Wayne State University was awarded a $296,254 Project Grant from the National Science Foundation Division of Electrical, Communications and Cyber Systems Engineering program to conduct research on physics-model-based neural networks from June 15, 2023 to May 31, 2026. Under this award, the university will develop a hybrid self-correcting physics-enhanced deep neural network framework called HYPHY-DNN. HYPHY-DNN aims to provide the performance benefits of deep learning models while incorporating analyzable behaviors and verifiable properties from physical models. This is intended to help ensure safety for applications involving cyber-physical systems, such as self-driving vehicles and drones. Specifically, Wayne State researchers will implement innovations in neural network architecture design, including physics augmentations of inputs, physics-guided network editing, and time-frequency representation filtering activations. It is expected this work will advance integration of deep learning and robust control methods to enable safety-critical cyber-physical systems to safely operate with high performance in dynamic environments.
Mod # | Description | Reason For Modification | Federal Obligation (Click to sort descending) | Date (Click to sort ascending) |
|---|---|---|---|---|
| Not listed | $296.3k | 6/8/23 |