Fast, Robust Real-Time Trajectory Generation for Autonomous and Semi-Autonomous Nonlinear Flight Systems

Recent military campaigns in the former Yugoslavia, Afghanistan, and Iraq have highlighted the importance of unmanned air vehicles (UAVs) to national defense and security. While UAVs have already begun to play an important role in military operations, the full impact of the technology will only be felt when autonomous capabilities of UAVs are fully developed and realized. This proposal addresses a key enabling technology for UAV autonomy: real-time path planning and trajectory generation. Automatic, real-time generation of paths and trajectories will free human operators from the details involved in low-level path planning and dynamic flight, and enable them to focus instead on higher-level mission design and objectives. In Phase~I of this project, real-time path planning and trajectory generation techniques for two-dimensional flight were developed and demonstrated in both software simulation and on a small UAV platform. In Phase~II we propose to refine, extend, and fully demonstrate the techniques developed in Phase~I. As part of this effort we will (1) add three-dimensional path planning and trajectory generation capability, (2) robustify the trajectory generation techniques with respect to wind and other environmental disturbances, and (3) augment the deliberative path planning techniques developed in Phase~I with reactive path planning and trajectory generation techniques, thereby facilitating obstacle avoidance and deconfliction with other UAVs. These extended techniques will be demonstrated by flight tests on small (1 to 6~foot wingspan) UAV. The demonstrations will be performed with sensing provided by a virtual, on-the-ground, sensor environment.