Bi-level programming based real-time path planning for unmanned aerial vehicles

This paper presents a novel real-time path planning approach for unmanned aerial vehicles (UAVs) based on bi-level programming (BLP), in which the planning problem is described as a leader-follower decision making model. The proposed approach can fulfill an integrated path planning requirement, including several realistic abilities of convergence to target, obstacle avoidance, path length optimization, flight path smoothing and adaptability to the changes of the UAV's kinematic and sensory properties. In the BLP model, opposition obstacles and UAV's interrelated performances are described to construct path searching constraints, and variable planning time intervals are introduced to generate navigable flight paths only when necessary. A discretization solution algorithm embedded with five heuristic optimization strategies is particularly designed to speed up the planning. Moreover, convergence and computational cost, as well as potential extensions, are discussed to expose the efficiency and applicability of the approach. Numerous simulations in stochastic and representative scenarios demonstrate not only the effectiveness of the approach in generating optimized flight paths for UAVs, but also its clear advantages through comparisons with four typical methods, while synthetically considering the crucial abilities.