Reentry Capsule Reachable Tube Boundary Prediction via Evolutionary Multiobjective Optimization

Abstract

In the field of aerospace, solving the boundary problem associated with the parachute-capsule system remains a big challenge. The conventional Monte Carlo method proves inadequate for acquiring comprehensive boundary information. To address this issue, this paper introduces a novel tube prediction scheme by leveraging the natural geometric characteristics of the reachable tube and employing a multiobjective optimization strategy. Initially, a multibody dynamic model with nine degrees of freedom was established and verified by the airdrop test data to ensure the accuracy and reliability of the model. Subsequently, the Sobol sensitivity analysis method was employed to assess uncertain factors that affect the deceleration phase of the reentry capsule. These factors are then utilized to determine the optimization parameters for the multiobjective optimization model. Ultimately, the multiobjective evolutionary algorithm based on decomposition was employed to solve the multiobjective optimization model, and the geometric boundary of the tube corresponds to the Pareto front of the multiobjective optimization. The proposed methodology was validated through a simulation experiment utilizing the Chang’e-5 reentry capsule as an engineering case. The experimental results unequivocally demonstrate the superior accuracy of our approach in predicting the boundary of the reachable tube compared to the Monte Carlo method. This research serves as a valuable reference for calculating reachable tubes in practical engineering scenarios and can be effectively applied to spacecraft search and rescue operations during the reentry phase.