TECH FOCUS AREAS: Cybersecurity; Artificial Intelligence/Machine Learning TECHNOLOGY AREAS: Information Systems OBJECTIVE: The objective of this topic is to apply existing AI/ML solutions to operational Air Force problems. This topic will reach companies that can complete a feasibility study and prototype validated concepts in accelerated Phase II schedules. This topic is specifically aimed at later stage development rather than earlier stage basic science and research. DESCRIPTION: The advent of modern AI/ML algorithms presents an opportunity for the Air Force to modernize various human-intensive and laborious processes that may be amenable to automation and autonomy. Areas of interest under this topic include the following: 1. Real Time Operational AI: Human operators are used across many airfield and flight operations that could potentially be performed or augmented by trusted AI systems. Intelligent systems could assist and reduce air traffic control workload, potentially reducing safe aircraft separation to increase sorties or reduce fuel consumption, while maintaining or improving safety of flight. AI could analyze flight paths, traffic volume/controller work load, range scheduling, etc. to determine the most efficient route/timing from airports to specific airspace, either for dynamic resource allocation or overall system optimization. Artificially intelligent precision approach radars could provide voice guidance for precision approach landings in agile combat operations without dedicated human controllers or modification of aircraft avionics. This emphasis area explores how real-time trusted AI could be applied to operational systems, either to augment and improve human and system performance or by acting autonomously. 2. Putting the Sec in DevSecOps for Machine Learning: PEO Digital seeks to take advantage of breakthroughs in machine learning in commercial technology with non-traditional partners. Under the hood, breakthroughs are enabled by leveraging common open-source machine learning frameworks such as PyTorch & TensorFlow. After training, these frameworks output serialized formats of a trained neural network that are optimized to run on hardware via model serving. However, these serialized formats are not well understood by common static code analysis (SCA) tools used in DevSecOps pipelines. This topic seeks to develop a static code analysis tool for neural network serialized graphs (ONNX, TensorFlow SavedModel format, etc) in order to automatically identify vulnerabilities and ensure that new and re-trained neural networks are secure to mitigate the risk of malicious neural networks within operations. 3. Automate Routine Battle Management Functions: Although modern architecture in AWACS and other platforms automates some of the functions, such as automatic track initiation, Air Battle Managers (ABM) still have to perform many routine functions. The AWACS SPO is interested in potential deploying AI to further automate many of the ABM's functions, allowing for ABMs to concentrate on functions that truly require humans in the loop. 4. Computer Vision Threat Detection: PEO Digital seeks ML/AI based computer vision technology to perform real-time threat detection and classification using full motion video provided by CCTV and electro-optical/infrared (EO/IR) sensors. In addition to threat detection, applications would also be applied to reduce current system false/nuisance alarm rates. Proposed technologies would preferably be device/sensor agnostic and capable of integrating with external C2 systems. 5. AI for Tactical / Cognitive Radios (CR): Optimized rogue RF signal detection, classification and response to detection and jamming signals in space can be significantly enhanced with a trained neural network using machine learning and AI techniques. Improvements are expected to reduce signal processing timelines from minutes and seconds to milliseconds. Applying AI techniques to a deployed compact system in a field environment are expected to yield real-time response and collection of signals intelligence without the need for significant offline processing and reduced power requirements. A demonstration of AI for tactical radios using COTS CR systems is preferred. PHASE I: Phase I should completely document 1) the AI-driven explainability requirements the proposed solution addresses; 2) the approach to model, quantify and analyze the representation, effectiveness, and efficiency of the explainable decision-making solution; and 3) the feasibility of developing or simulating a prototype architecture. PHASE II: Develop, install, integrate and demonstrate a prototype system determined to be the most feasible solution during the Phase I feasibility study. This demonstration should focus specifically on: 1. Evaluating the proposed solution against the objectives and measurable key results as defined in the Phase I feasibility study. 2. Describing in detail how the solution can be scaled to be adopted widely (i.e. how can it be modified for scale). 3. A clear transition path for the proposed solution that takes into account input from all affected stakeholders including but not limited to: end users, engineering, sustainment, contracting, finance, legal, and cyber security. 4. Specific details about how the solution can integrate with other current and potential future solutions. 5. How the solution can be sustainable (i.e. supportability). 6. Clearly identify other specific DoD or governmental customers who want to use the solution. PHASE III DUAL USE APPLICATIONS: The Primary goal of SBIR is Phase III. The contractor will pursue commercialization of the various technologies developed in Phase II for transitioning expanded mission capability to a broad range of potential government and civilian users and alternate mission applications. Direct access with end users and government customers will be provided with opportunities to receive Phase III awards for providing the government additional research & development, or direct procurement of products and services developed in coordination with the program. PROPOSAL PREPARATION AND EVALUATION: Please follow the Air Force-specific Direct to Phase II instructions under the Department of Defense 21.2 SBIR Broad Agency Announcement when preparing proposals. Proposals under this topic will have a maximum value of $1,500,000 SBIR funding and a maximum performance period of 18 months, including 15 months technical performance and three months for reporting. Phase II proposals will be evaluated using a two-step process. After proposal receipt, an initial evaluation will be conducted IAW the criteria DoD 21.2 SBIR BAA, Sections 6.0 and 7.4. Based on the results of that evaluation, Selectable companies will be provided an opportunity to participate in the Air Force Trusted AI Pitch Day, tentatively scheduled for 26-30 July 2021 (possibly virtual). Companies' pitches will be evaluated using the initial proposal evaluation criteria. Selectees will be notified after the event via email. Companies must participate in the pitch event to be considered for award. REFERENCES: 1. Myrbakken, H vard, and Ricardo Colomo-Palacios. "DevSecOps: a multivocal literature review." International Conference on Software Process Improvement and Capability Determination. Springer, Cham, 2017 2. Ventullo, Chris. JADC2 via ABMS. AIR UNIV MAXWELL AFB AL MAXWELL AFB United States, 2020; 3. Reichman, Dani l, Leslie M. Collins, and Jordan M. Malof. "Some good practices for applying convolutional neural networks to buried threat detection in ground penetrating radar." 2017 9th International Workshop on Advanced Ground Penetrating Radar (IWAGPR). IEEE, 2017.
