High Sensitivity Tracking for Event Based LEO Moving Target Indication

TECH FOCUS AREAS: General Warfighting Requirements (GWR) TECHNOLOGY AREAS: Space Platform OBJECTIVE: This topic seeks to design and develop an event-based sensing platform specifically optimized to the detection of ground moving targets from a small LEO payload. DESCRIPTION: The DoD's interest in a proliferated and hybrid constellation architecture to execute intelligence, surveillance and reconnaissance (ISR) missions requires us to rethink traditional sensing modalities and mature those which scale well with large volumes of data, supporting true autonomous sensing development. Space-based EOIR imagery has reached very high spatial resolutions and sensitivities but requires high format sensors which output large unchanging data volumes not useful for the mission. This limits the amount of imagery collected and stored, therefore inhibiting the ability to collect video frames of particular interest in the moving target indication (MTI) field. These problems will be amplified when moving to a hybrid satellite architecture where SWAP C demands are greater but the requirement to process and relay data on the edge puts greater strains on space-borne systems. Event-based sensors rely on asynchronous pixel response which only report information when changes in scene dynamics occur. The result is a sparse stream of high time resolution data where each event is in the format (t, x, y, p) where t is the time of the event, x and y represent the position of the pixel reporting the change, and p is a polarity term indicating positive or negative going changes. This results in inherently sparse data which maintains high time resolution. Event-based sensors, which were first designed for the machine vision applications are then ideal for space-based ISR missions such as MTI. While current state of the art event cameras is improving and well-suited for machine vision applications, they are not optimized for unique space-based remote sensing challenges. The goal of this research is the design and development of an event-based sensing platform specifically built and optimized to perform ground moving target indication (GMTI) from a small LEO platform ultimately well-suited for integration into a proliferated and hybrid satellite constellation. Successful design will require pixel-level considerations to maximize the trade-off between spatial resolution, field of view, and on pixel photon flux. The platform will also require robust GMTI algorithm development, leveraging the unique event camera dataset to monitor large numbers of targets while looking for anomalous behavior. This will be especially challenging in a constellation architecture, as persistent coverage requires handoff to maintain target tracks for meaningful time periods. PHASE I: Phase I requires a discovery study to inform the critical design parameters specific to the space-based MTI problem applied to event-based sensors. This includes an examination of pixel design for currently available cameras and improvements to optimize mission specific performance. Phase I will result in a recommended sensor design to be digitally engineered in Phase II. PHASE II: The Phase II will culminate in delivery of a full payload design including, optics, sensor, readout circuit and algorithms specific to event data for GMTI. Successful solutions will utilize digital engineering to the extent possible for the design process of a GMTI specific event-based sensing payload. Careful attention shall be paid to desired spatial resolution, and FOV required to accomplish the objectives from LEO. Sensor design should be informed by existing state of the art event-based sensors but specifically tailored to the scene dynamics associated with GMTI. Understanding scene background radiances and relevant contrasts for targets of interest will be key to the pixel design, optics selection, and success of developed algorithms. Payload and algorithm performance characterization will require high fidelity synthetic data use. Sensor design and performance will require all models be validated against physical observables in both the field and laboratory. PHASE III DUAL USE APPLICATIONS: The Phase III company will work with transition partners to identify mission specific use case. Build sensing payload into field and laboratory testable form factor. Use field and laboratory demonstration to evaluate MTI performance capability. Integrate tested payload into a small satellite form factor for flight demonstration. Further develop EBS exploitation algorithms for detection/tracking/counting of low contrast semi-resolved objects and generalize those methods for commercial applications. NOTES: The technology within this topic is restricted under the International Traffic in Arms Regulation (ITAR), 22 CFR Parts 120-130, which controls the export and import of defense-related material and services, including export of sensitive technical data, or the Export Administration Regulation (EAR), 15 CFR Parts 730-774, which controls dual use items. Offerors must disclose any proposed use of foreign nationals (FNs), their country(ies) of origin, the type of visa or work permit possessed, and the proposed tasks intended for accomplishment by the FN(s) in accordance with the Announcement and within the AF Component-specific instructions. Offerors are advised foreign nationals proposed to perform on this topic may be restricted due to the technical data under US Export Control Laws. Please direct questions to the Air Force SBIR/STTR Help Desk: usaf.team@afsbirsttr.us REFERENCES: G. Gallego et al., "Event-based Vision: A Survey," in IEEE Transactions on Pattern Analysis and Machine Intelligence, doi: 10.1109/TPAMI.2020.3008413. F. Barranco, C. Fermuller and E. Ros, "Real-Time Clustering and Multi-Target Tracking Using Event-Based Sensors," 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2018, pp. 5764-5769, doi: 10.1109/IROS.2018.8593380. Afshar S, Ralph N, Xu Y, Tapson J, Schaik Av, Cohen G. Event-Based Feature Extraction Using Adaptive Selection Thresholds. Sensors. 2020; 20(6):1600. https://doi.org/10.3390/s20061600 KEYWORDS: Event Based Sensing; Neuromorphic Vision; Target Tracking