Sequential Convex Programming using Hybrid Trust Region Method for Reentry Trajectory Planning

Abstract

The sequential convex programming (SCP) method is adopted to deal with the planning problem of reentry trajectory of a lunar capsule. However, the convergence rate of SCP is very slow for this reentry problem. To improve convergence performance, a hybrid trust region method is proposed. A fixed-radius trust region is first utilized to find the optimal solution coarsely. When the intermediate solution is close to the optimal solution, the trust region is changed to soft type to accelerate convergence rate. The effectiveness of the proposed method has been verified by numerical simulations.