Award Data

This effort will extend the RL-ICE capability for low-shot object detection to operate with zero real data within the MWIR band (zero-shot learning). Our goal is to 1.) mature our software, 2.) maximize the performance benefit of synthetic data in training an object detector and, 3.) assess the trade-space of real vs synthetic data by comparing performance to a detector trained on fully real data. ...

QUINN (Quantum INspired Neural Networks) is a cybersecurity implementation to be deployed alongside a machine learning system (especially a reinforcement learning, active learning, and computer vision system) to harden said system from both standard generative adversarial attacks (noise injection) and cyber-physical attacks. This protection is provided by three unique components that perform attac ...

Physical Sciences Inc. (PSI) will develop a Recycling Fast-Response Atom Interferometer for Navigation (ReFRAIN) as an atom interferometer (AI) accelerometer specially designed to improve inertial navigation systems (INS) on moving platforms.  The ReFRAIN provides sensitivity, dynamic range, bandwidth, and bias stability that matches or exceeds state of the art mechanical accelerometers.  PSI in ...

Execution of long-range weapons capabilities reduces risk and affords greater warfighter protection. Core enabling technologies for hypersonic projectiles include high-strength lightweight materials, precision avionics, and novel designs. System demands include the ability to withstand both high accelerations (up to 50,000 Gs) and aerothermal shock. Detailed structural engineering and shock modeli ...

SI2 proposes to leverage prior Government funded efforts to develop a transmit (TX) only, wideband (6:1 bandwidth), low-cost, scalable active electronically scanned array (AESA) that will utilize phase-only nulling to enable radar, Electronic Warfare / Electronic Attack (EW/EA), Information Operations (IO) and other capabilities on multiple platforms across DOD agencies. The array will employ digi ...

In this SBIR project, EMAG Technologies Inc. proposes to develop a compact, low-cost, scalable, transmit-only X-band active phased array antenna with phase-only nulling capability based on our proven VISAT architecture. The proposed AESA will use commercial PCB manufacturing platform and will utilize commercial off-the-shelf (COTS) parts and components for the entire multilayer stack-up. The propo ...

State of the art object detection in satellite imagery currently requires large quantities of hand-labeled satellite images. But what if there exists only very limited satellite imagery of the object, perhaps a single pass? Current deep learning solutions can not learn effective models with this extremely limited data. If, however, there exists model of the object that can be used to synthesize mo ...

Airborne radars operating over open water must classify maritime vessels by measuring and exploiting highly-variable radar signatures. Sources of signature variability include within-class ship construction and equipment differences, complex in-situ 6-DoF ship motion caused by ocean waves across a range of sea state conditions, acquisition geometry including grazing angle and maritime-specific RF ...

Deep Neural Networks (DNNs) are becoming widely used in the DoD for image classification, but recent research has shown DNNs are vulnerable to adversary attacks. Adversaries can monitor the DNN training and classification processes to learn attributes of the training data and the DNN. With this information, an adversary can gain valuable insight into the potentially sensitive data used to train th ...

Machine learning models are susceptible to adversarial attacks that make modifications to the input data in order to cause misclassifications. The root cause is the linearity of the decision boundaries of machine learning models in relation to their inputs. One promising direction is to represent the input data as a distribution. Quantum information science entails techniques for working with wave ...