Cold atoms are used to make precision measurements of rotation, acceleration, time (think atomic clocks), etc.  Atoms are typically cooled using laser cooling techniques to temperatures less than 100 micro Kelvin to reduce thermal noise and increase measurement time.  We are investigating new cold atoms sources that either reduce the size, weight, and power (SWAP) requirements of laser cooled sources or that can cool atoms without lasers.  Reducing SWAP is an important considering for making compact devices that can be transitioned from the laboratory into real world applications.  Cooling atoms without laser opens the possibility of using atoms that currently cannot be cooled but have properties that are of scientific and/or technical interest.   The assigned project will depend on the student’s interest and experience.

Cold Atom Sources

AFRL Scholars Program

Scholars

Virtually every spacecraft flying uses photovoltaics (aka solar cells) to provide electrical power. Advances in single crystal multi-junction, thin-film, and nano-technology based photovoltaics are important to achieving improved on-orbit performance. The current state of the art photovoltaic cells used for space applications are based on the III-V material systems (eg. GaAs, GaInP). However, innovative and novel material systems capable of more effectively utilizing the solar spectrum could provide tremendous advantages for space missions. For example, recent work examining nano-structures have shown interesting advancements. Some areas of interest for this topic include space environmental effects, electro-optical properties, and performance parameters for candidate material systems. Specific projects can be tailored to students’ interests and skills, but will require a basic understanding of electrical, optical, and materials properties. Students selected for this research opportunity will work with their mentor to develop a productive work plan which will be synergistic with their field of study and the Air Force Research Laboratory’s mission.


Advanced Photovoltaics for Space 

This project will involve the computer simulation of specific aspects of high-power fiber laser amplifiers. The project will compare and assess the reliability of different modeling approaches, especially in regards to fiber coiling effects. One goal will be complete a parameter sensitivity study for uncertainty quantification purposes. There is no prerequisite to understand lasers, fiber amplifiers, nor nonlinear optics - the relevant information will be taught as the project progresses.


Amplification of Lasers 

The Hybrid Architecture Demonstration is aggregating commercial, international, and government space assets into a single architecture to take advantage of the strengths and weaknesses of each part of the currently disjointed space domain. This project will focus on mapping, understanding, and reporting on the cutting edge exploitation and space-based sensor capabilities that are available, if they have the potential to fill capability gaps for the HAD, and where they fit in on the road map and in the architecture.


Analysis of Space-Based Hybrid Architecture 

Compressive single-pixel imaging is a signal processing paradigm wherein high-resolution scenes are 
recovered from highly subsampled, randomly spatially multiplexed data acquired on a single detector. 
While this allows for scenes to be imaged at sub-Nyquist rates, recovery becomes an iterative process 
with high compute requirements. For this reason, extracting information from the randomly multiplexed 
measurements themselves, without reconstruction, is highly desirable. One valuable piece of 
information is whether a static scene experienced any significant changes, called anomalies, over the 
course of the measurement. In this project, the student will develop novel methods for detecting and 
localizing anomalies from single-pixel measurements using time-series analysis and neural networks.

Anomaly Detection with Single-Pixel Cameras 

Fault detection systems are an essential component of satellites. The current method relies on expert knowledge identifying the likely failures along with long checkout windows. This project will be to investigate different techniques and create a code base that can be applied in our multi-agent quadcopter testbed.


Applying numerical methods to improve fault identification on satellites 

AFRL/RVS has established a dedicated architecture analytics testbed under the Space Performance Analytics and Computing Environment Research (SPACER) project. The objective of this project is to provide AFRL with an organic, in-house, capability to assess processing options for next generation mission applications. This addresses the increasing challenge of mapping mission requirements to hardware and software implementations for space computing applications. The topic provides several summer research opportunities for students interested in optimization and evaluation of mission application code on space hardware. The selected summer scholar will be given mission application algorithms/code and tasked to examine methods for optimization of the code (parallel constructs, optimized libraries, etc.) and compilation/execution of the software on hardware resources. This will allow for detailed analysis of the application’s performance on different hardware architecture alternatives, providing critical insight into the computational requirements for next generation mission applications. This effort will provide a valuable capability to the Air Force, guiding future science and technology investments decisions


Architecture Analytics for Next Generation Space Applications 

Low Earth Orbit (LEO) in Earth's atmosphere contains a number of atomic and molecular species, both uncharged and ions. As a spacecraft operates in LEO, the atmosphere serves as a source of drag on the velocity of the spacecraft leading to altitude variations over time. One significant source of this drag is the atomic oxygen concentration, which can vary significantly with various conditions (e.g. daytime vs nighttime) . The measurement of neutrals often involves instruments with significant size, weight and power requirements which often make it impractical to make in situ space measurements. In this project, the student will examine atomic oxygen collisions with various materials in a high vacuum environment so as to impart a response on either the surface or the oxygen collision partner that can lead to quantification of the number density of atomic oxygen. The student will compare and quantify the measured response to the state-of-the-art capabilities and assess whether these materials improve the state-of-the-art or should no longer be explored.


Atomic Oxygen - Surface Interactions 

Recent advancements in deployable structures are enabling mission concepts for large deployable apertures, i.e. antennas and solar panels. However, large deployable structures can have high flexibility, with modal frequencies approaching the bandwidth requirements for even low-precision attitude control. The fact that the flexible modes of deployable structures can be difficult to model prior to launch makes the controller-design problem even more difficult. The goal of this project is to develop control methodologies that exploit the large trade space in sensors, actuators, and control objective. Sensors can be limited to traditional attitude and angular-velocity sensors, or also include measurements of the flexible-body motion (i.e. strains or displacements). Actuators can be limited to applying a single torque to the spacecraft, or allow multiple actuators distributed across the deployable structure. Control objectives can be simply to avoid exciting the flexible-body motion, or to actively control these perturbations. Research projects may address one or more of these topics.


Attitude control of flexible-body spacecraft 

Air density variations around and airborne directed energy system distort a beam’s wavefront, resulting in degraded performance after propagation into the far field. Adaptive optics (AO) can be used to correct for these rapidly evolving aero-optical aberrations; however, in some conditions, the inherent latency between measurement and correction in state-of-the-art AO systems results in significantly reduced performance. Predictive AO control methods utilize future state predictions to compensate for rapidly evolving distortions and are promising techniques for mitigating this limitation. Accurate predictions of flow is a challenging problem, and any solution must be sufficiently light-weight to operate at the high bandwidths of an AO system. This research opportunity will involve characterization and development of new methods for producing short-term predictions of high speed flow to improve the performance of beam control systems for applications such as high energy lasers, communications systems, and astronomical seeing.


Beam Control Machine Learning Research 

Processing requirements for future spacecraft are expected to be very high. For example, high resolution sensors are expected to produce data at rates too high to transmit all the data to ground. Other applications generate less data, but require complicated decision making with little intervention from the ground. Size, weight and power restrictions associated with operation in space are forcing us to examine the capabilities of a broad range of processing architectures ranging from the traditional - e.g., multicore processing, many core processing with graphics processors (GPUs), Field Programmable Gate Arrays (FPGAs), Digital Signal Processors (DSPs) - to the novel - e.g., neural networks, memristor-based circuits, approximate computing, etc. Students working on this project will aid in the development and benchmarking of model processing kernels that will allow us to compare the capabilities of these architectures for various applications.


Benchmarking Processing Algorithms for Space 

With renewed interest in lunar exploration and commercialization, there is a need for novel methods and algorithms to promote space situational awareness in the cislunar regime. A software development environment is needed to process cislunar space situational awareness data efficiently. This research would focus on implementing and analyzing data structures, software architectures, and user interfaces to support an anticipated AFRL flight experiment flying to the vicinity of the Moon in the near future. Potential scholars are strongly encouraged to contact the mentor for more information.


Cislunar Data Processing/Software Architecture 

With renewed interest in lunar exploration and commercialization, there is a need for novel methods and algorithms to promote space situational awareness in the cislunar regime. This regime presents a variety of challenges including vast distances, unique exclusion zones, and chaotic system dynamics. This research effort is focused on the implementation and analysis of algorithms to process optical cislunar data products with a specific focus on data association, correlation, and state estimation. The research would support the data collection and processing for an anticipated AFRL flight experiment flying to the vicinity of the Moon in the near future. Potential scholars are strongly encouraged to contact the mentor for more information.


Cislunar Space Situational Awareness Analyst 

Our laboratory is developing inertial navigation sensors and clocks using cold-atom technology. The relevant experiments and prototypes for cold-atom sensors typically require specialized timing-control hardware, software, and computer integration. We are developing means to utilize micro-controllers, low-power computers, field-programmable gate arrays (FPGA), and open-source software for data acquisition and experimental control. The assigned project will depend on the student's interest and experience, but could include micro-controller programming, FPGA programming, low-power computer-experiment integration, and/or development and testing of new cold-atom control routines/hardware for use in future or current experiments.


Cold Atom Experimental Control and Data Acquisition 

The subject of our research is precision timing and inertial sensing enabled by advances in atomic physics and photonics. One of our major efforts pursues the development of robust, miniaturized optical clocks including optical frequency combs. These portable clocks will provide a cost-effective replacement for the atomic clocks aboard the global positioning satellite constellation and will enable new applications of precision timing such as free-space optical time transfer.

Another major effort uses atom-chip devices to develop confined atom interferometry, which offers the possibility to dramatically increase the interrogation time of atom-based inertial sensing devices. We use a rapid prototyping technique for atom chips that allows us to quickly customize and live-test atom chip structures. We are also exploring techniques for developing compact atomic devices, including large-diameter hollow-core fiber guiding, integrated atom-chip transport, and on-chip micro-optical cavity integration.

Cold-Atom Precision Timing and Inertial Navigation 

The applicant will apply a wide range of cyber tools to satellite flight software and sensor firmware in efforts to reduce attack surface and vulnerabilities to these systems. The target software architecture, source code, programming language, etc. will need to be investigated in order to apply the correct tool-set. The result will be a processes for application of said tools to specific targets and produce cyber-resilient software. As well, the applicant will need to perform simple Cyber Vulnerability Assessment of said targets to determine success of the tools.


Cyber-hardening of satellite flight software 

The Advanced SatNav Technologies (AST) program researches next generation satellite navigation (SatNav). Research areas include: advanced SatNav signals and signal exploitation, spacecraft SatNav payloads, and SatNav command & control systems. Technology areas include: digital and RF signal processing, software defined radios, RF signal generation and broadcast, encryption, and command and control technologies. Research is performed both in a simulated environment and in the laboratory.The AST program is developing signals for the next generation SatNav systems, including GPS. These signals will close weaknesses in existing GPS signals, such as provide authentication of civilian signals, rapid acquisition of military signals, improved performance of military signals in jamming situations, etc. Research is ongoing both in the development of new signals (and modification of existing signals) and the development of the receiver algorithms necessary to fully exploit the advanced signal features.


Data Structures for Advanced Satellite Navigation Signals 

The space domain asset population is in the midst of explosive growth, with
the number of space-based assets expected to quintuple by 2030. This is due
to a growing deployment of collaborating small satellites with on-board
sensing and computing technologies. However, as the number of
spacecraft-to-spacecraft interactions is expected to increase, there are
(and will be) limited human-in-the-loop resources available. Thus, these
space systems require autonomous capabilities on-board to carry out a
variety of operations in the space environment. This could include (but is
not limited to) rendezvous and docking, object-of-interest sensing/tracking,
satellite constellation reconfiguration, etc... All of these applications
require decision-making on the part of one or multiple agents to satisfy
mission requirements. This research effort is directed at the inclusion of
autonomy into these decision-making processes. High-level concepts include
information theory, belief fusion, expert reasoning, collaborative tasking,
and the incorporation of modeled and unmodeled uncertainties in the
operating environment. Potential scholars are strongly encouraged to contact
the mentor for more information, as well as to discuss specific research
ideas for the summer.

Decision-Making in the Space Domain 

The Advanced SatNav Technologies (AST) program researches next generation satellite navigation (SatNav). Research areas include: advanced SatNav signals and signal exploitation, spacecraft SatNav payloads, and SatNav command & control systems. Technology areas include: digital and RF signal processing, software defined radios, RF signal generation and broadcast, encryption, and command and control technologies. Research is performed both in a simulated environment and in the laboratory.

Design of Advanced Satellite Navigation Signals 

Recent advances in camera technology have led to the development of the
Event-Based Sensor (EBS), which has the potential to overcome limitations in
traditional frame-based cameras. Specifically, since these devices only
respond to changes, they offer the potential of dramatic improvements in
traits such as dynamic range, high-speed operation, and bandwidth
requirements. Because of the relatively recent technological advances, their
utility for many applications remains unexplored. This project will
specifically look at the potential of event-based sensors to outperform
traditional frame-based cameras in complex applications including target
detection and tracking for both military and consumer applications.
The Directed Energy Scholar would investigate the capabilities of the EBS
within these applications. Specifically, the work would be experimental,
theoretical or both. In the experimental track, the Scholar would design and
conduct experiments with an EBS, with options for both lab and field
experiments. In the theoretical track, the Scholar would support EBS
research by conducting modeling and simulation at either the system-level or
the component-level. In each case, the scholar would have the opportunity to
identify and develop new EBS applications.

Directed Energy Applications for Event-Based Sensors 

AFRL/RDHEC specializes in development of HPM sources via high fidelity simulation of device physics.  HPM sources we are currently developing include conventional magnetrons, relativistic magnetrons, recirculating planar magnetrons (RPMs), and others.  We are also interested in gigawatt-class amplifier HPM sources.  New ideas are welcome, students will have the opportunity to virtually prototype their own source using the DoD ICEPIC simulation code and DoD supercomputers if they wish, or can contribute by modifying existing designs.  Opportunities exist also for students that are fluent with programming to help improve the tools used for HPM design, which we are also actively developing.


Design of High Power Microwave Sources Using High Performance Computing 

All major military operations rely on space. Space operations are inherently software centric (arguably the one domain where software is most critical). The software world moves entirely too fast for our traditional acquisition paradigm (5+ years timeline from concept to operational use). In order to be relevant in software we need to go from “wild idea” to deploying in operations in 18-24 months. This requires streamlining and automating the process from code level to human level, including integration, security, safety, test, and training. We are seeking to implement a process of Continuous Operational Validation (DevSecDT/OTOps) by bringing people, processes and software technologies together on a two-week development and operations cadence for testing/stress-testing all of the above (people, processes, technologies). We are developing a live and live-virtual-constructive environment for operationally testing Space Control, Space C2 and Space Domain Awareness capabilities. Experience with software languages/developer tools/applications: JavaScript, NodeJS, Python, C++, Unity, Unreal, C#, Grafana, Tableau, Kubernetes, GitLab, Kibana, ElasticSearch, Amazon Web Services, Atlassian suite (JIRA/Confluence), Relational Databases (MySQL), NoSQL, Kafka, REST services, Salesforce CRM Experience with agile software development methodologies Experience with DevOps concepts Experience with CI/CD pipelines (Continuous Integration/Continuous Delivery) Systems Engineering mindset and desire to optimize operational workflows Aerospace Engineering experience, coursework in orbital mechanics Desire to gain experience in space operations (space surveillance, space object tracking, command and control) (The ideal candidate will have a combination the skills/experience listed above)


Dragon Army Operations: Continuous Validation for Operational Space C2 Technologies 

We are developing a laboratory plasma device called EMPD (ExB drifting Magnetized Plasma Device. EMPD is a 4' dia. by 6' long vacuum chamber with 0-200 Gauss axial magnetic field "bottle" for scaled plasma experiments. We are currently developing plasma sources, experiments, and diagnostics to explore plasma-plasma coupling and Langmuir Probe physics in the positive bias flowing plasma regime which is still poorly explored, with relevance to applications as the Electro-dynamic Space Tether.


Dynamic Plasma Coupling in Laboratory, Computer, Space 

[1] Light-electron interaction and optical response of semiconductors;
[2] Effect of electron scattering on transport in semiconductors;
[3] Controlling propagation of electromagnetic waves in conducting materials;
[4] Coupling of lattice vibrations and electron transport;

Effects of Light Interaction and Defects on Photo-Transport Properties in Quantum Wells and Superlattices 

The goal of the research is to investigate the physics of electron field emission under a variety of conditions.  Carbon fiber cathodes will be used for the research.  First, emission characteristics with pulse width from 100ns to 500ms and DC will be examined.  A new cathode test bed will be built for this purpose.  Second, different numbers of emitters (1<n<100) will be used to study the turn-on effects of cathodes, and the phenomena will be modeled using  a predator-prey relationship.  ICEPIC will be used to predict the effects of pulse width on electron field emission, in addition,  various numbers of fibers placed on the cathode will be subjected to the same voltage pulse, and ICEPIC will be used to model the cathode geometry, and comparison in terms of the total current obtain and the value of the field enhancement factors will be made with theory and experiment.  Lastly, a predator and prey relationship model will be developed to study the physics of emission when large number (1<n<100) of fibers are present on the cathode.  During the course of the program, Students will be involved in using ICEPIC to design the geometry of the cathodes, as well as to set up ICEPIC model to study the emission of the cathode under different pulse width.  At the end of the internship, students will gain knowledge in the emission physics from a carbon fiber cathode which is highly relevant in future technology.


Emission Physics of Carbon Fiber Field Emitters 

TEST PREVIEW Recent advances in camera technology have led to the development of the
Event-Based Sensor (EBS), which has the potential to overcome limitations in
traditional frame-based cameras. Specifically, since these devices only
respond to changes, they offer the potential of dramatic improvements in
traits such as dynamic range, high-speed operation, and bandwidth
requirements. Because of the relatively recent technological advances, their
utility for many applications remains unexplored. This project will
specifically look at the potential of event-based sensors to outperform
traditional frame-based cameras in complex applications including target
detection and tracking for both military and consumer applications.
The Directed Energy Scholar would investigate the capabilities of the EBS
within these applications. Specifically, the work would be experimental,
theoretical or both. In the experimental track, the Scholar would design and
conduct experiments with an EBS, with options for both lab and field
experiments. In the theoretical track, the Scholar would support EBS
research by conducting modeling and simulation at either the system-level or
the component-level. In each case, the scholar would have the opportunity to
identify and develop new EBS applications.



Event-Based Sensor Research for DE Applications 

Primum igitur, quid aut officiis debitis aut ad eam non quo quaerimus, non recusandae itaque negat opus esse expetendam et quale sit voluptatem et dolore suo sanciret militaris imperii disciplinam non recusandae itaque negat opus esse expetendam et quale sit voluptatem et dolore suo sanciret

This project will be on modeling high-power fiber lasers/amplifiers. Specifically, the goal is to develop techniques for overcoming nonlinear and thermal effects that inhibit scaling to higher output powers while maintaining good beam quality, and usually narrow linewidth. One method for achieving this goal is through novel fiber designs, which may suppress deleterious processes that occur at high powers in the fibers, but are also complex and challenging to model on a computer. Through this modeling effort various fiber laser/amplifier configurations will be explored in order to optimize the output.


High-Power Fiber Laser/Amplifier Modeling & Simulation 

One of the major quantities of interest for airborne laser systems is the unsteady variation in the pointing direction of the beam, or jitter. In airborne laser systems, the beam jitter is usually thought of as sourced from two components. The component of the jitter caused by mechanical vibration of the optical table, optical elements, etc. is called mechanical jitter. The component of jitter caused by flow structures on the order of the aperture size is called the aero-optical jitter. In typical experiments, information pertaining to the aero-optical component of the jitter is almost always corrupted by mechanical disturbances and is typically removed from the data. For this reason, much is unknown about the nature of aero-optical jitter disturbances. In the Aero-Effects Laboratory at AFRL we study, among other topics, the aero-optical component of the jitter and its relationship to target tracking performance. This study often involves both experimental work, as well as high fidelity simulations of the aero-optical environment. This project will involve closely guided and mentored research into the aero-optical jitter generated by canonical flows.


Investigation into the Aero-optical component of the jitter for airborne directed energy systems

Fault detection on satellites 

We will assemble a team of DE Scholars and AFRL scientists to model/simulate cutting edge gigawatt-class high power microwave (HPM) source technologies. We will utilize the DoD's massively parallel supercomputing capabilities to perform high fidelity simulations using the AFRL's Improved Concurrent Electromagnetic Particle-in-Cell Code (ICEPIC). We will be virtually prototyping and improving the world's most advanced HPM sources, leveraging the latest published results to produce virtual prototypes that represent the best of the world's HPM sources. Team members will work closely with each other and their mentors to learn the basic and advanced features of ICEPIC to create and run successful HPM models in a high performance computing environment.


Gigawatt-class High Power Microwave Source Modeling 

Over the past few years, an increasing amount of interest has been building towards returning to the moon. A second space race of private companies have pledged to send satellites to the Moon within the next decade. However, the dynamics involved in this regime are much more complicated than a typical Earth-orbiting satellite. Scholars applying to this topic will assist in advancing the capabilities of the Air Force in this new regime, developing high-fidelity three-body motion models and assisting in deploying them in a simulation environment.


High-Fidelity Modeling of Three-Body Spacecraft Dynamics 

The Gridded Retarding Ion Drift Sensor (GRIDS) instrument is a CubeSat compatible sensor designed to
characterize ion density, drift, and temperature in the ionosphere. The sensor has a variety of
mechanical and electrical components. Previous iterations of this instrument are will be flown on two
NASA CubeSats, petitSat and Dione. This project will focus on further miniaturizing the instrument by
combining some mechanical components with one existing electrical PCB into a single PCB. Knowledge
and/or experience with schematic capture software is preferred.

Ion Collector Circuit Board Design 

The laser effects modeling and simulation branch (AFRL/RDLE) is tasked with studying the physics and associated phenomenology of laser material interaction to support next generation laser system requirements. Materials of interest include but are not limited to advanced aerospace materials such as metal alloys, high temperature ceramics, composites and semi-conductor materials. AFRL/RDLE constantly strives to develop innovative and novel simulation models to develop predictive capabilities for pre-test analysis and to support military utility studies. This project will focus on developing high fidelity and reduced order laser material interaction models. The research will investigate laser induced failure modes and mechanisms. Numerical methods of interest include but are not limited to Finite Element/Finite Difference (FE/FD), Fluid Structure Interaction (FSI), Magneto Hydro-Dynamics (MHD)/Arbitrary Lagrangian Eulerian (ALE), and Computational Fluid Dynamics (CFD). Interns are not required to have a working knowledge of all of these methods, but rather the mentor will identify the appropriate software based on the intern’s skills and educational background. Custom scripting may be required to modify/update existing models. Development of interaction codes may also require literature reviews for appropriate material properties and development of material constitutive equations. The student will conduct verification and validation analysis with provided analytic, numeric and experimental results. The developed models may be integrated into larger frameworks for wargaming and mission planning.


Laser Effects Modeling and Simulation 

The Local Intelligent Networked Collaborative Satellites (LINCS) Laboratory is a multi-agent robotics laboratory at the Space Vehicles Directorate. It is interested in developing and implementing novel algorithms for collaborative satellite autonomy. In this project, high school scholars will work collaboratively in the LINCS laboratory in a team environment. Prospective scholars should be independent and motivated to learn and gain new skills. They will work closely with senior researchers and other interns to develop an understanding of the robotics platforms, build and design new robotics platforms and sensors to test, learn embedded systems programming and robotics skills.


Local Intelligent Networked Collaborative Satellites Lab 

The Local Intelligent Networked Collaborative Satellites Laboratory is a multi-agent robotics laboratory at the Space Vehicles Directorate. It is interested in developing and implementing novel algorithms for collaborative satellite autonomy. In this project, Phillips Scholars will work collaboratively in the LINCS laboratory in a team environment. Prospective scholars should be independent and motivated to learn and gain new skills. They will work closely with senior researchers and other interns to develop an understanding of the robotics platforms, build and design new robotics platforms and sensors to test, learn embedded systems programming and robotics skills.


Local Intelligent Networked Collaborative Satellites Lab (Phillips Scholars) 

The Local Intelligent Networked Collaborative Satellites Laboratory is a multi-agent robotics laboratory at the Space Vehicles Directorate. It is interested in developing and implementing novel algorithms for collaborative satellite autonomy. In this project, Phillips Scholars will work collaboratively in the LINCS laboratory in a team environment. Prospective scholars should be independent and motivated to learn and gain new skills. They will work closely with senior researchers and other interns to develop an understanding of the robotics platforms, build and design new robotics platforms and sensors to test, implement advanced distributed guidance, navigation, control and autonomy algorithms.


Local Intelligent Networked Collaborative Satellites Lab (Space Scholars) 

This project will explore machine learning/deep learning approaches to performing image processing tasks that are traditionally done using computationally intensive and memory intensive algorithms. Among the tasks to be explored are image classification and feature extraction using simulated image data. Projects will make use of neuromorphic processor chips or graphical processing units (GPUs) depending upon the student’s interest and level of expertise. Software frameworks include PetaVision, CUDA, OpenCL, etc. Machine learning frameworks might include PetaVision, cuDNN, Caffe or TensorFlow.


Machine Learning for Space 

Due to harsh and inaccessible operating environments, systems for spacecraft are subject to many unique challenges and constraints. However, the increasing need for real-time sensor and autonomous processing, coupled with limited communication bandwidth with ground stations, is increasing on-orbit computing demands for next-generation space missions. To address these challenges, AFRL's Space Electronics Technology program has established a dedicated architecture analytics project called the Spacecraft Processing Architectures and Computing Environment Research (SPACER) Laboratory to provide the DoD with the capability to assess on-orbit computing solutions for spacecraft. This project will explore software fault injection on neural networks to assess their reliability in a space environment as part of this effort. This project will introduce students to programming machine learning applications with an emphasis on space applications. The basics of neural networks and computing reliability will be covered along with a light introduction to software fault injection. Fault injection campaigns will then be executed and the results will be analyzed under mentor supervision.


Machine Learning Reliability Studies 

Emerging 5G communications and network technologies can be leveraged to enhance military communication capabilities. In particular, 5G-enabling technologies are envisioned to provide higher data rates, lower latency, lower power consumption, security enhancements and ubiquitous access including non-terrestrial links. The three major use case domains of 5G— enhanced mobile broadband (eMBB), ultra-reliable low latency communication (URLLC) and massive machine type communications (mMTC)—provide the opportunity to harness commercial technology for future AF use cases such as smart bases, self-driving vehicles, augmented and virtual reality technologies for training, dynamic spectrum management and sharing technologies to facilitate coexistence of commercial and military spectrum dependent systems (SDSs). The 5G research areas of interests for this topic include but not limited to:
• Dynamic spectrum management and sharing with unlicensed and shared bands
• Internet of Things (IoT)
• Waveform design for enhanced security and high mobility
• Small cell mission scenarios
• AI and ML enhanced/incorporated spectrum management, dynamic sensing and sharing
• Smart base/smart port use cases with small cell, V2X, low power and localization
technologies
• Advanced physical layer techniques such as carrier aggregation, full-duplex and massive
MIMO
• Beamforming and adaptive nulling for interference tolerance and spectrum sharing/co-
existence
• Millimeter-wave and terahertz band communications
• The application of unmanned aerial vehicles (UAVs) and Non-Terrestrial Networks
(NTN) in the next generation of wireless networks and beyond
• Leveraging the mobility and 3D positioning of UAVs to enable inherent communications
at the physical layer, i.e., Physical Layer Security (PLS)
• Channel modeling of air-to-air and air-to-ground channels

5G Enhanced Communication Capabilities

Griffiss

Machine learning models, especially deep neural networks, have been shown to be vulnerable to multiple attacks, e.g., dataset poisoning, backdoor attacks, and adversarial examples, that apply near-imperceptible perturbations to training or test data but yield undesirable outcomes in the machine learning models. The majority of research in this area focuses on benchmark image classification datasets. In natural images, the cost associated with each perturbation is generally calculated by taking some measure of distance between the original and perturbed samples. In this context, an attack is successful if the attack fools a target model and its cost does not exceed a specified threshold. This project is interested in exploring adversarial machine learning techniques in domains where the cost associated with “perturbations” is more restrictive. Interns on this topic will help explore the effectiveness of the existing attack literature on problem domains with additional constraints. Results will be used to explore the generalizability of current adversarial machine learning methods. In the case of adversarial examples, perturbations are typically generated for a specific data sample, however recent work shows that some perturbations may be universal across images and architectures. A goal of this effort is to reconcile this universal approach with the standard individual approach as a method of increasing the generalizability and lowering the cost of generating adversarial attacks against machine learning models.
This effort will proceed by implementing various existing perturbation and adversarial attack methods, performing numerical experiments by applying the perturbation techniques to new datasets in non-imagery domains, and exploring the mathematics underlying generalizability of the perturbation and adversarial attack methods.

Generalizing Adversarial Machine Learning to Non-Imagery Domains

The AFRL/RI Resilient & Secure Computing on Untrusted Clouds (RESCU Clouds) program keeps close collaboration ties with highly skilled students, professors, engineers, and researchers from academia and industry to conduct different basic and applied research projects. The peculiar diversity of talents working with and cooperating under RESCU Clouds investigate, design, and implement novel methodologies to securely and efficiently outsource data and distribute computations across heterogeneous hostile computing environments and untrusted Cloud Service Providers (CSPs). The research topics of interest focus on zero trust security and include, but are not limited to, (1) decentralized identity and access control mechanisms and protocols, including those that support anonymity. (2) Novel application of existing cryptographic primitives and protocols to zero-trust computing paradigms. (3) Design cross- cloud, CSP-independent, privacy-aware protocols and frameworks that operate in the presence of emerging zero-trust security mechanisms. (4) Enable secure and transparent migration of application and data across heterogeneous CSPs, and facilitate multi-objective optimization in the security-mission trade space. (5) End-to-end data protection, concurrency and consistency for multi-user multi-cloud environments.

Resilient and Secure Computing on Untrusted Clouds (RESCU Clouds)

The Quantum Information Sciences branch (AFRL/RITQ) performs cutting-edge experimental and theoretical research in a wide range of topics at the frontiers of quantum computing and quantum networking. With a unique interdisciplinary team composed of physicists, engineers, mathematicians, technicians, and computer scientists, RITQ has numerous active research efforts in quantum algorithms and leading quantum technologies such as trapped-ion systems, integrated quantum photonics, and superconducting quantum devices. Below are the application spaces and
skillsets desired for four different projects available to choose from broken out by active research efforts.
The Integrated Quantum Photonics team centers on both the theoretical and experimental aspects of photon-based qubits. The team develops the technology to generate, manipulate, measure, and quantify entangled photons. Typical summer projects include the production and quantification of entangled photons, programming to automate test equipment, integrated photonic device modeling, and theoretical models for quantum transduction.

Quantum Information Sciences: Integrated Quantum Photonics

The Quantum Information Sciences branch (AFRL/RITQ) performs cutting-edge experimental and theoretical research in a wide range of topics at the frontiers of quantum computing and quantum networking. With a unique interdisciplinary team composed of physicists, engineers, mathematicians, technicians, and computer scientists, RITQ has numerous active research efforts in quantum algorithms and leading quantum technologies such as trapped-ion systems, integrated quantum photonics, and superconducting quantum devices. Below are the application spaces and
skillsets desired for four different projects available to choose from broken out by active research efforts.
The Quantum Algorithms team tackles understanding, characterization and exploration of commercially available quantum hardware systems. Application spaces of algorithm development include graph theory optimization and quantum machine learning.
Typical summer projects may include programming on a commercially available quantum hardware platform, quantum simulations, software analysis, and algorithmic development.

Quantum Information Sciences: Quantum Algorithms

The AFRL/RI C-sUAS team is comprised of a unique combination of program managers, cybersecurity specialists, hardware/software engineers, UAS flight testing operators, and UAS Traffic Management (UTM) subject matter experts. This combination of individuals allows them to tackle a wide spectrum of topics including:
• basic and applied research such as Commercial Off-the-shelf (COTS)
• small Unmanned Aircraft System (sUAS) Command and Control (C2) link exploitation
• sUAS detection/tracking/identification and defeat capability development
• UTM capability development for ensuring the safe operation of UAS within the National Airspace System (NAS)
and the application of Artificial Intelligence (AI) and Human-Machine Teaming (HMT) to speed up decision making to assist C-sUAS and UTM Operators in executing their missions.

Typical summer topics may include C2 protocol analysis, C-sUAS capability development, AI/ML/HMT programming and algorithm development, and UTM capability development. Systems include the Ninja and Paladin C-sUAS systems, and the Collaborative Low-altitude UAS Integration Effort (CLUE) UAS Service Supplier (USS) and other UTM capabilities under development.

Small Unmanned Aircraft Systems (C-sUAS)

The AFRL/RI SIGINT (Signal Intelligence) Group would like to transform the traditional methods for encoding, modulating, transmitting, receiving, exploiting, demodulating and decoding the information carried by man-made signals. We are looking for smart, low complexity, efficient, real-time, reliable and multifunctional transceivers and techniques that take advantage of novel signal processing and machine learning methods. We are also interested in extending methods beyond their conventional usages. Typical summer projects include classification, detection, estimation, classification, coding, and other areas of telecommunications applied to hostile and complex environments. The projects are designed to promote brainstorming and allow the student to implement guided solutions or propose and develop their own depending on technical merits.

Software Defined Radio and Machine Learning for Exploitation

The AFRL/RI RIGC Branch is comprised of a mathematicians, physicists, DSP engineers, software engineers, and intelligence operators. This combination of individuals allows them to tackle a wide spectrum of topics from basic research to the challenging aspects of real-time implementation of the results of research efforts. Typical summer topics may include software analysis, database development, programming, machine learning in IOT type systems and sensor data signal detection and processing. The branch interest range over topics linked to increasingly sophisticated techniques in signal detection, characterization, tracking, and classification, all with the goal of signatures via array processing. State of the art techniques such as deep neural networks and other machine learning algorithms are used to pursue solutions for increasing signal exploitation to enable warfighters with significant situational awareness and
knowledge. Key to successful development and use of ML techniques is robust knowledge of ML and insights into the inner workings of these systems.

Machine Learning Interpretability and Explainability

The AFRL/RIGC ELINT programs include onsite/in-house signal generation, capture, and processing capabilities. Anticipated internship activities include signal processing and data science applications. Specific tasking/topics involves radar waveform definition, generation, and collection, as well as feature extraction, clustering, and tagging.

ELINT Signal Processing & Data Science

The AFRL/RI Audio & Acoustic Group is comprised of a unique combination of engineers, mathematicians, and software engineers. This combination of individuals allows them to tackle a wide spectrum of topics from basic research such as channel estimation, acoustic detection and tracking, acoustic identification (ID), language ID and dialect ID. To the challenging aspects of real-time implementation, GUI design, and concepts of operations. Typical summer topics may include software analysis, database development, programming, and signal detection and identification. State of the art techniques such as deep neural networks and other machine learning algorithms are used to pursue solutions for real-time and offline problems.

Audio Processing & Signal Analysis

AFRL/RI is in the process of setting up an Internet of Things (IoT) Living Lab (IoTLL), which will be located in the Innovare Advancement Center (IAC). The IoTLL will be a device and infrastructure test ecosystem that will provide security, interoperability, resilience, and data governance opportunities for AFRL/RI and other entities with IoT interest. Typical summertopics may include software and data analysis, integration of multiple sensors, programming,testing of smart devices within Griffiss Park. Topics may also include machine learning andartificial intelligence research.

Internet of Things Living Lab

The AFRL/RI Battle Damage Assessment (BDA) and Critical Infrastructure (CI) Group is focused on modeling, simulation and analysis of resilient Power Grids Networks and developing algorithms for Battle Damage Assessment that could be associated with CI. Summer research topics include but are not limited to:

- Modeling resilient power grids, critical nodes analysis, collateral damage analysis, and optimal mitigation analysis.

- Image recognition, damage quantification, using incomplete, sparse data.

Battle Damage Assessment Algorithms

A trained classification model can be used to automatically sort items of interest, such as images or signals, into discrete categories or classes. Modern machine learning algorithms typically require a vast amount of labeled data with which to learn such a classification scheme. Unfortunately, this quantity of data is sometimes difficult to acquire. We are interested in techniques that help users efficiently label datasets to reduce the amount of labeled data required to achieve a useful classification model. Your work will focus on implementing state of the art techniques for data efficient machine learning, running simulation experiments, and helping develop the next generation of efficient labeling procedures. Other projects will be introduced over the summer with opportunities for additional involvement.

Data Efficient Machine Learning

The AFRL/RI Text Analysis group is seeking to leverage Publicly Available Information (PAI). PAI includes a multitude of digital unclassified sources such as news media, social media, blogs, and other sources. One task is to identify resources related to trust, for example the Media Bias Fact Check (MBFC), and use these resources for categorizing news. Second task is to explore how news stories with a clear perspective (example pro/anti science, pro/anti COVID vaccine and so on) are propagating online i.e. which users are reposting it, liking it (allows to form communities of users that are aligned to a particular viewpoint). Third task is to utilize topic modeling and network characteristics to differentiate communities aligned to a particular viewpoint (need to identify keywords as well as network characteristics of trustworthy individuals). Final task is to utilize the crowdsourcing power of these communities for labelling new articles (for example the articles that are liked by scholarly/scientific community and not by anti-science community may have a higher probability of being authentic).

Leveraging Publicly Available Information

The AFRL/RI Conversational AI (CAI) Group is comprised of a combination of naturallanguage processing engineers, data scientists, machine learning experts, and developers. Theteam jointly focuses on developing and delivering the next generation on CAI techniques to AirForce Operations. As a team, we focus on researching novel methods of building task-oriented
CAI systems that can communicate with complex information sources, analytics, and microservices to provide users comprehensive and details answers. Traditionally, summer assignmentsvary from working on machine learning problems within CAI to developing micro services usingone-to-multiple information sources. Some of the high-level research areas we are interested inare dialogue state tracking, natural language understanding, few shot entity linking, unstructuredlanguage representation learning, and natural language generation.

Conversational AI

The AFRL/RI Phantom Arrow group is comprised of computer scientists, and active duty members. This group is an all-volunteer group dedicated to enhancing RI’s ability to conceptualize, build, integrate and deliver realistic and useful applications of technology to warfighters through wargaming. To meet this goal, the Phantom Arrow group is working on a digital wargame in the Unity Game Engine. The current effort is to digitize the commercially available wargame Persian Incursion, with the goal of modularizing its well-developed ruleset to elevate the wargame to a near-peer wargame in future iterations. Typical summer efforts have been primarily development focused and building upon the work previously completed in an agile software development style.

Phantom Arrow Wargaming

The Gridded Retarding Ion Drift Sensor (GRIDS) instrument is a CubeSat compatible sensor designed to characterize ion density, drift, and temperature in the ionosphere. This project will focus on the FPGA firmware development aimed at extending the capability of the sensor. Previous iterations of this instrument are will be flown on two NASA CubeSats, petitSat and Dione. The firmware developed for this project will be used in future flight units. Knowledge of HDL is required, but expertise and fluency is not.

VHDL Firmware Design for Ion Drift Sensor 

The University Nanosatellite Program (UNP) was founded in 1999 to foster satellite development expertise in young engineers and to generate new small satellite technologies on small satellite platforms. UNP works with various universities across the country to guide students in the end-to-end design and fabrication of small satellites. UNP typically selects 1 to 5 students to participate in the AFRL Scholars program to assist in this objective. Selected candidates will be required to assist in the analysis, test or documentation for the current nanosatellites preparing for launch. Often times this involves hands-on activities with the satellite. The candidates will also participate in program activities including spacecraft testing, troubleshooting, requirements verification, and providing technical expertise to other university participants within the program. Scholars may have the opportunity to accompany the UNP Office to official small satellite events during the summer. Scholars will have the opportunity to gain and swap systems engineering experience and expertise with other students and professionals in various small satellite lab programs across the US.

University Nanosatellite Program 

Current imaging, wavefront-sensing, and adaptive-optics solutions are inadequate when in the presence of distributed-volume atmospheric aberrations and extended objects (aka the "deep-turbulence problem"). This shortcoming requires that we innovate towards a new solution. In leveraging a recent journal-article publication [https://doi.org/10.1364/JOSAA.36.000A20], this research opportunity will develop imaging, wavefront-sensing, and adaptive-optics approaches that sense and correct for the disturbances found all along the laser-propagation path. Overall, this volumetric solution will enable advanced remote-sensing and directed-energy functions at extended standoffs.


Unconventional Imaging, Wavefront Sensing, and Adaptive Optics 

The ultrashort pulse laser (USPL) group is seeking young scientists and engineers to join the USPL team. Topics under investigation include laser wakefield acceleration, filamentation, and laser-solid or laser-gas interactions. Students will assist with the design and implementation of experimental hardware, build diagnostics, and perform data acquisition and analysis.

Ultrashort Pulse Laser Research 

Modeling the global solar magnetic field is critical for forecasting space weather events. Individuals interested in working with the input data used to drive solar wind models are encouraged to apply for this opportunity. The project involves working with a variety of ground and space based solar disk observations, as well as in-situ data from multiple spacecraft.

Space Weather Modeling 

A spacecraft in geosynchronous Earth orbit (GEO) is subject to the harsh weather of space. Weather in GEO comprises primarily fluxes of high energy electrons, protons, and ultraviolet light. The Spacecraft Charging and Instrument Calibration Laboratory (SCICL) on Kirtland AFB (Albuquerque) is building the capability to accurately simulate the harsh GEO environment in a laboratory-based vacuum chamber. Toward that end, the candidate will construct an electron flood gun based on an in-house design and characterize the performance of a proton gun in order to simulate different GEO weather conditions such as a coronal mass ejection. Some of the activities associated with the project will be working with our in-house machine shop to create the components from which to build the electron source, instrument construction, programming of instrument control software using LabVIEW software, and initial investigations into the effects of GEO weather on some common spacecraft materials.


Space Environmental Simulation Chamber 

Thermal modeling and analysis of spacecraft is hard. Thermal models are highly optimized for a specific satellite design and generate large amounts of data. As a result, analyzing thermal model data is time-consuming and complicated. These complications can be mitigated if a software tool is created to interface with industry standard programs and condense large amounts of data into easily communicable results. The student will be responsible for the following: designing/writing software tools to analyze thermal modeling data; determining methods to interface with industry standard thermal modeling software; using data analysis techniques to write data analysis algorithms; design/implement effective data presentation tools.

Spacecraft Thermal Analysis Software 

This project aims to develop new software tools for improving current and future mission planning activities. The goal is to reduce the time required to produce mission planning products while also improving their quality. Software developed has the potential to be used in current spacecraft operations. Tasks will involve developing or improving programs using a variety of languages to schedule activities, perform coordinate transformations, find fuel efficient trajectories, process and display data, perform optimization, and other. As part of these tasks, you can expect to learn about spacecraft mission planning, spacecraft operations, orbital mechanics, basic astronomy, and space situational awareness. Depending on the task, platforms will include Linux and Windows. Languages may include Perl, Python, Java, C, C#, Matlab, or application specific scripting languages depending on what system the software is integrating with.

Spacecraft Mission Planning Software Development 

Modeling the global solar magnetic field is critical for forecasting space weather events. Individuals interested in working with the input data used to drive solar wind models are encourged to apply for this oppertunity. The project involves working with a variety of ground and space based solar disk observations, as well as in-situ data from multiple spacecraft.

Solar Driven Space Weather Models 

The study of relative motion between multiple resident space objects allows for the modeling and simulation of their interactions. Recent pushes toward spacecraft operating in this regime, such as the Gateway project developed by NASA, will require a new suite of relative motion models to be developed and tested for the new operating environment. Scholars applying to this topic will assist in advancing the capabilities of the Air Force in this new regime, developing and testing new relative motion models.

Relative Motion Models under the Effects Three Body-Dynamics 

The AF and DOD as a whole would like to increase the operating temperature of VIS-LWIR detectors. The goal of this topic is to better understand the fundamental mechanisms that drive the overall dark current in III-V based barrier architecture detectors such as, but not limited to, nBn, pBp, complementary barriers, etc. By developing an in-depth understanding of the mechanisms that limit performance it is anticipated that improved barrier architecture detector designs will result in detectors that can be operated at higher temperatures resulting in an eliminated or reduced need for cryocoolers. Summer research thrusts can be in several areas including: dark current, QE, and noise measurements. These detectors will ultimately be utilized in a space environment so there is keen interest in understanding their performance when subjected to ionizing radiation. Deliverables from this research will ultimately dictate the path forward in future detector growth-characterization campaigns. OPSEC Approval AFRL-2022-3887


Radiometric and Radiation Characterization of III-V Barrier Architecture IR Detectors 

While the majority of radio based studies of the ionosphere use active transmissions, the goal of this project is to explore passive techniques that have only recently been made available, thanks to large, low frequency radio telescopes such as the Long Wavelength Array (LWA) in New Mexico. With unparalleled frequency coverage, field of view, duty cycle, angular and temporal resolution, the LWA radio telescopes are the world’s foremost instruments for passively studying the ionosphere. Students can pick from multiple topics related to this project which include:

1. Imaging riometry for studying atmospheric gravity/acoustic waves in the ionospheric D layer
2. Geolocating, tracking, and measuring electron density of sporadic E structures using unintentional radio noise from the electrical grid as a wide spread radar source
3. Studying the connection between Meteor Radio Afterglows and persistent trains

Radio Astronomy Techniques for Studying the Earth/Space Environment 

Big solar eruptive events often result in the production of highly energetic particles that propagate from the Sun out to the Earth and beyond. In some cases these solar energetic particles (SEPs) reach GeV energies, cause serious communications and spacecraft problems, and can be detected with ground based detectors. The general understanding is that SEPs are accelerated in shock waves in the solar corona that are driven by mass ejections in the eruptive events. The SEPs propagate along magnetic field lines away from the Sun and spread throughout interplanetary space. Our work attempts to understand the origin of the SEPs - what seed particles are accelerated, and when and where in the eruptive events? What eruptions or solar conditions are favorable or unfavorable for SEP production? We use data sets of spacecraft in situ observations of SEPs and space and ground-based optical, radio, EUV, and X-ray observations of solar eruptions to characterize and understand SEP production. Because of the great diversity of SEP events, the work is often statistical in nature, with analysis of properties of many observed SEP events.


Propagation of solar energetic particles in space 

Controlled magnetic fields are critical to the operation of confined atom interferometers, and associated inertial sensors. This project involves further developing in-house capabilities for laser etching so-called 'atom chips'. We use inexpensive DBC substrates: aluminum nitride with a layer of copper bonded to each face. By laser cutting specific wire shapes into the copper, we can create well-controlled magnetic fields simply by varying the applied currents. These chips already enable a variety of experiments involving ultracold atoms, and were used to create a Bose-Einstein condensate. Our goal is to progress to more advanced structures and processes, carefully characterize the laser mill cuts, and reduce the need for post-processing due to trace defects.


Precision magnetic traps for atomic physics

Applicants will port and optimize standard unclassified object tracking and detection algorithms on embedded processing hardware including low power CPUs, GPUs or FPGAs, depending on the applicants hardware expertise. They will then perform benchmarking studies of these algorithms to assess their performance (speed), scalability (problem size) and power utilization using unclassified synthetic data. The student will work closely with both algorithm development and hardware benchmarking and analysis teams in the Space Vehicles directorate.


Porting and Benchmarking of Object Tracking and Detection Algorithms 

Our group studies a wide array of plasma chemistry in fast flow tubes. Our goal is to improve chemistry models of plasmas of interest to the Air Force and DoD. Typical examples include the natural ionosphere, high speed combustion, reentry, solar, and trace gas detection. There are four separate apparatuses in the laboratory for studying reactions at extreme temperatures and pressures as well as a variety of processes including ion-molecule, electron-molecule, ion-electron, and ion-ion. New areas of interest include solar fuels, laser ignition, and catalysis for advanced propulsion.

Plasma Chemistry for Space Related Operations 

We study a broad range of plasma chemistry including electron attachment, ion-molecule reactions, dissociative recombination, and mutual neutralization. We specialize in studying difficult species, radicals, and extended temperature ranges. Several fast flow plasma reaction apparatuses are used. The ion-molecule temperature range is 90-1800 K, while for the other plasma processes temperatures up to 1400 K can be studied. Recent successes include measuring the only product distributions for mutual neutralization, electron attachment to fluorocarbon radicals, and the discovery that electrons catalyze mutual neutralization. A recent upgrade has allowed for studies of radicals with electrosprayed ions, which is important for solar fuels research.

The data support a wide variety of AF/DoD applications including the natural ionosphere, hypersonic vehicles, plasmas assisted combustion, high power lasers, conversion of gaseous to liquid fuels, trace gas detection, high energy density materials, and other catalytic processes.

Plasma Chemistry for Space Applications 

The integration of learning and control in planning for autonomous spacecraft has enormous promise, however rigorous assurances, that typically take the form of mathematical guarantees, are often difficult to obtain. The stochasticity that arises from machine learning algorithms is often resistant to traditional control theoretic techniques. This project investigates the integration of data-driven methods into a stochastic optimal control framework. The goal is to capture the stochastic effects of machine learning algorithms in onboard autonomy in manner amenable to traditional approaches for optimization, planning, and control. The proposed work involves investigation of empirical characteristic functions, which can be embedded into a chance constraint framework, in space robotics scenarios.


Planning and control under uncertainty for autonomous spacecraft 

The normal operation of electric propulsion devices results in the generation of atomic and molecular ions of various charge, often with significant velocities. These expelled ions can collide with other atomic and molecular components (e.g. Earth's atmosphere, un-ionized propellant) and impart a significant amount of internal energy into the collision partners. This internal energy can be rotational, vibrational and/or electronic and often leads to emission of photons at specific wavelengths. These specific wavelengths occur because of the unique energy levels of the collision partners. While the thruster environment can be relatively complicated, these collisions can be generated in the laboratory under controlled conditions such that quantitative determination of the populations in the various energy levels can be determined. These populations can lead to specific optical signatures that can be used for thruster diagnostics and safety-of-flight purposes.


Optical Signatures of Ion-Neutral Collisions 

This topic involces conducting research and development in the field of optically pumped lasers. There various thrusts in this area including diode pumped alkali lasers as well as optically pumped rare gas lasers.  Experimental area includes small scale laser demonstrations,and pressure broadening and shift rate explorations. There is also oppurtunity to develop and test advanced diagnostics for enhancing the understanding of these systems.


Optically Pumped Lasers 

One way to detect and identify spacecraft is to gather the sunlight reflected from the spacecraft as it orbits overhead. These data contain a huge amount of information about the position, composition and intent of the spacecraft that is difficult to obtain in any other way. However, these light curves are dependent on the spacecraft surface materials and they are know to change in response to space weather exposure. So to make the most use of these types of data, we must understand how the reflectivity of spacecraft materials change. 

During this project you will interface with the spacecraft observing community and lab based scientist to quantify the degree of optical changes to common spacecraft materials under simulated GEO exposure in the lab. These data will then be applied to observations of spacecraft in modeling software to recreate the complex light curve of real space objects.

Optical dynamics of spacecraft materials under simulated GEO space weather exposure 

The goal of this study is to characterize how the optical properties of commonly used spacecraft materials change due to exposure to energetic particles. In space, energetic particles break bonds within a material such as polyimide (Kapton®), often forming free radicals. These radicals can then reform the original bond, or they can form new bonds, creating a new material. Understanding the chemical mechanisms for bond breaking and forming processes individually will allow for development of a model to predict the material’s chemical composition as a function of environmental exposure and time. We will simultaneously characterize the optical properties and chemistry of the materials, and develop understanding and correlations between material chemistry and optical signatures.


Optical Characterization of Spacecraft Materials Damaged in a Simulated GEO Environment 

The AFRL/RI Trusted Software group is composed of computer scientists and engineers that tackle research challenges in the areas of scalable formal methods; model-based engineering and validation; compositional verification techniques for resilience; and automation for abstraction validation, and synthesis. These research topics support security, resilience, and reliability in

modern software development processes (Agile, DevSecOps, etc). Typical summer topics may include: developing and testing of models, analysis, and software development/programing. State of the art techniques for machine learning and automation are used to pursue solutions for developing and verifying learning-enabled autonomous systems.

Trusted Software

Despite advances that have enabled deep learning classification methods to achieve human-level performance, significant research is still necessary to ensure trusted and robust operation. The presence of vulnerabilities which can be attacked by a variety of methods, i.e., adversarial machine learning, necessitates a multi-faceted approach to algorithms and methods that ensure adversarial robustness.
The AFRL/RI Robust and Secure Machine Learning team focuses on analyzing plausible adversarial attack methods in order to develop and evaluate mitigations and techniques for trusted, robust, and efficient ML. Topics of interest cover both the basic and applied research categories, and include algorithm implementation and evaluation, dataset curation and analysis, and vulnerability analysis. State of the art techniques using deep neural networks and emerging computing architectures for video, image, and radar signal processing are explored for detection and classification applications.

Robust and Secure Machine Learning/Deep Learning

Mathematical optimization deals with developing models of problems and methods for finding the best available solutions. For example, removing noise from an image or extracting moving objects from video can both be cast as optimization problems. Research in this area includes studying theoretical properties of models, developing algorithms for solving them, and characterizing these solutions. Models arising from applications are often inherently nonconvex or nonsmooth, but may also inherit useful structure from the application in question. This research focuses on identifying that useful structure within the problem which can be utilized to provide theoretical analysis and algorithmic convergence guarantees. Examples of this structure include sparsity, generalizations of convexity, and metric regularity. Some areas of interest are sparse optimization, image and signal processing, variational analysis, and mathematical foundations of machine learning.

Application-driven Structure in Optimization

Position Description:
Researches, designs, and develops computer architectures with military grade cyber security. Works as part of a team to develop and implement a custom microprocessor and its support environment. Implements digital logic designs in a Hardware Description Language (VHDL), and simulates the design to demonstrate the desired functional and security features. Synthesizes the VHDL into a gate level netlist and creates the laid-out design suitable for generation of a mask set for wafer fabrication. Uses fundamental knowledge of computer engineering, digital logic design, and integrated circuit layout, to quickly master ever changing commercial Electronic Design Automation (EDA) software such as Cadence Innovus. Documents work in tech memos and reports, and provides briefings to the team and customers. Develops test benches and assembly or C programs for validation of circuit designs. As a member of a multi- disciplinary in-house team you will have the opportunity to work with leading researchers across academia and industry on cutting edge technologies.

Computer Architecture- Hardware Development

Researches, designs, and develops software integrated development environment for computer architectures with military grade cyber security. Works as part of a team to develop and implement microcode, compilers, assemblers and linkers for custom military microprocessors and its support environment. Create firmware in unison with hardware design team in order to develop and test new embedded processor features. Develop low level interface drivers for open communication protocols like RGMII, SPI, I2C, SATA, etc. Documents work in tech memos and reports, and provides briefings to the team and customers. Develops test benches and assembly or C programs for validation of circuit designs. Plans, initiates, and conducts research and development programs to exploit new technology in the area of computer science and systems to support Air Force command, control, communications, and intelligence missions. As a member of a multi-disciplinary in-house team you will have the opportunity to work with leading researchers across academia and industry on cutting edge technologies.

Software Integrated Development Environment for Computer Architectures

The high-profile applications of machine learning(ML)/Artificial Intelligence (AI), while impressive, are not suitable for a) Size, Weight, and Power (SWaP) limited systems and b) systems without access to “the cloud” and high end computer resources via
networking. Neuromorphic computing is one of the most promising approaches for low-power, unconnected ML. This computing can potentially operable down at the sensor level, and implements aspects of biological brains, e.g. trainable networks of neurons and synapses, in non- traditional, highly-parallelizable, reconfigurable hardware. As opposed to typical ML approaches today, our research aims for “the physics of the device” to perform the computations and for the reconfigurable hardware itself to be the ML algorithm. This research effort encompasses mathematical models, hardware characterization, hardware emulation, hybrid VLSI CMOS architecture designs, and algorithm development for neuromorphic computing processors. We are particularly interested in approaches that exploit the characteristic behavior of the physical hardware itself to perform computation, e.g. optics, memristors/ReRAM, metamaterials, nanowires.

Neuromorphic Computing

Hyperdimensional computing (HDC) or Vector Symbolic Architectures (VSA) are potentially the mathematically rigorous engineering design rules sought by the machine learning (ML) community to stitch together disparate artificial neural network (ANN) frameworks. In HDC, information is represented by high dimensional vectors (1,000 – 10,000 elements long), which may be added (superimposed) or multiplied together to create sets or dictionary key-entry pairs, respectively. A similarity metric measures the correlation between any two vectors. These formalisms follow a connectionist approach: linking concepts together such as in sequences, graphs, and trees. This research topic considers the “edge computing” potential of these methods, e.g. sensors fusion for robot navigation, methods for computing with sparse hyperdimensional vectors, implementing a resonator network in hardware, and memristor crossbar implementations. Additional interest is in exploring the application of sheaves from topological geometry with respect to hyperdimensional vectors derived from sensor data.

Hyperdimensional Computing (HDC)

The Quantum Information Sciences branch (AFRL/RITQ) performs cutting-edge experimental and theoretical research in a wide range of topics at the frontiers of quantum computing and quantum networking. With a unique interdisciplinary team composed of physicists, engineers, mathematicians, technicians, and computer scientists, RITQ has numerous active research efforts in quantum algorithms and leading quantum technologies such as trapped-ion systems, integrated quantum photonics, and superconducting quantum devices. Below are the application spaces and skillsets desired for four different projects available to choose from broken out by active research efforts.
The superconducting team’s research focuses on the investigation of new quantum devices, new qubit control and measurement techniques, and the exploration of fundamental physics relevant to quantum networking architectures, with an emphasis on hybrid superconducting systems – i.e. interfaces between superconducting quantum circuits and other leading quantum modalities, such as trapped-ions and quantum photonic circuitry. Typical summer projects may involve one or more of the following: numerical simulations; programming for data acquisition and data analysis applications; design and assembly of laboratory hardware, including mechanical, electronic, and cryogenic components; and participation in quantum measurements.

Quantum Information Sciences: Superconducting Quantum Devices

Machine learning is a powerful technique that can enable many next-generation applications for the DoD. This is why I and our team have great interest in this topic. I work in the Space Electronics Technology (SET) program in the Space Vehicles directorate of AFRL, specifically the Spacecraft Performance Architectures and Computing Environment Research (SPACER) laboratory. We work with various satellite processing architectures and based around the algorithm being looked at, complete trade-studies to produce the best recommendations and/or predictions for newer technologies. The idea of on-orbit processing could expand the capabilities greatly of next-generation space systems, but there are strict power limitations and the harsh effects that space radiation has on electronics. That is why neuromorphic and bio-inspired hardware is of interest to our group, as they complete highly parallelized processing at a fraction of the cost of traditional modern processors in specific cases, and have shown a good resiliency to bit flips during processing. We will explore many aspects of these systems and how well they can perform, which could broaden out into spiking neural networks or possibly lifelong learning training methods. Another possibility could involve improving and experimenting with neural network models for space related imagery and optimizing them for edge processing on-orbit.

On-Orbit Neuromorphic Processing for Spacecraft 

Due to harsh and inaccessible operating environments, embedded systems for spacecraft are subject to many unique challenges and constraints that limit on-orbit computing performance. However, the increasing need for real-time sensor and autonomous processing, coupled with limited communication bandwidth with ground stations, is increasing on-orbit computing demands for next-generation space missions. To address these challenges, AFRL’s Space Electronics Technology program has established a dedicated architecture analytics project called the Spacecraft Processing Architectures and Computing Environment Research (SPACER) Laboratory to provide the DoD with the capability to assess on-orbit computing solutions for spacecraft. This topic provides scholars with the opportunity to analyze a variety of the latest commercial-grade and space-grade processors including multi/many-core CPUs, DSPs, GPUs, FPGAs, SoCs, and neuromorphic architectures using a combination of metrics, benchmarks, simulations, and emulations. These architectures will be analyzed in terms of their performance, power efficiency, memory usage, reliability, programmability, or other factors using several compute-intensive space applications. Results will help determine how to best optimize architectures for various applications and which architectures are best suited for which missions, and can be used to guide future DoD investment decisions. Projects will be tailored to the interests and expertise of the scholars to provide a meaningful experience.


On-Orbit Computing for Spacecraft 

This project seeks to develop improved methods for spacecraft guidance, navigation, and control (GNC) through improved understanding of the spacecraft nonlinear dynamics. Spacecraft translational motion is dominated by orbital dynamics, and the control is often constrained by a limited fuel supply. Therefore, translational GNC methods generally must be designed to work with these dynamics instead of fighting them. Spacecraft attitude motion is governed by the particular structure of rotational dynamics, and robust performance of attitude GNC methods depends on careful adherence to this structure. Additionally, spacecraft operations are subject to significant nonlinear control-estimation interactions, an example being the lack of observability of control-free relative motion in close proximity when using angles-only measurements. Research projects may address one or more of these topics.


Nonlinear dynamics in spacecraft guidance, navigation, and control 

The Quantum Information Sciences branch (AFRL/RITQ) performs cutting-edge experimental and theoretical research in a wide range of topics at the frontiers of quantum computing and quantum networking. With a unique interdisciplinary team composed of physicists, engineers, mathematicians, technicians, and computer scientists, RITQ has numerous active research efforts in quantum algorithms and leading quantum technologies such as trapped-ion systems, integrated quantum photonics, and superconducting quantum devices. Below are the application spaces and skillsets desired for four different projects available to choose from broken out by active research efforts.
The trapped ion team focuses on implementing trapped ion systems using barium 133 and/or ytterbium 171. These isotopes have a combination of properties that make them perform particularly well as a qubit. Past summer projects contributed to the effort at a variety of levels and include a wide range of topics. Examples include: the investigation of a new material for optics mounting, the development of electrical and mechanical ion trap demonstrations, the construction of key pieces of control electronics, and the creation of new versions of existing lab tools at new wavelengths.

Quantum Information Systems: Trapped Ion Systems

With upcoming emergent technologies in satellite control and communications, crosslinking communications provide a novel method for transfer of information from multiple heterogeneous platforms across large distances. This domain brings numerous challenges at many levels due to the vast distances required in communications, the complexity of the dynamics and the on-board computational power. This research effort is focused on control design and analysis related to coordination of distributed and decentralized multiagent space systems. Potential scholars are strongly encouraged to contact the mentor for more information, as well as to discuss specific research ideas for the summer.

Multiagent Autonomous Systems 

The Infrastructure Management Branch of AFRL administers the RV and RD directorate's facility’s needs, through planning, programming, design, and construction including providing architectural solutions. This includes the development of graphic studies in the areas of sustainable design (identifying and implementing energy cost strategies to existing facilities), new facility projects, facility remodeling projects, facility condition inspections (ICI program), and other misc architecture and engineering design solutions including developing Energy Sustainability and Assurance for AFRL's campus. Our program for 2020 Scholars will offer those involved in Architecture and Engineering disciplines the opportunity to have hands on experience within their given fields. The Infrastructure Management Branch oversees all PRS facilities and these positions support multiple RV and RD programs.


Misc. Architecture, Engineering, and Construction Projects 

Electrospray thrusters using ionic liquid propellants are next-generation propulsion systems that are well-suited for in-space propulsion for small spacecraft. These systems provide unique high-density, low power propulsion capabilities and are envisioned in such missions as the upcoming gravity wave observatory mission (LISA, NASA). The physics occurring at the emission sites of high performance electrospray systems is still ripe for experimental and theoretical exploration designed to improve our understanding of the operation of the devices. 

In this project, the selected student will either undertake experimental or theoretical approaches to further the understanding of the plume of ions emitted from either (or both) a capillary or externally wetted emitter. Theoretical methods, if the student’s primary interest, might include the use of quantum mechanical and molecular dynamics approaches to the prediction of emitted ion clusters, the lifetime of the clusters, prediction of the plume shape and other observables, modelling in support of experimental data or other related work.

Students focused primarily on experimental investigations will utilize mass spectrometry, current-sensing and mass-sensing tools to examine the plume generated from various different propellants or emitter types at relevant volumetric flow rates. Analysis of the experimental measurements will yield insights into trends among the different propellants and improve understanding of the microphysics occurring post-emission in a laboratory thruster.

Micro-physics of Electrospray Thrusters 

The AF has interest in developing catalytic species for applications such as solar fuels; however, development is hampered by a lack of understanding of the mechanisms by which these processes occur. We aim to decipher these mechanisms through a combination of gas phase experimental studies, quantum chemical calculations, and kinetic modeling. Experiments will focus on either biomimetic species or metal cations, measuring the temperature dependent kinetics for systems of interest using a Selected Ion Flow Tube apparatus equipped with a variety of ion sources (electrospray, electron impact, chemical ionization, solids probe, laser vaporization). The (generally complicated) potential surfaces of these systems will be determined using quantum chemical calculations (e.g. Gaussian), and finally information on the mechanisms will be extracted by kinetic modeling of those surfaces to fit to the experimental results.


Mechanistic studies of catalysis in the gas phase 

The propagation of laser beams through turbulent flows has been an important topic with applications ranging from missile defense to target designation and tracking. The turbulent air disturbances are severe enough to severely distort the light, preventing it from properly focusing. The study of these interactions has been described as "aero-optics". The opportunities for the aero-optic research at AFRL include experimental studies of turbulence such as turbulent boundary layers and shear layers as well as the development of appropriate diagnostic instrumentation, and water-table visualization. Computational (CFD) opportunities also exists to design aero-optic experiments and to develop accurate aero-optic CFD solutions.

Aero-optics research

Machine learning is a bound and a step over traditional image processing for classification accuracy, meaning it should be used in all domains possible, even space. I work in the Space Electronics Technology (SET) branch in the Space Vehicles directorate, specifically the Spacecraft Performance Architectures and Computing Environment Research (SPACER) laboratory. We work with various satellite processing architectures and try to fit specific algorithms to each one. Going back to image processing, their is an idea of processing on-board the satellite while in space, the problem is power limitations and the harsh space environment. Specifically the power requirements and the radiation effects on hardware. that is why bio-inspired hardware is of interest to our group, as they complete highly parallelized processing at a fraction of the cost of modern processors. We will explore many aspects of these systems and how well they can perform, which could broaden out into spiking neural networks or different online training methods. On another route, neural network models could be improved for space related imagery or optimized by being combined with different image processing solutions/algorithms, pre-processing methods, or etc.


Neural Networks for Image Processing on Spacecraft 

Highly motivated students will receive laboratory experience while assisting  engineers and research scientists in advanced materials for aerospace applications.  Potential skills:  MATLAB, Solidworks, lab experience, programming.

Student Research Program

Soche

The Space Communication and PNT (SCPNT) program researches next generation satellite navigation (SatNav). Research areas include: advanced SatNav signals and signal exploitation, spacecraft payloads, SatNav control systems and PNT situational awareness. Technology areas include: digital and RF signal processing, software defined radios, RF signal generation and broadcast, encryption, machine learning and command and control technologies. Research is performed both in a laboratory simulated environment and in the field.
The SCPNT program is developing new SatNav concepts. These concepts need to be further defined, their feasibility analyzed, and initial performance characteristics estimated before the concepts can be advanced. An intern project would involve building a model of the concept in an appropriate environment (possibilities include Matlab, Simulink, software defined radio, etc.), calculating link budgets, estimating the size, weight, and power (SWaP) of a transmitter/receiver, estimating receiver performance in the presence of an appropriate noise model, developing machine learning tools to identify and characterize contested environments, etc.

Advanced Satellite Navigation Concepts

The Griffiss Institute aims to develop the next generation of STEM talent and defense technologies that will strengthen U.S. national security and create economic opportunity for our region, state and nation. We elevate talent, we empower innovation, and we enable experience. 

Formed in 1967, SOCHE is a consortium of colleges and universities in Ohio that promotes educated, employed, and engaged citizens. SOCHE is a leader for higher collaboration, engaging with colleges, universities, and industry to transform the economy through education and employment.

AFRL Science and Technology Fellowship Program

The CRFP is a leading fellowship program that recruits, employs, funds, and mentors student and post-doctoral researchers. Through the CRFP, graduate and undergraduate students and post-doctoral Fellows are matched with government agencies that provide applied research experience in support of the U.S. Department of Defense. With the supervision of mentors and specialists, research Fellows develop research, analytical, and technical writing skills by assisting experts while maintaining classes and pursuing a degree. Research fellowships give students and post-doctoral researchers 1-3 years of experience in applied research settings, making their fellowship as beneficial for their future career as it is for their education. 