Near Optimal Obstacle Avoidance Using Exploring Trees and Waypoint Guidance
J. Willard Curtis, J.R. Cloutier · 2009
A tree-growing routine is embedded in a receding horizon optimal control formulation to produce a discrete near optimal, flyable, obstacle-avoiding flight path for attacking a moving ground target. As the near optimal branch is being constructed, the nodes of the branch become three-dimensional waypoints. Unit vectors are determined from central dierences that define the desired flight path through each waypoint. A guidance law is developed that is capable of guiding the vehicle through each waypoint while satisfying the pre-determined flight path constraint associated with each waypoint. After passing the last waypoint, the guidance law is then operated in a proportional navigation-like homing mode to hit a moving ground target.