基于定制内点法的多无人机协同轨迹规划

To improve the efficiency of cooperative trajectory planning (CTP) for multiple unmanned aerial vehicles (UAVs), a customized interior-point method is proposed for efficiently solving convex subproblems within the decoupled sequential convex programming (SCP) method. Firstly, slack variables are introduced to construct an equivalent form of the subproblem, and the optimality conditions of the subproblem in this equivalent form are derived. Then, under the frame of the predictor-corrector primal-dual interior-point method, an efficient computation procedure for solving the equations of optimality conditions is developed. And through exploring the matrix characteristics of constraints to reduce the complexity of solving subproblems, a fast generation method for primal-dual search direction is proposed to improve the planning efficiency. Simulation results demonstrate that the customized interior-point method can decrease the computation times of CTP by one order of magnitude to several seconds.