Trajectory Planning of UAVs with Fault Tolerance Based on Monte Carlo Sampling

Abstract

Due to the limitation of system positioning accuracy, it is necessary to correct an unmanned aerial vehicle’s (UAV) error by some correct points arranged in advance, in order to successfully accomplish their tasks. Planning the optimal trajectory of UAVs considering error correction with fault tolerance is a great challenge that almost has not been broken out because corrections may fail by these points. To address the issue, this paper proposes two trajectory planning methods based on “the definite rule” and Monte Carlo sampling. By constructing five calculation models of the two methods in detail, designing the algorithms appropriately, and developing computational programs based on MATLAB, the results of trajectory planning of UAVs from a practical problem are obtained. The results achieve the optimization objectives that the total trajectory length is as short as possible and that the correction points UAV passes are as few as possible, and meanwhile, the optimal trajectory satisfies all the constraints, which illustrate the feasibility and reasonability of trajectory planning based on the definite rule and Monte Carlo sampling. The comparable results show that computation time is 104 and 195 times that of this paper when the Dijkstra algorithm with ant colony algorithm and the greedy algorithm with tabu search algorithm are respectively used, which illustrate the high efficiency of the proposed method. This work provides a feasible solution for UAVs’ trajectory planning that considers the error corrections and failure probability on certain correction points and achieves trajectory planning of UAVs with fault tolerance.