Initial value selection strategy of glide trajectory based on Legendre pseudospectral method

Abstract

In order to improve the flight efficiency of the gliding section of the vehicle and ensure that the mission can be accomplished more accurately. In order to solve the discrete aircraft trajectory optimization problem using sequential quadratic programming (SQP) algorithm, the problem that the iterative solution problem will be scattered due to the poor selection of initial values is easily caused. In this paper, an initial value selection strategy for solving the trajectory optimization problem based on Legendre pseudospectral method is proposed. In this paper, the gliding segment trajectory of X-51A similar aircraft model is used as the validation object, and the problem is solved in advance by introducing genetic algorithm (GA) and particle swarm algorithm (PS) in bionic algorithm, and the corresponding optimization result is used as the initial value of the problem solved by SQP algorithm. The experiments show that both the initial values of the genetic algorithm and the particle swarm algorithm can ensure the completion of the task, reduce the possibility of divergence, and indirectly improve the accuracy of problem solving. Under the condition of convergence of the solution results, the comparison between the initial values of the genetic algorithm and the particle swarm algorithm and the optimization results of the traditional initial values, the absolute error of the parameters of the genetic algorithm initial values and the traditional initial values is smaller and the results are more accurate. It is proved that the genetic algorithm is more suitable for the initial value screening of the vehicle trajectory. The results of the corresponding solutions can satisfy the set initial and termination conditions, and the task is successfully completed.