Improved Sequential Convex Programming Based on Pseudospectral Discretization for Entry Trajectory Optimization
Shoudong Ma,Yuxin Yang,Zheyu Tong,Hua Yang,Changju Wu,Weifang Chen
DOI: https://doi.org/10.1016/j.ast.2024.109349
IF: 5.457
2024-01-01
Aerospace Science and Technology
Abstract:Sequential convex programming (SCP) is widely used to solve entry trajectory optimization problems. However, challenges persist in scenarios with strict constraints, such as unsolvable convex subproblems, iterative solution oscillations, and slow convergence rates. This study introduces an enhanced SCP algorithm designed to address these limitations. First, hp Radau pseudospectral discretization is used instead of trapezoidal discretization to improve the efficiency in solving subproblems while maintaining discretization accuracy. Second, the trust region is adaptively updated on the basis of trajectory information during iterations. Additionally, constraint relaxation and virtual control are introduced to facilitate smooth iteration at the initial stage. The regularization technique is also utilized to improve the convergence rates. Finally, the proposed algorithm is validated through two examples: maximum-terminal-velocity entry and maximum-terminal-longitude entry. The results show that these two problems cannot be effectively solved using the basic SCP algorithm. However, the proposed algorithm, along with the trust-region SCP algorithm and GPOPS, can solve them efficiently. With comparable accuracy in the obtained solutions, the algorithm proposed in this paper requires only half the CPU time of the trust-region SCP algorithm and just 5 % of the computing time of the general-purpose open-source solver GPOPS.