. Fabrication facilities include a general purpose sputtering lab, oxidative chemical vapor deposition, shadow mask design and fabrication, and general purpose wet chemistry lab. Characterization capabilities include Hall effect systems, probe stations, ellipsometry, FTIR, SEM, TEM, femtosecond lasers tunable from 280 nm- 20 m, variable temperature luminescence, and blackbody radiators. Access to an on-site, fully capable, Class 100 microelectronics cleanroom as well as high performance computing systems, can be arranged.
SF.35.21.B10048: Computational Modeling of Non-linear Processes in the Ionospheric Plasma
Ionospheric plasma supports high-frequency and low-frequency modes which can be excited concurrently by large-amplitude, high- frequency pump waves (either electromagnetic (EM) or electrostatic (ES)). The coupling of the pump wave and the low-frequency plasma wave results in parametric instabilities. The goal is to study wave-ionosphere interaction and underlying linear and non-linear processes in ionospheric plasmas in the very low frequency (VLF) and extremely low frequency (ELF) range of wave frequencies. Wave localization, electron and ion heating, filament type structure formation, and explosive-like growth of wave amplitude leading to generation of density irregularities are the problems of interest for the proposed research topic. Impact of excited turbulent density pulsations on propagation of high frequency waves based on the obtained short scale description of irregularities can also be a subject of the proposed research.
SF.35.21.B0002: Heterogeneous Integration for Photonics Enabled RF Microsystems
This topic seeks to develop optimized chip-to-chip and chip-to-fiber interconnects for mixed technology photonic RF microsystems utilizing agile fabrication techniques to account for device-to-device misalignment during assembly. Techniques that may be explored through both modelling and experiment may include but are not limited to 2-photon lithography, grayscale lithography, digital projection lithography, alignment tolerant design, and MEMS based alignment correction [1].
The wide bandwidth of photonic systems present a significant opportunity for modulation and detection of RF signals, low-loss, low noise transmission over optical link, as well as showing early promise for certain computational tasks such as recurrent neural network implementation and analog signal processing [2-3]. However, the limited size scaling of photonic devices for a given wavelength and the wide variety of materials and processes utilized for photonic integrated circuits and active photonic devices (emitters, amplifiers, and detectors) limits the flexibility of design for monolithically integrated devices [4]. It is anticipated that a heterogeneous approach will be required for secure, timely, and cost effective development of state of the art, low to mid volume RF photonic devices by allowing components with optimal performance characteristics for each system subtask to be selected and integrated using standardized methods similar to electrical RF interconnects today [5].
In order to provide true flexibility of design without excessive costs and low throughput of traditional optical assembly methods, it is desirable to develop an optical chip-to-chip interconnect that can adapt or be immune to misalignments naturally caused during die placement and assembly steps while providing low loss connections. Similarly, optical assembly for interface into optical fibers for module-to-module interconnection is a primary driver of cost and throughput for optical module production. Improved methods for creating fiber to chip or fiber to module interconnects would enable increased access to create custom optical modules for system designers.
[1] Blaicher, Matthias, et al. "Hybrid multi-chip assembly of optical communication engines by in situ 3D nano-lithography." Light: Science & Applications 9.1 (2020): 1-11.
[2] Denis-Le Coarer, Florian, et al. "All-optical reservoir computing on a photonic chip using silicon-based ring resonators." IEEE Journal of Selected Topics in Quantum Electronics 24.6 (2018): 1-8.
[3] Minzioni, Paolo, et al. "Roadmap on all-optical processing." Journal of Optics 21.6 (2019): 063001.
[4] Abrams, Nathan C., et al. "Silicon Photonic 2.5 D Multi-Chip Module Transceiver for High-Performance Data Centers." Journal of Lightwave Technology 38.13 (2020): 3346-3357.
[5] Estrada, José Antonio, et al. "Metal-Embedded Chip Assembly Processing for Enhanced RF Circuit Performance." IEEE Transactions on Microwave Theory and Techniques 67.9 (2019): 3537-3546.
SF.35.21.B0001: Nanoscale Vacuum Field Emission Devices
This research topic seeks to extend the state of the art in achieving robust 2-terminal (diodes) and 3-terminal (transistors) nanoscale vacuum field emission devices, that are suitable for low power digital, RF amplification, and high power switching circuits intended to operate in austere and space environments and/or exceed performance of existing solid state technologies. Predictable device operation over long time scales and elevated temperatures is a key metric to enable technology transition. We are interested in concepts that span both traditional and non-traditional materials but are largely amenable to leveraging existing wafer-scale microfabrication technologies. Modelling and simulation based design, fabrication, and experimental characterization of fundamental emission mechanisms, devices, and fundamental circuits is all within the scope of this topic.
Keywords: Field emission devices, transistors, vacuum
SF.35.20.B0006: Heterogeneous Integration Technology
This research initiative is pursuing novel approaches towards implementing advanced 2D, 2.5D, and 3D integrated System in Package (SiP) and System on a Chip (SoC) microsystem solutions. Primary interest is on the advancement and minaturization of microwave and mm-wave applications as the preponderance of commercial investment is towards high-speed digital. Towards this end, integration of individual components including microelectronics, photonics, MEMS, micro-antennas, and RF passive components are all of interest. Efforts to enhance electrical, RF, thermal, and mechanical performance of the interconnects and packaging methods used for existing and novel heterogeneous integration approaches are all of interest.
References:
Bayraktaroglu, B., Heterogeneous Integration Technology Technical Report, AFRL-RY-WP-TR-2017-0168
Keywords:
Heterogeneous Integration, System in Package, System on a Chip, Packaging
SF.35.20.B0005: Sample Size Requirements and Method for Estimation in Hierarchical Linear Models
Traditionally, the complexity of autonomous human-machine systems are qualitatively assessed or quantitatively simplified with the use of linear regression due to limitations in achievable sample sizes, subject availability, cost, and access. These approaches do not address the complexity of human-machine systems and their inherent hierarchical structure. Hierarchical Linear Models (HLMs) can analyze the complex variability inherent in the multiple levels of organizational structures.
Prior to this research, complex human-machine systems, such as ISR PEDs, have only been assessed up to a Level-2 HLM. To determine the appropriate sample sizes per level needed to model and understand performance of human-machine systems, this research derived the standard errors for Level-2, Level-3, and Level-4 HLMs under different conditions. This research seeks to further develop novel extrapolation techniques for small samples to provide inference for performance on higher level and complex hierarchical model.
Assessment of human-machine system performance is of most importance to the Air Force and the need for autonomous systems is growing. Due to rising operational costs, changing missions, small populations, complex environments, and absence of necessary data, the need for HLM assessments through simulations are crucial. Prior sample size recommendations and small sample size extrapolation are a step towards advancing the Air Force`s capabilities and mission understanding for human-machine based systems.
SF.35.20.B0004: Hyperspectral Imaging
AFRL is interested in developing hyperspectral imaging (HSI) technology for automated target detection and identification. This area includes both hardware and software research covering the visible through longwave infrared (0.4-14.0um) spectral region. We are interested in novel imaging spectrometer technology that can reduce cost, size, weight, and power (C-SWaP) with respect to existing technology. Similarly, we are interested in designs demonstrating improvement in sensitivity, spectral resolution, spectral range, field-of-view, amongst other measures of performance.
AFRL is also interested in development of associated exploitation algorithms to improve target detection and identification capabilities. Specific areas of focus would include calibration, non-uniformity correction, atmospheric compensation, bad pixel correction, false alarm mitigation, target detection, material identification, classification, dimensionality reduction, and unmixing.
SF.35.20.B0003: Switchable constituents integrated with exploitable nanophotonic cavity elements
The objective of our research program is to develop novel electro-optic sensing and infrared components by exploring light-matter interactions of photonic and plasmonic structures in the mid-infrared (MIR) and long-wavelength infrared (LWIR). Of particular interest is the development of angle-insensitive MIR filters, selective perfect absorbers, and switchable/tunable devices using modeling and simulations as well as experimental investigation. Spectrally selective transmission or absorption is enabled via critical photonic coupling into gap plasmon, epsilon-near-zero, or other resonant cavity modes. The particular materials to be integrated include, but are not limited to, those with known phase-change or actively tunable optical properties such as germanium antimony telluride, vanadium oxide, and indium tin oxide. Such tunable materials and resonant cavity structures remain relatively unexplored in the mid- to long-wave infrared, these regimes are of specific interest to the Air Force for identification of heat signatures and specific emissions as part of countermeasure systems. The revolutionary science being undertaken will provide the ground breaking tools and nanophotonic structures that lead to light-weight integrated agile IR countermeasures, threat warning, tracking & identification systems. Keywords: nanophotonics, nanoplasmonics, metasurfaces, infrared, sensing, detection, phase change materials
SF.35.19.B0002: Organic/Inorganic hybrid structures for quasi-phase matching frequency conversion devices
High power broadly tunable laser sources operating in the mid and longwave IR region are in great demand for various civilian and military applications, including IR countermeasures, laser radar and reliable IR communications. However, the small number of available direct laser sources which, in addition, are limited in power and tunability has led to seeking alternative approaches, such as phase and quasi-phase matching (QPM) conversion of frequency. A QPM frequency conversion device is based on a thick structure from a nonlinear optical (NLO) material such as GaAs, GaP, ZnSe, or GaSe, grown on the “so-called” orientation-patterned (OP) templates. Recently, heteroepitaxy of GaP grown on OPGaAs templates and GaAsP grown on both OPGaP and OPGaAs templates turned out to be a successful approach, resulting in thick device quality QPM structures. Showing that combing materials could be fruitful made us to take the next step towards combining two different (types of) materials – organic with inorganic materials. Although that to date the research has been focused mostly on inorganic materials, some organics have been also studied for the same applications. What are the major differences: First, both the template preparation and the growth of an inorganic material are difficult, expensive and hardly reproducible. They suggest the usage of poisonous, flammable or corrosive precursors, extreme growth conditions and high-tech machineries such as MBE and HVPE. In contrast, the fabrication of the most organics is usually a harmless process that can be conducted at moderate growth conditions in a low-tech growth environment. Second, organics may have much higher nonlinear susceptibility than the related inorganic alternatives. However, many of them cannot be grown in larger sizes. Many of them are, also, barely transparent in a large portion of the IR region, cannot handle high enough temperatures, have lower thermal conductivity or are soft and volatile to chemicals, including photoresists. By growing/fabricating hybrid organic-inorganic QPM structures we aim to get to a point when the two materials mutually will compensate their shortcomings taking advantage of their most useful properties. For example, learning how to grow or introduce an organic material in one already grown or fabricated inorganic half-matrix we eliminate the need of growing large organic crystals. At the same time we protect chemically and mechanically the organic areas. We also allow the inorganic half with its high thermal conductivity to act like a heat sink, to cool down the organic half and thus to keep high enough the required laser threshold. As initial steps in this research efforts we are planning, first, to determine what kind of organic materials are more suitable -- crystalline, semi-crystalline, polycrystalline, amorphous, including polymers, or even liquid crystals. Do they have transparency range that is similar to the related inorganic partnering material? Can they handle the heat coming from the pumping beam without deteriorating their optical properties? Can they wet the inorganic surfaces and what surfactants we shall use to improve this property? In-situ or post growth polarization schemes shall be also investigated in order to achieve such polarization of the organic half which is comparable to the natural polarization of the inorganic half of the matrix. From here, good transparency throughout the oppositely oriented domains of the matrix should be pursued as a precondition for realizing wave mixing and frequency conversion. The plan for the later stages of this research is improving the IR transparency of chosen organic materials by introducing heavy ions in the molecular structure in an attempt to balance their molecules which in many cases consist of a heavy and a light part.
SF.35.19.B0001: Two- and Three-Dimensional Data Alignment of Measured and Synthetic SAR Data
In order to incorporate synthetic aperture radar (SAR) imagery into a multi-sensor detection and recognition system, two fundamental problems need to be solved: 1) we need to be able to use synthetic data to develop fusion algorithms to enable fusion at the feature and signal level and 2) we need to understand and quantify the uncertainty associated with SAR detection and recognition processing. To attack the first problem, one of the main challenges we face in contested, or anti-access/area-denial (A2/AD) environments, is being able to augment the small set of available measured data with a large set of synthetically generated data. Due to the assumptions made about scattering phenomenology and background clutter, synthetic data often serves as a misrepresentation of measured SAR data. We aim to use two- and three-dimensional sparsity-based techniques in a Bayesian framework to discover a representation of our data that best aligns corresponding synthetic and measured data collects. This data alignment procedure will need to be autonomous and allow for adaptable pre-processing of data before training and testing of machine learning (ML) classification algorithms. To attack the second problem, we will seek to understand the uncertainty associated with the decisions being made by ML classification algorithms leveraging the Bayesian representation developed to align the measured and synthetic data. This uncertainty understanding is a necessary step before incorporating SAR automated target recognition (ATR) outputs into a decision or feature level fusion network that may combine ATR output from other sensors such as electro-optical (EO) and laser radar (LADAR).
SF.35.18.B0005: Multimodal machine learning for Aided target recognition (AiTR)
Multimodal machine learning aims to build models that can process and relate information from different modalities such as EO, IR, Hyperspectral and RF. In recent years, machine learning has been successfully applied to multimodal learning problems, with the aim of learning useful joint
representations in data fusion applications. In this effort, we are interested in developing the state of the art machine learning algorithm which will include but is not limited to deep learning based approaches to address the challenging problem the US Air Force faces in the realm of Multi-INT Aided Target Recognition (AiTR), and publish our findings in peer reviewed Journals and conference proceedings. Delivery of any source code developed is expected.
SF.35.18.B0004: Scaled Stellar Interferometry Laboratory Research
Comprehending the physics-based relationships involved in optical stellar interferometry in order to perform scaled laboratory experiments investigating novel concepts on topics such as fiber based beam relay effects, beam combination for fringe detection, atmospheric mitigation and interferometer component risk reduction research. The design of experiments using physical modeling and simulation tools is desired in developing methods for investigating novel concept research in a scaled laboratory environment. The formulation of optical interferometry experiments will be applied to future defense applications. Recent technology advancements in fiber design and optical sensor technology have revealed opportunities to investigate new component choices applied to long baseline optical interferometry not previously studied which could yield low cost improvements to existing long baseline optical interferometer design. Our long term research goal is to enable low cost solutions to optical stellar interferometer component design that support defense applications in space object characterization and space situational awareness.
Keywords: Atmospheric turbulence and propagation; Optical stellar interferometry; Optical design; Optical fiber design; Optical sensing; Research test and evaluation; Space Situational Awareness; Synthetic Image Reconstruction
SF.35.18.B0003: Growth and characterization of two dimensional materials and heterostructures
Our research focuses on the development of promising two dimensional (2D) electronic materials and their heterostructures for the next wave of nanoelectronics. The wide variety of available 2D materials gives a full tool box of materials properties to explore. We are actively developing growth, fabrication and characterization methods of 2D hBN, graphene, phosphorus materials, heterostructures and devices. Research emphasis is 1) to develop methods, models, and understanding of vapor phase growth of these atomically thin van der Waals layers and 2) to study the effects of interlayer interactions between 2D and 3D materials on electronic transport and optical properties. A combined experiential and theoretical approach to study growth and properties will help realize the fully potential of these materials. Our laboratory operates several CVD and MOCVD reactors for growth of 2D layered materials. These growth techniques are complemented by a variety of structural, optical, and electrical characterization tools, as well as, fully equipped nano fabrication facilities.
SF.35.18.B0002: Adaptable Cyber Security
The ability to sense, learn, and adapt to cyber security threats is required to provide mission assurance to the warfighter. This project will focus on research and development of one or more key enabling technologies necessary to build self-protecting cyber security systems, namely: (1) the development of cyber sensors that will identify malware and/or architectural or implementation defects in the target platform, including, but not limited to developing active sensor probes to accelerate contextual learning for use in subsequent adaptive response, (2) the ability to determine access to discovered defects (e.g., via trace-based analysis) and associated triggering mechanisms , and (3) the ability to respond to unanticipated attacks by providing self-adapting mechanisms, including the ability to repair or replace defective software or firmware, or to disrupt malware triggering mechanisms in order to provide resiliency to ongoing attacks. The end-goal of this project is to develop a software or hardware-based proof-of-concept that demonstrates one or more of the above capabilities.
SF.35.18.B0001: Applying Immune System Methodologies to Embedded System Malware Detection and Response
This proposed research project focuses on exploring biological innate and adaptive immune systems (e.g., human or bacteria) and applying the underlying methodologies being used by these systems to develop a cyber immune system. The focus of this research project is to explore the underlying mechanisms that immune systems use to distinguish self from non-self and the application of those mechanisms to detect and respond to malware that has been inserted within the software and firmware supply chain of embedded systems.
Research in the first phase of this project should lead to a comprehensive understanding of the mechanisms that immune systems use to discriminate host DNA from foreign DNA (i.e., self vs. non-self) in order to apply the underlying process or approach to distinguish computer malware (Trojans) from the host software in which it is surreptitiously embedded. Research in the second phase of this project should then focus on using the previously discovered methodologies to develop a bio-sequencing or other binary analysis approach that can recognize foreign DNA sequences in a software or firmware applications that could contain embedded malware. Finally, research should investigate an approach to malware response that involves precise editing of the foreign DNA to destroy, disable, or prevent the execution of the malicious Trojan, in much the same manner as genetic engineers and biologists are using genetic editing to correct DNA defects that cause disease in humans and animals. The desired end-goal of the project is to develop a software and/or hardware prototype that uses the underlying mechanisms found in immune systems to perform software/firmware supply chain malware detection and response.
SF.35.17.B0006: Deep Generative Models for Closed Loop Sensing
The principal focus of current deep learning research has relied on supervised learning where an extensive number of labeled training examples (many 100s per class) are required. However, AFRL is interested in settings where knowledge representations can be developed from sensed data that can be shared across multiple related sensing tasks and which are robust to unseen mixtures of nuisance parameters. For such representations disentanglement of the causal factors in measured data is required. A number of unsupervised learning approaches have been proposed in the literature for learning these representations to include variational auto-encoders and generative adversarial networks. The objective of this research is to explore such unsupervised approaches for learning representations for closed loop sensing. Publication of the resulting work (both on the sensor's directorate internal ATRPedia and in public venues) is expected, along with the delivery of all source code developed. Keywords: unsupervised learning, deep generative models, attention, active vision.
SF.35.17.B0005: Cognitive Captioning for AF Data
Keywords: Deep learning, captioning, meta-data, video processing, image processing, recurrent neural network
Deep learning techniques are currently producing state of the art (SotA), results in common image and video exploitation tasks. Of particular interest are existing image and video captioning systems. Such results are attained through the use of great quantities of representative data for training consisting of pairs of imagery/video and human-written descriptions.
To date, the data that has been utilized is public in origin. Both the motivation and goal of this effort is to ascertain the performance of these existing SotA tools on AF mission-relevant data and, consequently, improve on the SotA. This particular effort will apply deep-learning approaches to a combination of data-types (e.g. imagery and text) to learn relationships between representations, ultimately producing accurate and improved semantic descriptions of scenes observed from AF-relevant platforms.
Publication of the resulting work (both on the sensors directorate`s internal ATRPedia and in public venues) is expected, along with the delivery of all source code developed.
SF.35.17.B0003: Synthesis and Surface Characterization of Si1-x-yGexSny Alloys
The goal of this research project is to develop Si1-x-yGexSny for sensor applications: 1) remote sensing applications such as IR countermeasures, remote hyperspectral imaging, and high sensitivity chemical/biological weapons detection; 2) as a potential broad NIR, MWIR, and LWIR detector material for use in laser RADAR; 3) as a CMOS-compatible material for development of electro-optic integrated circuits (EOIC) including smart pixels for advanced focal plane arrays (FPAs); and 4) for free-space optical communication. The large size difference of these group IV elements presents unique challenges which require the development of a new approach to allow synthesis over a broad range of composition. Achieving this goal will require substantial effort in process development. Investigation of the gas phase chemistry and associated growth kinetics will provide a fundamental understanding of the surface physics and reaction chemistry involved in this process. Synthesis will be closely coupled with characterization of the basic structural, surface, electrical and optical properties of the resulting films to inform the growth optimization process.
SF.35.17.B0002: Investigating a-Sn (Grey Tin) for semiconductor Devices
The focus of this lab task is measuring the electrical and optical properties of a-Sn and a-SnGe alloys to assess their suitability for electrical and optical devices. Tin is a group-IV element that can assume the diamond crystal structure (the a-phase) and become a zero-gap semiconductor with a band gap opened with strain. Moreover, strained a-Sn has been shown to be a topological insulator, a newly discovered class of materials that have an insulating bulk but metallic, spin-polarized surface bands. These properties make a-Sn a candidate for several applications, such as high-speed, low-power electronics; long-wavelength IR detectors; spintronics; and, quantum computing. Our technical approach is to synthesize a-Sn films by molecular beam epitaxy (MBE) on nearly lattice-matched substrates to produce highly crystalline films in well-defined strain-states and, consequently, adjust the band gap and associated electronic properties. We measure these properties using magneto-transport and optical methods.
We require expertise in the modeling of band-structure of narrow-gap semiconductors and the interpretation of multi-carrier transport measurements. We also require expertise in the modeling of far IR optical measurements.
SF.35.16.B0006: Metamaterials, Plasmonics and Photonics for Novel and Enhanced Sensing
Research opportunities are currently available for the theoretical and experimental investigation of novel sensing schemes. Optical metamaterials, plasmonics and photonics technologies will be utilized to provide enhanced sensing that could ultimately be the backbone of the next-generation Detectors and Focal Plane Arrays (FPAs). The wavelength regimes of interest are short-wave, mid-wave and long-wave infrared. We are interested in the study of novel sensing mechanisms that would lead to more sensitive and lower noise devices. The study could include any aspect of the design and synthesis of the materials to the architecture of the device. The study at the very minimum should include simulations with the quantitative proof-of-principle, but ideally experimental demonstration should be included through fabrication and characterization. Simulation, Fabrication, and Characterization facilities will be provided.
Keywords: Infrared, Sensing, Detectors, Focal Plane Arrays (FPAs), Metamaterials, Plasmonics, Photonics
SF.35.16.B0004: Aspects of Robust Sensing via Cognitive Radar
A cognitive radar (CR) framework is defined by its exploitation of feedback, memory, and prediction on multiple scales or levels in order to enhance the traditional adaptive/fully adaptive/knowledge-aided radar frameworks. In addition, a cognitive approach often features coupled or jointly designed system architectures. These features increase the computational and design complexity of a given system in order to increase the adaptivity and robustness of the sensor(s). Therefore, care must be taken to properly mitigate the risks of a coupled design featuring feedback (e.g. stability, learning saturation, etc.). However, such a design methodology offers the potential to adapt to a dynamic environment, spectral or otherwise, in a robust, efficient manner.
Our interest is in applying the CR framework to enable robust sensing in a distributed, netted cognitive radar network (CRN). As such, we are interested in designing fully adaptive radar networks that can maintain sensitivity in the face of degraded sensors, intentional and unintentional RFI, heterogeneous clutter, as well as numerous other challenging scenarios. The distributed nature of the network also requires adaption to the difficulties inherent in multistatic sensing (e.g. multistatic clutter, waveform orthogonality/separability, communication/collaboration requirements, etc.). Finally, these networks should show the capability to learn and adapt to unforeseen circumstances based on past experience.
Keywords: Cognitive radar; Fully adaptive radar; Distributed sensing; Netted systems; Radar network; Multistatic radar; MIMO radar; Information fusion; Decentralized processing
SF.35.16.B0003: Photonic Integrated Circuits for Photons with Orbital Angular Momentum
Investigations are underway to exploit orbital angular momentum (OAM) of light with a focus towards quantum based photonic integrated circuits. Current OAM communication links utilize liquid crystal based spatial light modulators which are relatively bulky and expensive. We are exploring alternative, chip scale platforms for creating integrated OAM transmitters and receivers with an eye towards information transfer with improved security and with increased robustness to transmission through turbulence. Novel methods of incorporating active tuning into an OAM photonic integrated circuit will provide increased functionality of the devices. Design, numerical modeling, fabrication, and characterization are being pursued, as well as exploration of potential algorithms and protocols for quantum information and quantum processing.
SF.35.16.B0002: Optical Properties of 2D Materials
Atomically thin transition metal dichalcogenide materials, with their remarkable optical and electronic properties, are proving to be promising candidates for quantum based photonic integrated circuits. Several research groups have recently demonstrated single photon emission from two dimensional tungsten-diselenide (WSe2) [1,2,3,4], with second order correlation function values well below 0.5 [1]. The origin of the single photon emission comes from defect states in the material, although little is currently known about the nature of these defects. It is plausible to assume that such localized emitters exist also in other transition metal dichalcogenides (TMDCs). In order to fully realize the potential of these materials, a more detailed study of the optical and electronic properties of these defect states is required. We are studying such defect states in a variety of two dimensional materials, such as MoS2, MoSe2, MoTe2, WS2, WTe2, both theoretically and experimentally. In particular, we are going to investigate whether the operating temperature of the single-photon source can be increased above 10 K, and whether the emission of the single photons can be triggered on demand for the potential use in quantum information technology. One of the main advantages of 2D materials for single-photon sources is that defects can be created by means of standard lithography methods. In addition, while single photons from 3D materials need to travel through the host media with a large refractive index, single photons can be extracted from 2D materials much more efficiently.
References:
[1] A. Srivastava et al., Nature Nanotechnology 10, 491 (2015).
[2] Y.-M. He et al., Nature Nanotechnology 10, 497 (2015).
[3] M. Koperski et al., Nature Nanotechnology 10, 503 (2015).
[4] C. Chakraborty et al., Voltage-controlled quantum light from an atomically thin semiconductor, Nature Nanotechnology 10, 507 (2015).
SF.35.16.B0001: Highly Tunable RF Resonant Microsystems
This research seeks to extend the state of the art in achieving highly tunable narrowband (high-Q) passive and active resonant devices (including bulk acoustic wave, surface acoustic wave, suspended MEMS technologies). We seek to improve system performance characteristics for RF filtering and signal generation/mixing including tuning range, insertion loss, power, and operating spectrum. Addressing considerations for the operating environment including elevated temperature and vibration is also of interest. Exploring and leveraging non-traditional and engineered materials solutions to develop solutions is within scope, with preference for approaches that are amenable to direct integration with RF microelectronics. In-house research involves modelling and simulation of device performance, device fabrication, and characterization.
Keywords: RF filtering, Tunable, MEMS, Microsystems, Oscillators, Mechanical Resonators
SF.35.14.B0811: Polycrystalline YAG Fiber Characterization
This research seeks to demonstrate a viable Polycrystalline YAG fiber for use in amplifier and laser applications. Opportunities exist primarily in contributing to the characterization of these fibers but also in improving these AFRL-made fibers and working toward the goal of demonstrating a viable laser source using these fibers.
Keywords:
Novel Fiber Characterization; Laser theory; Nonlinear optics
SF.35.14.B0810: Injection-Locked Semiconductor Lasers
This research focuses on studying the injection locking of quantum-well and quantum-dot semiconductor lasers primarily in order to generate narrow-linewidth tunable microwave signals. Here different topologies as well as more fundamental issues (such as the suppression of amplitude and frequency fluctuations) are of critical interest and currently being investigated. Depending on background and interest level opportunities to contribute to one or more of the different areas we are investigating including theoretical studies, experimental work, and/or numerical simulations exist.
Keywords:
Quantum-dot lasers; Mode locking; Laser dynamics; Laser theory; Semiconductor lasers; High-speed test and measurement; Nonlinear optics; Integrated photonic development
SF.35.14.B0809: Micro-Mechanical Oscillators
This research seeks to push the performance of micro-mechanical oscillators for applications including the generation of microwave signals, diverse waveforms, and entangled photons. We are currently characterizing resonators made out of a novel material which seeks to improve on what has be traditionally demonstrated. Opportunities exist to contribute to theory, experiment, and/or simulations depending on background and interest level.
Keywords:
Micro-Mechanical Oscillators, Laser theory; Semiconductor lasers; High-speed test and measurement; Nonlinear optics; Integrated photonic development
SF.35.14.B0808: Printed Electronics and Photonics
This research focuses on using novel fabrication techniques such as ink jet and aerosol jet printing and nanoimprint lithography to print electronic and photonic circuit and device elements. Focus will be on novel materials such as polymers, CNTs, graphene, etc. and on flexible substrates. In-house research involves device and material fabrication, characterization and testing of fabricated structures. Current areas of interest are in field effect transistors and mid-wave IR detectors.
References:
Ha, Mingjing, et al., Nano Lett. 13, 954-960 (2013)
Lee, Seoung-Ki, et al., Nano Lett. 12, 3472-3476 (2012).
Keywords:
Printed electronics/photonics, aerosol jet printing, ink jet printing, flexible electronics
SF.35.12.B1101: Fully Adaptive Radar
Research opportunities exist in physics and phenomenology-based adaptive signal processing methods for enhanced radar target detection and estimation, tracking and classification involving closed loop radar operation. The onerous challenges of harsh environments, difficult targets, and a rapidly shrinking electromagnetic spectrum necessitate a systematic treatment for developing closed-loop radar operation. The concept of fully adaptive radar (FAR) seeks to exploit all available degrees-of-freedom on transmit and receive in order to maximize target detection, tracking and classification performance. This area has received increased interest in recent times and builds upon a rich history of prior research. Of key importance is the concept of closed loop radar operation via feedback. Specifically feedback from the receiver and tracker to the transmitter for guiding the next illumination to better concurrently detect, and track targets of interest in computationally demanding and training data starved scenarios is required. A first step is the development of the feedback signal from the receiver and tracker to the transmitter via prescribed metrics such as mean squared error, entropy, or mutual information. The next step is to develop analytical and computer simulation methods for determining the detection, tracking, and classification performance with respect to single and multiple radar waveforms. Furthermore, due the large number of degrees of freedom, the number of unknown nuisance parameters incurs a substantial increase. Consequently the curse of dimensionality prevails. Novel approaches for overcoming this issue are of considerable interest. Concepts of machine learning can be brought to bear in a powerful manner in this context. Relevant performance metrics include the tracking error and computational cost. Extension of this approach to handle distributed and MIMO radar performance must be undertaken. Performance validation for both single and distributed radar needs to be analyzed using simulated and measured data sets.
Keywords:
Fully adaptive radar, closed-loop radar operation, concurrent detection, tracking and classification, feedback control design, performance benchmarking and validation
SF.35.12.B0126: Electro-Optical Devices and Technologies for Infrared Sensing Applications
Light sources and detectors are essential components of infrared sensing systems. Also electro-optical systems require various other beam control devices and technologies that are necessary for directing photons from the light sources to the detectors. Our research effort is focused on optical beam control components such as diffractive optical elements, spatial light modulators, optical phased arrays, devices utilizing phase change materials. Though many electro-optical near infrared and shortwave infrared devises are available commercially-of-the-shelf, further work is required to improve the performance of these devices. In the midwave infrared and longwave infrared, the choice of the optical beam control devices and technologies are limited. Therefore, we continue to search for new and innovative ideas beam control technologies for infrared sensing applications. Our applications of interest are: laser detection and ranging, infrared countermeasures, infrared scene projectors, multispectral and hyperspectral imaging systems.
References:
[1] Sarangan, A. et. al., “Broadband Reflective Optical Limiter Using GST Phase Change material”, IEEE Photonics J. 10(1), 2018.
[2] Kononchuk, R. et. al., “Photonic Limiters with Enhanced Dynamic Range”, Proc. SPIE 10513-0W, 2018.
[3] Gulses, A. et. al., “Lasers with Intra-Cavity Phase Elements”, Proc. SPIE 10513-18, 2018.
[4] Anisimov, I. et. al., “Infrared Dynamic Scene Projectors: Technical Challenges and System Requirements”, 2016 GOMAC Tech Conference (Unlimited distribution version of the paper is available from the author upon request).
[5] Anisimov, I., “Developing Innovative Solutions for Infrared Sensing Applications in Collaboration with Air Force”, 2016 IEEE NAECON.
[6] “Laser Radar: Progress and Opportunities in Active Electro-Optical Sensing”, National Research Council, 2014, ISBN 978-0-309-30216-6
SF.35.02.B7398: Scattering and Propagation in Complex Environment
Our primary interests are (1) the statistical models for radar environment such as terrain, ocean, hilly region, forest, atmosphere, ionosphere, and urban region; and (2) relation of the statistical properties of the scattered signals with that of the environment model.
Topics of interest include waves in random media (discrete and continuum models), radiative transfer theory and remote sensing, scattering from randomly rough surfaces, polarimetric scattering, inverse scattering and parameter retrieval from measured data, physics-based models for scattering from terrain and ocean, scattering from media with space-time fluctuations, monostatic and bistatic models for radar clutter, combined random media and rough surface scattering, spectral density of scattering from ocean surfaces, scattering and propagation of radar signals through turbulence in atmosphere and ionosphere, and models for subsurface sensing. Most studies in the literature on these topics are on the derivation of the average scattering coefficients or propagation constants. However, we are interested in more detailed statistics such as probability density function and spectral density of the scattered signals. Often the signals will be of wide bandwidth, so we are interested in the characteristics of scattered signals over a wide range of frequencies. We are also interested in studies on the scattering of targets embedded in complex environment and polarimetric techniques for detecting such targets. The targets may be stationary, mobile, or fluctuating. Of particular interest is the location and imaging of targets embedded in a complex environment.
Keywords:
Random media; Rough surfaces; Imaging; Clutter models; Turbulence; Radiative transfer; Sea clutter; Polarimetry;
SF.35.02.B7397: Sensors Aboard Hypersonic Vehicles
Hypersonic flights lead to high temperature flows, air dissociation, and cumulative heating of air-frames. Consequently the performance of all on-board sensor systems such as GPS, telemetry, communication, command and control, radar, ladar, and electro-optical sensors are all adversely affected to varying degrees by the hypersonic environment. Further, the dynamic range of parameters that characterize the environment is quite large and is strongly influenced by many factors including altitude, velocity, duration of flight, geometry of the vehicle, airframe, and heat-shield material. For instance, the electron density can vary by several orders during the course of a trajectory. Hence, sensor systems encounter a variety of situations. Some of the issues encountered include signal attenuation, communication blackout, signal distortion due to turbulent flow, radiation from heated optical windows, and emission from hot flows. Communication blackout although old is still a problem and is encountered when the signal frequency is well below the plasma frequency. An adaptive sensor system which uses a diagnostic tool to sense and adaptively match to the environment will be desirable. Even in the case when the signal is above the plasma frequency the boundary layer flow can be dispersive, inhomogeneous, fluctuating, and lossy. This poses challenges to wideband RF system using conformal arrays. With arrays we also face the problem of mutual coupling and impedance mismatch effects on beam forming. Moreover, the signal transmitted from the vehicle may be sufficiently intense to initiate nonlinear processes in the flow. The rapid maneuvers and high velocity place limitations on the integration time of the processing algorithms of the receivers. Although optical sensors are not similarly affected by hypersonic flow as RF sensors they have their own share of issues. The hot window can radiate at infrared frequencies and the hot flow fields can emit and absorb at optical frequencies thereby seriously affecting the optical and EO/IR sensors on board. Two major problems of concern are beam pointing error and wave front distortions. We are particularly interested in assessing the impact of environment on spectral measurements and imaging. Research opportunities exist in the analyses and mitigation of the above-mentioned issues confronted by sensors aboard hypersonic platforms.
Keywords:
Communication blackout; hypersonic boundary layer turbulence; Diagnostic tools; Adaptive sensors; Nonlinear processes; optical windows
SF.35.02.B5790: Phenomenology-Based Adaptive Radar Signal Processing
Research opportunities exist in physics and phenomenology-based adaptive signal processing methods for enhanced radar target detection and estimation. Classical adaptive signal processing methods for radar rely on the formation and inversion of a sample covariance matrix. However, nonstationary reflectivity properties of the scanned areas, dense target environments, and strong clutter discretes tend to introduce heterogeneities in the training data for covariance estimation. These tend to have a deleterious impact on detection and false alarm performance. Furthermore, as the dimensionality of the problem increases, the training data support for forming the covariance matrix and the computational cost of the matrix inversion are prohibitively high. To address these issues we seek to exploit a priori information from the scattering physics or phenomenology underlying a given scenario. For example, in many instances clutter can be viewed as the resultant of the scattered power from a small number of strong interference sources, thus rendering it low rank. This information can then be advantageously used to reduce the training data support and computational cost of the resulting adaptive processing algorithm. The problem of target detection is further complicated by the presence of a large number of nuisance parameters. These effects are exacerbated by systems and environmental considerations pertaining to the operational scenario. We seek novel approaches based on either a priori knowledge of the clutter scenario or on principles of invariance for this problem with a goal to maintain a constant false alarm rate and achieve robust target detection performance. Development and performance analysis of MIMO radar signal processing algorithms for the above described scenarios is of particular interest.
Keywords:
Adaptive radar signal processing; Physics-based methods; Constant false alarm rate; Sample support; Invariance; Robust performance; Knowledge-based methods
SF.35.02.B5355: Integration of Knowledge and Sensor Data for Intelligent Actions
Research opportunities exist in the area of advanced methods for integrating knowledge and data coming from a variety of sources and sensors with the goal of developing intelligent systems capable of extracting actionable information from data and knowledge. Sensor development and data extraction is only one part of enabling the decision support processes. Discovery and disambiguation of the system of interacting agents coordinated within the set of goals is the essence of information construction required for successful decision-making. Between data and actions there are processes of information extraction based on knowledge. Networks of knowledge, interwoven with intentional systems of goal-oriented agents, ought to account for the available options in decision-making based on data, knowledge, phenomenology, and computational processes.
Knowledge includes physical models of sensors, wave propagation and scattering, statistical models of object properties, dynamical models of motion, linguistic text models, semiotic models of meaning, and cultural models of human behavior in a society of interest. Corresponding to the available knowledge, the models might be detailed or approximate, reflecting precisely known physical laws or uncertain intuitions about undiscovered phenomena or human nature. The integrated functioning of an intelligent system is not a one-time deal but a continuous loop of operations in which sensors and data collection are directed based on the current-moment results; models and actions are continuously refined.
The relevant systems have substantial affinity with the known mechanisms of brain and intelligence of living creatures. Whereas, the past algorithms combining knowledge and data often encountered prohibitive combinatorial complexity, the mind can do it. The mind avoids combinatorial complexity by combining conceptual understanding with emotional evaluation. We hope to develop algorithms utilizing non-combinatorial measures of similarity between models and data, resembling affective-emotional capabilities of the mind in combination with modeling-conceptual capabilities.
Biology suggests that intelligent systems are evolving systems. Evolution is a desirable property of the AF systems, so that they improve with experience, rather than become obsolete. Genetic evolution inspired the development of genetic algorithms. Other evolutionary mechanisms suggested by cultural evolution operate faster than genetics. We are interested in evolving network-centric sensory systems inspired by biological evolution of communication and language systems.
Keywords:
Integrated systems; Intelligent agents; Evolution; Evolving systems; Similarity measures; Combinatorial complexity; Affective computation; Emotional computation; Conceptual computation; Models of mind; Semiotic models;
SF.35.02.B0229: Plasmonic and Metamaterial Enhancements for Multispectral Sensing
Our interests include electro-optic, magneto-optic devices, components, sensors and focal plane arrays operating in the near infrared to the millimeter wave spectrum. Our current research focuses on improved signal detection, polarization control, or other enhancement to photodetectors and focal plane arrays operating in the mid-wave to long-wave infrared with specific interest in multispectral sensing. For example, we are studying the use of metamaterials and plasmonics to enhance the performance of quantum dot photoconductors and focal plane arrays for multispectral sensing. We are using the metamaterial and plasmonics enhancements to reduce the background radiation and increase the selectivity of the photodetectors and focal plane arrays. The work involves experimental investigation as well as modeling of the photodetectors and focal plane arrays. Our facilities include high-end workstations and personal computers, design and characterization capability in the near to long wave infrared. Our characterization capability includes a mid to long wave FTIR, a femtosecond laser tunable from 1 to 20 micrometers for spectral response, photocurrent, dark and noise current measurement setups, and a Photoluminescence measurement setup with sample temperature control form 15 to 300 K. There is access to a full microelectronics cleanroom with nanofabrication capability like high resolution electron-beam lithography and nano-imprinting. There is also access to a high performance computing center, and in-house epitaxy and materials synthesis.
Keywords: Quantum Dots, Plasmons, Plasmonics, Metamaterials, and Spatial Dispersion
SF.35.02.B0220: Advanced Antenna Technology
Research opportunities exist in the area of high-gain and multibeam antennas to support Air Force air- and spaceborne radar systems. We are particularly interested in wideband (>1 octave) and multiband systems, and in developing both new antenna designs and new efficient analysis methods. This applies to directly radiating planar and conformal arrays, as well as array-fed reflector and lens configurations, which offer the potential for an electrically scanned, high-gain beam at low cost.
Other topics of interest include wide-scan wideband planar arrays, arrays for hemispherical coverage, low-profile UHF/VHF arrays on conformal platforms, planar and conformal frequency selective surfaces, efficient electrically small antennas, and applications of metamaterials in novel antenna and lens designs.
A final topic centers on computational electromagnetics for electrically large bodies. We need to improve both frequency and time domain methods, which eventually will permit us to numerically analyze large finite planar and conformal arrays, including mutual coupling and edge effects on large complex platforms, possibly of exotic nonconducting materials including metamaterials.
References
Josefsson L, Persson P: Conformal Array Antenna Theory and Design. Hoboken: John Wiley and Sons, 2006
Bhattacharyya AK: Phased Array Antennas. Hoboken: John Wiley and Sons, 2006
Keywords:
Electronically scanned antennas; Computational electromagnetics; Conformal arrays; Metamaterials;
SF.35.01.B7635: Discovery for Difficult Targets
AFRL is interested in discovering new ways to assemble exploitable difficult target signature attributes from the layered-sensing construct--explicitly, from field-able sensors into products indicative of object type or purpose. Targets of interest include dismounts and in particular, vehicles. Understanding and organizing target data is the first step toward this goal. The goal of this research is to exploit data from Full Motion Video (FMV), infrared (IR), LiDAR/LADAR, Hyperspectral, and Polarimetric IR sensors, individually or in concert, to determine what measurable characteristics, or combinations of measurable characteristics, can be used to classify entities in the data. Once characteristics are identified that allow for classification, sensed data can be tailored to capture the measurements necessary for exploitation and identification. This effort may include a data collection using fixed sensors from a standoff range (greater than one kilometer). In addition to publications (both on AFRL1s internal ATRPedia and public venues), delivery of documented source code will be expected.
Keywords: Feature modeling; Object Classification; Task-driven Sensing; Machine Learning
SF.35.01.B7634: Dual-System Based Formalism for Data to Decisions
AFRL requires innovative approaches to the problem of developing robust subjective representations of sensory data to apply to many Data to Decisions (D2D) problems. Systems that rely on creating an objective representation of the world are prone to fragility as a result of sensor degradations and dramatic differences in input. These systems are characterized by the attempt to capture attributes of the environment under consideration in a physically accurate representation. In contrast, nature has developed a dual system approach that is able to robustly deal with degradation of sensors and incomplete information. System 1 is fast and reflexive in nature, while System 2 is slow and logical. We seek an implementable mathematical formalism that can model these two systems along with their blending--allowing for learning and decision-making based on a subjective representation that is not tied to maintaining fidelity with physics-based reality. Application areas include D2D domains such as target identification and tracking, sensor management, sensor fusion, cyber, and ISHM. Publication of the resulting work (both on ATRPedia and in public sources) is expected, along with the delivery of any source code developed.
Keywords:
Knowledge representation; ISR; Biologically-inspired; Mathematical formalism; Robust decision making; Subjective representation; Information fusion; Data to decisions
SF.35.01.B7606: Multi-agent Sensing Systems
The primary objective of this research is to develop and demonstrate the capability of multi-agent systems to sense an environment and build a semi-coherent and synchronized model of that environment. In particular, we are interested in improvements in two primary areas described below.
First, while there has been significant research in the area of distributed sensing, many practical examples that have been put forward deal with very simple models of the environment. One example is a sensor network detecting the temperature of an area. Because temperature is a single number and highly correlated over space, building a global model is relatively straight-forward. Imagery sensors (cameras), on the other hand, collect huge amounts of data making the communication of this data to all agents in the network problematic. In addition, as the network grows, keeping a full-resolution global model precisely synchronized across all nodes in the network also becomes impractical. Therefore, a hierarchical, semi-synchronized model may be required. Research into generating and maintaining the global model across a distributed multi-agent system when the global model is very large is a significant research problem for multi-agent sensing systems.
Second, when sharing data across agents in a network, overcoming the track-to-track correlation problem while still improving on the performance of any one sensor is fundamental to achieving distributed data fusion (DDF). However, most approaches that overcome the track-to-track correlation problem provide very conservative estimates of the fused data uncertainty. Research into DDF techniques that achieve more accurate estimates without becoming over-confident is of interest when designing multi-agent sensing systems.
SF.35.01.B7518: Heterogeneous Integration Packaging
"The goal of this research project is to develop 3D heterogeneous integration (3DHI) packaging techniques using either traditional or non-traditional fabrication techniques (such as additive manufacturing techniques, polymer materials, etc.). We are expanding in-house 3DHI fabrication techniques, and the development of improved packaging techniques and methodologies for 3D heterogenous integration devices is critical before they can be integrated into functional devices and applications. The performance of these devices will be compared to that of current state-of-the-art devices.
This proposed project can explore the use of traditional fabrication techniques and the use of newer additive tools and processes, such as ink jet printing, aerosol jet printing, nanoimprint lithography, and nScrypt printing, can be used. Existing procedures for electronic device testing will be used to analyze and compare these devices to the current state of the art. It is anticipated that a willingness to obtain multi-disciplinary academic excellence, drawing primarily from electrical engineering, materials science, physics, and chemistry, will be required for research success.
key words
3DHI; heterogeneous integration; fabrication; packaging"
SF.35.01.B7368: Intracavity Laser Dynamics and Diverse Waveform Generation
This research primarily focuses on manipulating light within laser cavities in order to generate diverse waveforms outside of the laser. Work is currently being conducted experimentally by investigating the mode locking and modulation of multi-section quantum-dot lasers. Computationally we are solving systems of partial differential equations based on traveling-wave models which are able to incorporate not only carrier dynamics but spatial effects in the cavity. We have found that the use of delay differential equations provides a compact model still able to capture some spatial effects and our analytical work has focused on this type of model. Opportunities exist to contribute to one or more of the above listed areas depending on one’s background and interest level.
Keywords:
Quantum-dot lasers; Mode locking; Laser dynamics; Laser theory; Semiconductor lasers; Optical waveform generation; High-speed test and measurement; Nonlinear optics; Integrated photonic development
SF.35.01.B7367: Computational Methods for Antenna Designs
This opportunity involves the application of advanced numerical methods for computational electromagnetics towards the modeling of antenna systems. Methods of interest include finite element tearing and interconnect (both time and frequency domain methods), time domain boundary integral methods, and Discrete Galerkin methods. These numerical methods are to be utilized on high performance parallel computers with distributed memory architectures. Applications of interest include antennas on platforms, large phase array antennas, wideband antennas, and conformal antennas. These techniques are to be applied to electrically large systems with the goal of obtaining accurate solutions for multiscale and highly heterogeneous structures. Balancing the burden between user inputs, meshing, and computers to obtain idealized throughput for design studies is also a desired goal.
Keywords:
Computational electromagnetics; Antennas; Parallel computers; Phased arrays; Time domain; Frequency domain; Green’s functions; Fast solvers; Domain decomposition;
SF.35.01.B7313: Waveform Agile Radar Processing (WARP)
Radar systems transmit an electromagnetic signal into a volume of space containing both objects of interest (targets) and objects not of interest (clutter). The signal reflects from both kinds of objects and these reflections are received by the radar along with signals produced by other sources (interference). The radar must then separate the target returns from the clutter and interference. There are currently effective techniques for adaptively suppressing clutter and interference at the receiver. Salient examples include constant false alarm rate (CFAR) detectors, adaptive antenna beam forming techniques, and space-time adaptive processing (STAP). In theory, clutter and interference suppression could be enhanced if the transmit signal was tailored to the target/clutter/interference environment. However, despite potentially significant performance gains, this kind of transmit adaptivity has not yet become a mature technology. This is principally due to the insufficient computing power and limited RF waveform generation hardware that have rendered adaptive transmit techniques impractical to implement. However, the coupling of Moore’s Law with recent advances in arbitrary waveform generation has dramatically improved the prospects of transmit adaptivity as a viable technology. We say that a system is "transmit adaptive" if it is capable of altering its transmit waveform in response to knowledge about its environment. By "environment" we mean those elements that affect system performance, such as targets, clutter, and radio frequency interference (RFI). Knowledge of the environment could be acquired a priori, estimated online, or both. There are three main reasons to investigate transmit adaptivity in radar systems: (1) performance improvement, (2) resource management, and (3) novel missions. The first reason involves improving performance subject to constraints on the transmit waveform (e.g., peak power, nice autocorrelation); the remaining two reasons focus on maintaining a minimum level of performance while either minimizing resources or performing multiple functions. For example, a transmit-adaptive system might be able to compute a transmit waveform that maximizes the probability of detection in a given RFI environment (Reason 1). Such a system might also be able to maintain the probability of detection at a desired level while using a minimal amount of transmitted power and bandwidth (Reason 2). Efficient spectrum usage afforded by such a system would provide an immense benefit in a spectrally congested environment. Ideally, an ATx system would also be capable of performing two missions simultaneously--e.g., spotlight synthetic aperture radar (SAR) and digital communications--while maintaining a minimum level of performance for each function (Reason 3). The multifunction capability afforded by WARP would be of considerable benefit in realizing one or more attributes of the Layered Sensing paradigm. Research opportunities exist in every aspect of waveform design and optimization. This includes theoretical design of waveforms, optimization algorithms, and efficient computation. Transition opportunities to investigate practical aspects of implementation and proof of concept demonstrations are available through other laboratory resources.
Keywords:
Waveform agile radar; Transmit-adaptive radar; Waveform design; Waveform optimization; Simultaneous multifunctional operation;
SF.35.01.B7108: Photonic Circuits for Quantum Information Processing and Computing
Quantum encoding photons and entanglement are non-classical features unique to interacting quantum systems and are considered to be the fundamental resources that underlie the power of quantum computation, quantum communication, and quantum memory and storage. In general, quantum information processing can be considered as the exploitation of engineered quantum systems. Our goals are to exploit quantum encoded photons and entanglement for the purpose of developing quantum information processing systems with the potential to exceed conventional computing capabilities. Quantum computing systems with unprecedented levels of security and computational capability for select applications promise decreased computation times for complex AF problems such as faster optimization of complex AF C2 systems, ultrahigh-speed signal and image processing, and informational data base searches with a quadratic speed-up.
Current research focuses on the development of photonic integrated circuits that will enable the processing of quantum encoded photons and entangled photon pairs with significantly scaled down dimensions from table top experiements. Generation of single and entangled photons, coupling them into photonic waveguides and devices with polarization maintaining and low optical lossess are of interest.
References:
J.C. F. Matthews, A. Politi, A. Stefanov and J.L. O’Brien, “Manipulation of mulitphoton entanglement in waveguide quantum circuits,” Nature Photonics 3, 346 (2009).Baehr-Jones T, e t al: Optics Express 13(14): 5216, 2005
J. W. Silverstone, D. Bonneau, K. Ohira, N. Suzuki, H. Yoshida, N. Iizuka, M. Ezaki, R. H. Hadfield, V. Zwiller, J. G. Rarity, J. L. O'Brien1 and M. G. Thompson, “On-chip quantum interference between two silicon waveguide sources,” arxiv:1304.1490
Keywords:
Integrated photonic circuits; Quantum Information Processing, Quantum Computing, Entangled Photons, Single Photons
SF.35.01.B7107: Computational Electromagnetics and Electronics
This research involves topics in computational electromagnetics and charge thermodynamics-treated both separately and as a fully coupled system. By itself, the electromagnetics component emphasizes efficient full wave methods for electrically large linear problems. Domain decomposition algorithms applied to full wave discretization schemes are considered along with sufficiently accurate field expansion approximations. Applications for these approaches include optical cavities, waveguides, and nanophotonics. The electronics component focuses on charge transport, particularly in the Fermi gas/drift-diffusion approximation. Accurate representation of hot electron effects and the role of semiconductor bandstructure within this formalism are of interest, as well as the coupling of the classical and quantum confined regions of quantum devices. Applications include quantum well and quantum cascade lasers, detectors, and wide bandgap HFETs. Research interests in such opto- and micro-electronic components extend to high frequency wave effects requiring a fully coupled treatment of vector field electromagnetics and charge dynamics.
Keywords:
Full wave electromagnetics; Charge transport; Drift diffusion; Charge thermodynamics; Semiconductor devices; Quantum well laser; High frequency; Hot electrons; Computational physics;
SF.35.01.B7063: Computational Physics for Radar Signal Processing
Research opportunities exist in computational science to design, implement, and test algorithms for enhanced radar designs. Advances in radar signal processing techniques are increasingly using the physical details of the signal environment. We are particularly interested in advancing novel computational electromagnetic algorithms for use on the most challenging radar problems. These challenges include using radio frequency signals for underground sensing as well as for sensing applications throughout the ionosphere.
Reference
Sotirelis P, et al: Proceedings of the IEEE 2009 National: DOI 10.1109/NAECON.2009.5426633, 2009
Keywords:
Computational methods; Physics-based methods; High-performance computing; Computational electromagnetics; Full-wave; Multiphysics;
SF.35.01.B7062: Distributed Netted Radar Signal Processing
Research opportunities exist in distributed, netted radar systems for detecting difficult (low cross section, low Doppler) targets embedded in severe clutter backgrounds. A distributed network of radars systems, including multiple input multiple output (MIMO), consists of multiple radar stations netted together through data communication links. Traditional netted radar systems normally comprise several monostatic radar systems, each operating at a different carrier frequency to avoid the interference and detection confusion among the radar stations in the system. As a result, the multiple radar stations in the system are incapable of operating in a bistatic or a multistatic mode. The strong capabilities of bistatic and multistatic radar systems are well known. An improved netted radar system is one that can simultaneously operate in both bistatic and monostatic modes. In this case, all radar stations are assumed to operate at the same carrier frequency and are coordinated and controlled by a coherent, signal-level fusion processing unit. Such a system not only significantly improves radar performance in target search, tracking, and recognition, but also effectively addresses emerging radar challenges. However, many fundamental challenges exist and need to be addressed. We seek novel methods of waveform design (specifically orthogonal or quasi-orthogonal waveforms) for proper radar functioning. New and robust adaptive processing techniques capable of operating in both bistatic and multistatic environments are also sought. For airborne and spaceborne applications, the geometry-induced Doppler dispersion significantly degrades the performance of adaptive processing techniques such as Space-Time Adaptive Processing (STAP). Multistatic clutter characterization techniques are also desired for better understanding of the impact of clutter on detection performance in these environments. Finally, depending on whether all data is sent to the fusion processing unit for detection purposes (centralized processing), or individual detections made at the sensor level are sent to the fusion center (decentralized processing), new signal processing techniques are desired that cohere signals across multiple distributed radar apertures.
References
Fishler E, et al: IEEE Transactions on Signal Processing 54(3): 823, 2006
Chen CY, Vaidyanathan PP: IEEE Transactions on Signal Processing 56(2): 623, 2008
Keywords:
Distributed sensing; Netted systems; MIMO radar; Waveform design; Centralized processing; Decentralized processing; Bistatic/multistatic STAP; Fusion processing;
SF.35.01.B7061: Integration of Multicomponent Oxide Materials and Sensor Applications
Our research focuses on the development of oxide materials and the incorporation of these oxide materials in high performance transistor applications. Current investigations include beta-Ga2O3 MOSFETs, ZnO and InGaZO thin film transistors, and impurity doped ZnO thin films for transparent conductors.
The research entails device design, fabrication, and testing, as well as optimization and analysis of material microstructural and morphological properties to achieve enhanced device performance with particular emphasis on ALD gate dielectric, ohmic contact and associated interface development. Research is conducted in a fully equipped ISO-5 device fabrication clean room. Materials characterization capabilities include electron microscopy, ellipsometry, x-ray diffraction, atomic force microscopy, Hall effect and photoluminescence.
References:
A.J. Green, K.D. Chabak, E.R. Heller, R.C. Fitch, M. Baldini, A. Fiedler, K. Irmscher, G. Wagner, Z. Galazka, S.E. Tetlak, A. Crespo, K. Leedy, and G.H. Jessen, “3.8-MV/cm Breakdown Strength of MOVPE-Grown Sn-Doped beta-Ga2O3 MOSFETs,” IEEE Electron Device Letters 37, 902 (2016).
M. Schuette, A. Green, K. Leedy, A. Crespo, S. Tetlak, K. Sutherlin, and G. Jessen, “Ionic Metal-Oxide TFTs for Integrated Switching Applications,” IEEE Trans. Electron Devices 63, 1921 (2016).
B. Bayraktaroglu, K. Leedy, and R. Neidhard, “High Frequency ZnO Thin Film Transistors on Si Substrates,” Electron Device Letters 30, 946 (2009).
Keywords:
Thin film transistor; MOSFET; Zinc oxide; Beta-Ga2O3; Atomic layer deposition; Pulsed Laser Deposition; Nanotechnology; Power Semiconductor Devices
SF.35.01.B6994: Semiconductor Laser Source Phenomenology and Development
Applications in active sensors and infrared countermeasures may often be significantly improved by unique sources. We theoretically and experimentally investigate high-brightness semiconductor lasers with unique properties such as broad tunability, mid-wave and long-wave operation, fast modulation, and multi-wavelength operation. Devices such as vertical external cavity surface emitting lasers combine the inordinate semiconductor optical gain with a classic macroscopic laser cavity, which results in a device that can achieve signal spatial-mode operation under many watts of optical output powers approaching in any wavelength achievable by semiconductor lasers. We are interested in longer wavelength semiconductor lasers in uncommon operation regions, such as the 2-4-micron band through interband Type-I and Type-II lasers, large conduction-band offset intersubband lasers. It is essential to understand material systems (such as GaAs- and GaSb-related compounds) to appropriately design a laser cavity. For example, Auger recombination and carrier confinement both become extremely influential at high power and temperatures. Mitigation of these deleterious effects is essential for efficient operation.
Interest areas include material design and development, fabrication, coupled resonator designs, passive mode-locked lasers, spatial temporal waveform generation schemes for semiconductor lasers, as well as nonlinear frequency conversion. Fabrication facilities are available including semiconductor growth, typical microfabrication techniques, electron-beam lithography, as well as numerous metal and dielectric film deposition techniques. A full testing capability suite is available for cryogenic through high-temperature measurements for device operation.
Keywords: Lasers; Semiconductor; Infrared; High-brightness; Resonator; Unstable beam; Nonlinear;
SF.35.01.B6593: Nonlinear Optical Materials and Devices
We investigate and develop new nonlinear optical devices, primarily for frequency conversion applications. Specifically we are seeking tunable coherent sources in the mid-infrared (2-5 micron) and far-infrared (8-12 micron) spectral regions. We are also interested in beamsteering, modulation, and other techniques that improve the performance and utility of frequency conversion devices. Current projects focus on quasi-phasematched (QPM) materials, specifically orientation-patterned semiconductors, in both bulk and waveguide form; although other QPM materials, innovative birefringently-phasematched crystals, and techniques for enhancing performance in more established materials are also of interest. Experiments address optical characterization of candidate materials, demonstration and performance characterization of nonlinear devices, and performance modeling, with the goal of improving material fabrication processes and device designs to optimize performance. Facilities are available for measuring absorption, fluorescence, and scatter, as well as thermo-optical properties. A variety of lasers are available for use both in material characterization and as pump sources for frequency conversion and other nonlinear devices. Optical parametric oscillator testbeds are available, along with equipment for measuring device output power, spectral content, beam quality, and other performance parameters. AFRL also has a class 100 clean room and facilities for fabricating periodically poled ferroelectric materials, for bonding semiconductors and other optical materials, and for MBE growth of GaAs.
Keywords:
Nonlinear frequency conversion; Nonlinear optical materials; Quasi-phasematching;
SF.35.01.B6351: Multichannel Radio Frequency Sensor Processing
We develop radio frequency sensor processing hardware and software techniques. The use of multiple channels on transmit and receive enables multiple Degrees of Freedom (DOF) processing and Digital Beam-Forming (DBF). Research interests fall into the following categories:
(1) Novel multichannel architectures. The goal is to achieve broad spatial and frequency coverage utilizing multiple simultaneous beams. Architectures should be scalable in number of channels and capability (number of beams, bandwidth coverage).
(2) Sensor signal processing techniques. We are particularly interested in techniques that reduce the per-channel computational requirements, DBF processing requirements, and DOF reduction processing requirements for wide instantaneous bandwidth systems. We are also interested in performance versus cost tradeoffs for adaptive and non-adaptive techniques.
(3) Software compensation schemes. Many multichannel signal processing techniques assume perfectly matched hardware channels. It is desirable to have digital/software compensation schemes that correct for mismatches, or to develop processing techniques that are impervious to channel mismatch. We are interested in research in compensation schemes for Analog (Amplifiers, mixers, transmission lines), and mixed signal (Analog-to-digital Converter) component non-linearity, as well as inherently robust processing algorithms.
Keywords:
Digital receivers; Waveform generation; Digital beam forming; Radio frequency signal processing; A to D converter; Sensor processing;
SF.35.01.B6350: Investigation and Optimization of Existing and Novel Electronic Devices
The goal of this project is to investigate and optimize the electrical performance of existing devices, devising and investigating novel electronic devices (e.g.,nanoelectronic devices). Investigation of these devices through electrical characterization, limited physical characterization, and modeling will enable a better understanding of their limitation and potential. The information produced in this project will push the limits of the current state-of-the-art in electronics.
An example of this research includes, but is not limited to, the optimization of AllnN/GaN HEMTs by investigating changes in electrical device performance due to changes in device fabrication. For this example, components that could be studied are silicon nitride dielectric passivation, ohmic and Schottky contact formation with variations in metal stack components, and surface treatments prior to metal deposition.
References:
Gillespie A, et al: Solid-State Electronics 49(4): 670, 2005
Jessen GH, et al: IEEE Electron Device Letters 28(5): 354, 2007
Keywords:
Transistor; Power; Compound semiconductors; Nanoelectronic; Device; FET; HEMT; Microwave;
SF.35.01.B6349: Radio Frequency Micro-electromechanical System Design, Fabrication, Packaging, and Characterization
Ebel, J - 937-713-8698
Micro-electromechanical systems (MEMS) switches offer numerous advantages for radio frequency (RF) systems over competing technologies. Several MEMS technology advantages include low insertion loss, high linearity, and reduced size. Although the reliability of RF MEMS devices has improved dramatically over the last several years, deficiencies in several fundamental areas associated with the technology persist. Areas that we are currently working on include (1) wafer-level encapsulation of the switches, (2) dielectric charging characterization, and (3) metal contact damage evolution. Wafer-level thin-film encapsulation provides discrete device packaging to protect the switches from the environment. Current research efforts focus on the sealing of thin-film encapsulants and characterization of the package hermeticity. The electrical performance and reliability of RF MEMS capacitive switches is dependent on the properties of the switch’s dielectric material. We are conducting experiments to quantify the charging and discharging behaviors of the dielectric materials for capacitive switches. We are also interested in modeling the contact wear mechanics of the mating surfaces for metal contact MEMS switches. Model validation is anticipated through examination and testing of fabricated devices. Device processing and test facilities are available to fabricate and characterize the RF MEMS switches. Device fabrication is completed in an on-site clean room facility. RF probe stations and an environmental chamber are used for electrical characterization and hermeticity experiments.
Keywords:
MEMS; Radio frequency; Dielectric charging; Thin film; Metal contact; Encapsulation; Packaging;
SF.35.01.B5915: Autonomous RF Front-End Enabling Components and Techniques
With the increasing proliferation of tightly integrated digital control algorithms, sensors and actuators, and corresponding circuitry, reconfigurable/multifunction RF circuit components are viewed as enabling components for future autonomous systems. Future system applications in contested environments depend on the capability to sense, learn and adapt in real time based upon mission and environmental factors. Envisioned transmitter operating conditions and scenarios will certainly require differing metrics with regard to frequency, bandwidth, gain, output power, transmitted noise levels, linearity, efficiency, and peak-to-average power ratio (PAPR). For receivers, the ability to negate unwanted jamming signals, in addition to frequency, bandwidth, gain, noise levels, linearity, dynamic range, DC power consumption and survivability are important parameters that can be traded off depending on the scenario.
To this end, research focuses on enabling components and techniques for autonomous front-ends. Waveform and frequency agility are desired characteristics. For transmitters, maintaining efficient operation is vital for reducing thermal constraints and prime power requirements. For receivers, the ability to operate in the presence of jamming signals would be advantageous. Potential topic areas of interest include, but are not limited to, novel architectures, switching elements, signal injection/cancellation techniques, sensors, actuators, control algorithm implementation, and modeling at both the device and system sub-system level.
SF.35.01.B5858: Bio-Inspired Systems for Intelligent Information Processing (IIP)
Bio-inspired passive radar systems incorporate intelligent information processing (IIP) to identify a wide radio frequency (RF) signal processing application range from detection to image identification. Bio-inspired IIP focuses on mixed-signal radar, microwave, coherent source, wavelets, millimeter wave, polymorphic, self-organization, self-assembly, nanoprocessors, and nanosystems of systems architectures to enable real-time information extraction and context processing. The objective of this research is to address a broad spectrum of information fusion, algorithms, and embedded processing for developing IIP passive radar architectures for both receiver and transmission technologies.
Keywords:
Bio-inspired; Radar; Real time; Mixed signal; Wavelets; Polymorphic; Fusion algorithm; Context processing; Nanosystem; Radar processing; Radio frequency signal processing; Millimeter wave; Microwave and passive sensing;
SF.35.01.B5814: Precision Navigation Reference Sensor Techniques
Our goal is to research and develop the latest technologies to provide precise Postion, Navigation, Attitude, and Time (PNAT) for military systems operating in any environment at any time. PNAT is the cornerstone in providing an assured reference for distributed sensing and warfare. Research includes investigating inertial sensor technologies and their integration with secondary sensors (preferably passive) that provide precision measurement updates, as well as optimization of the reference solution for the specific distributed or staring sensing task. Potential secondary sensors include jam resistant global positioning system (GPS) technologies, optical sensors using passive image data, light detection and ranging (LIDAR) sensors using ranging data, and software based receivers and navigation using natural signals (e.g., the Earth’s magnetic field).
Current in-house research partially focuses on exploration of military use of global navigation satellite systems (GNSS) and sensor-inertial fusion in constrained environments where GPS may not be available. A critical piece of this research effort involves determining the optimal method for integrating various navigational aiding sensor data into a precision navigation system.
Keywords:
Assured reference; Navigation techniques; Precision navigation; GPS receiver technology; Navigation aids; Software defined radio; Inertial navigation; Global Navigation Satellite Systems (GNSS)
SF.35.01.B4960: Advanced Digital Receiver/Aperture Development
Our ultimate goal is to realize fully digital multifunction active arrays with programmable functionality. In this architecture, only the high power and low noise amplifiers at the aperture are analog circuits. Wide-bandgap semiconductors are candidates for these high-power and linear amplifiers. All the amplitude and phase-shifting functions, as well as the beam-forming, are done digitally. Transmit waveforms are generated by direct digital synthesizers and the received signals are captured with very high-speed analog-to-digital converters with deep dynamic range and wide frequency response. True time delay for steering very large wideband agile arrays is also digitally implemented. This architecture requires very high-speed, high-throughput data processing. Commercial simulation tools can facilitate the design of the array aperture and radiators, as well as the components used in the active phased array. Behavioral simulation methodology allows performance prediction of experimental components used in the architecture. Affordability, size, and power consumption are important factors to include in the trade space, but may be considered secondary to advanced performance capabilities. Military RF sensors must operate in extremely dense signal environments (including intentional, unintentional, and co-channel interfering signals). UHF and VHF radars for Foliage Penetration and ground penetration are especially vulnerable because of the dense array of commercial emitters, limited aperture spatial discrimination, wide instantaneous bandwidths, and long integration times required. Research in all areas of RF system design mentioned above is sought.
Keywords:
Microelectronics; Microwave antennas; Microwave electronics; Solid-state electronics;
SF.35.01.B4480: Electro-Optical Sensor Research
Our research focuses on the following:
(1) Algorithm development and signal processing for laser radar. Techniques of interest include, but are not limited to: turbulence mitigation, phasing of coherent apertures, streamed processing high bandwidth waveforms, photon counting arrays with reduced computation, micro-scanning, stitching of imagery, and data exploitation.
(2) Eye-safe transmitters for coherent laser radar. The objective is to develop efficient laser-based systems capable of gathering multidimensional images, where the dimensions can include spatial (x,y,z), spectral, polarization, and vibration characteristics. Transmitters with agile wavelength and with waveform agility are desirable, as they can be used for multiple imaging modes. Transmitters need to be capable of high peak powers with minimal stimulated Brillouin scattering effects for continuous wave and especially pulsed applications.
(3) Shot noise limited coherent receivers for laser radar. Full waveform reception multi-mode laser radar signals are desired. Also of interest are techniques which can help to implement conformal/distributed coherent apertures, including non-mechanical beam steering and photonic integrated circuits among others. Receive apertures that can be used for transmitting or passive imaging at other wavelengths as well are preferable due to the increased functionality.
Keywords: Electro-optical imaging; Laser radar; Ladar; Optical phased arrays
SF.35.01.B2220: Radar Systems Theory
This task involves research and theoretical analyses concerning the development of new techniques and radar system concepts (spatially and spectrally diverse radars), which provide greatly improved performance for ground, air, and space surveillance (passive and active) radars; synthetic aperture (and “all target” imaging) radars; sparse arrays; multi-static radar; distributed RF sensors; and multiple aperture interferometric radars. Current topics include suppression of range-angle-Doppler ambiguities, analysis of ambiguity induced limitations from an information theoretic viewpoint, interferometric clutter cancellation and interferometric height discrimination, automatic focusing, preprocessing for clutter and electromagnetic interference suppression prior to space-time adaptive processing, and multifrequency waveform diverse radars for persistent surveillance combined with high fidelity imaging and feature extraction for processing, exploitation and dissemination (PED). Extensive computer modeling, simulation facilities, and experimental data sets are available to support this research.
Keywords:
Adaptive process; Clutter; Dual frequency; Modeling; Radar; SAR; Simulation; Target recognition;
SF.35.01.B1528: Active Electro-Optical Automatic Target Recognition
Description: Recent advances in airborne synthetic aperture lidar (SAL) systems have allowed for the collection of higher-resolution 2D/3D images and point clouds from longer stand-off ranges. This technology is highly beneficial for operations in denied environments but currently has a need for advanced automatic target recognition (ATR) and classification algorithms. This topic seeks to develop effective algorithms for advanced ATR and classification for active electro-optical sensors (i.e. Laser radar, SAL, etc.). Data can be provided and delivery of any developed source code is expected.
SF.35.01.B1520: Plasmonic and micro-/nano-resonant structures for sensing applications
The major objectives of this work are to design, simulate and fabricate plasmonic and micro-/nano-resonant structures and devices for optical and photonics sensing applications. Examples of such structures may include metallic media, semiconductor patterned structures, Fano-resonant antennas, plasmonic metamaterials, or resonant nanocavities. Theoretical models and computational simulations will be developed for these structures to describe the electromagnetic behavior as well as to optimize these devices. The resulting fabrication techniques and modeling methods will lead to new technology for devices that can be tailored for specific sensing schemes for optical signal enhancement and detection. It is anticipated that the modeling/fabrication/characterization efforts will be iterative towards the development of a highly sensitive optical sensing platform that is robust and wavelength scalable. Keywords: Plasmonics, Nonlinear Optics, Resonant sensing, Photonic devices,Metamaterials, Near-IR, Mid-IR
SF.35.01.B1123: Effective Autonomous Fusion Multi-Sensor Multi-Look Data and Autonomous Learning
Research opportunities in autonomous information fusion, autonomous learning and the optimal decision making.
Effective autonomous modeling and understanding of many real word situations and scenes also requires complex hierarchical representation of the sensor information. Data in different space and time scales and modalities is coming from a variety of sensors and other types of information collection methods with different limitations and levels of plausibility. The meaningful information could be extracted after analyzing the hierarchy of multiple facts and concepts. For typical complex learning algorithms, especially multi-layered neural networks, the processes in the inner layers are usually not transparent and have no clear relationship with the low level patterns they recognize. In contrast to current machine learning algorithms, a successful human learning experience produces a clear explanation for the facts or hierarchy of linked concepts that have clear meaning and can be "recycled" in other learning and explanatory processes. Perhaps the absence of clear associations with meaningful concepts in the machine decision steps cause resistance toward acceptance and trust in the results produced by machine learning algorithms.
We are trying to formulate a solution based on transparency principle for hierarchical machine learning algorithms which can make result inferences produced by the algorithm more trustworthy and understandable by humans even in case of lack of training data. Hence, it is our hope that this will lead us toward better synergy and cooperation with machines.
Keywords: unsupervised learning, hierarchical learning, visualization
SF.35.01.B0915: Categorical Uncertainty Modeling for Situational Reasoning
Recent advances in mathematical modeling include the development of novel models for knowledge representation and reasoning based on category theory (e.g., the work of Spivak on OLOGS) and the emerging field of homotopy type theory. Incorporating uncertainty modeling into these abstract models is important for these methods to be useful for Air Force layered sensing applications, but this requires generalizations of classical probability theory. We seek theoretical advances and tractable, implementable algorithms for modeling and reasoning with uncertainty in abstract relational mathematical models that can be used to support situational representation and understanding. Publication of the resulting work (both on ATRPedia and in public sources) is expected, along with the delivery of any source code developed.
SF.35.01.B0118: Solid-State Laser Source Development
We investigate and develop new infrared (1.5-12 μm) materials, devices, and technologies that have potential for high lasing efficiency, broad wavelength tunability, and robust operation. Research includes exploration and development of new solid-state laser media such as Cr2+ doped into ZnSe and analogous semiconductor host materials. Lasers with direct emission throughout the infrared spectrum and room-temperature operation are sought. Current projects include demonstration of high-power modelocking, Q-switching and CW operation of Cr2+:ZnSe, demonstration of sub-ns pulse operation of a Tm μ-chip oscillator-fiber amplifier, and demonstration of > 2-μm fiber lasers. Rapid tuning, beam control via guiding structures, and beam switching are also performance issues. Infrared lasers (bulk, fiber, and semiconductor types) are also being developed for use as pumps of new quasi-phasematched nonlinear media such as orientation patterned gallium arsenide. Areas of interest include spectroscopy of infrared laser materials, beam transport in infrared fibers and waveguides, and infrared nonmechanical beam steering. Facilities are available to perform spectroscopic measurements of absorption, excited-state absorption, fluorescence, fluorescence lifetime, and ultrashort pulse characterization. A variety of pump lasers are available including Nd:YAG lasers, Tm fiber lasers, Tm,Ho:YLF Q-switched lasers, and semiconductor lasers for measuring spectral content, power, beam quality, and thermo-optic properties of new laser materials.
Keywords:
Solid-state lasers; Tunable lasers; Infrared lasers; Fiber lasers; Infrared fiber transmission;
