Collision-free Path Planning and Trajectory Generation for MAVs Flying in Urban Terrain

This paper presents an optimal algorithm for finding collision-free path and trajectory for microair vehicles (MAVs) in urban terrain where multiple ground obstacles are distributed. Regarding vertices of polygonal obstacles as input points of the Delaunay triangulations, a Delaunay-based path planning is proposed to determine an optimal waypoint path while considering criterions of minimum fuel expenditure and collision risk. A trajectory generation method with corresponding optimal constraints is further developed to yield feasible fly trajectory along the planned path. The off-line pre-planning and on-line updated planning are integrated to enhance autonomy of vehicles flying in dangerous situations, with tight maneuvering and operational capability. Simulation results of a mission scenario are finally presented to demonstrate effectiveness of the proposed approach