A Framework for UAS Collision Avoidance Systems

Dr. M. Ilhan Akbas, Embry-Riddle Aeronautical University

Advances in unmanned aircraft systems (UAS) have led to their presence in various domains, ranging from recreational use to commercial applications. However, ensuring the safety of UAS operations, particularly regarding collision avoidance, remains a paramount concern. Our research, supported by the Boeing Center for Aviation and Aerospace Safety, introduces an open-source framework designed to enhance the validation of UAS collision avoidance systems. The framework will empower developers and researchers to iteratively refine and validate collision avoidance algorithms, ultimately enhancing the safety and reliability of UAS operations.??

The proposed framework employs a two-tiered simulation approach: first, a low-fidelity multi-agent simulator for rapid scenario testing, and then a higher fidelity 3D simulator for in-depth analysis. This framework streamlines the transition between the two types of simulators aiding comprehensive scenario exploration and in-depth analysis. It also allows seamless integration with physical testing environments. ?

The low-fidelity agent-based simulator facilitates the testing of diverse scenarios, enabling wide-ranging coverage of potential collision situations. Through this approach, users can efficiently explore a wide spectrum of scenarios, evaluating the performance of algorithms under varying conditions. This approach also allows for the utilization of AI-validation tools, developed by Embry-Riddle students Quentin Goss and John Thompson, in the UAS domain. Making use of the extensive dataset produced by these simulations, the framework employs selection criteria to identify the most critical scenarios for further evaluation. Subsequently, the selected scenarios undergo rigorous testing within a higher fidelity 3D simulator, offering a more detailed and realistic environment.?

Alejandro Gonzalez, a Ph.D. student in the Embry-Riddle Aeronautical University Electrical Engineering and Computer Science Department, conducts the experiments and manages the integration of simulation layers. He has been using modeling and simulation tools in his research and presented his earlier M.S. thesis work at an Institute of Electrical and Electronics Engineers conference in 2023 (image below). ?