UAV path planning in dynamic and uncertain environments

This paper presents a genetic-algorithm-based approach for uninhabited air vehicles (UAVs) path planning in dynamic and uncertain environments. We model the UAV path as a sequence of speed and heading transitions occurring at discrete times, and this model in particular takes some vehicle dynamic constraints into consideration in the generation of trial solutions. The sequence of speed and heading transitions is variable-length, and it is encoded as variable-length chromosomes accordingly. Simulation studies have shown that the proposed algorithm is effective in finding a near-optimal obstacle-free path in dynamically changing and uncertain environments, and it can guarantee that all candidate solutions lie within a feasible and reachable path space.