Unmanned Combat Air Vehicles: A Case Study in Multi-agent Hybrid Systems

In addition to the theoretical work in the areas of multi-agent architecture integration, multi-modal control, and hybrid model real-time code generation, the Berkeley Aerobot (BEAR) research team is currently developing unmanned air vehicle (UAV) experimental capabilities as well as a capability for modeling and simulating UAV missions. The BEAR experimental platform consists of modified Yamaha helicopters (with a main rotor diameter of 11 ft. and 60-lb. payload). The Yamaha helicopters, developed for commercial crop spraying in Japanese agriculture, provide a reliable and affordable platform with similar maneuverability and agility to that envisioned for small, tactical UAVs.We successfully modified the Yamaha helicopters for autonomous flight to achieve waypoint navigation.
The research conducted with the Office of Naval Research award further enhanced our UAV capabilities to serve as experimental platforms for new technologies involved in the following research areas:
multi-agent coordination, planning, and distributed decentralized control;
coordinated surveillance using multiple UAVs;
formation flight;
pursuit/evasion games;
autonomous landing on simulated ship deck motion platforms; ground-based target tracking;
hybrid control design and verification;
safe and efficient flight mode switching;
design and reliability assessment of fault tolerant control systems;
sensor fusion algorithms for multiple sensors, including INS, GPS, and active vision; and vision-based navigation.