Study On Trajectory Optimization For Hypersonic Vehicle With Complex Constraints

Motivated by the emerging needs for long-range precise strike and military delivery, a hypersonic vehicle is developed that provides a safe approach to the time critical targets. With the test platform of hypersonic vehicle, this research tends to work out the multi-objective and multi-constraints trajectory optimization problem based on Pseudospectral Method(PM) and classical Multi-Objective Algorithms based on Decomposition.Firstly, mathematical models of the multi-objective and multi-constraints trajectory optimization problem for hypersonic vehicle are established. Based on optimal control theory, the multi-objective optimization problem(MOP) is transformed into multiple objective optimal control problem(MOOCP). With mission requirements under different actual scenario, a modeling method of no-fly zones is proposed by composing simple shapes into a new one. Four typical no-fly zone models are built and tested which shows the veracity of this method. Then, a serial iteration approach based on PM for trajectory optimization strategy is proposed to solve the problem. In order to deal with the complex constraints, a successive approximation strategy is used. The constraints are added to a simple problem one by one to approximate the original hard problem. Solutions obtained during the process are used as initial guess for the more difficult problems. Numerical examples for re-entry trajectory optimization with multiple constraints are simulated to demonstrate and verify the method developed. Finally, a solving frame work for multi-objective problem based on PM is designed. Studying multi-objective trajectory optimization problem, this frame work first aggregates subobjectives and then uses PM to solve the single optimal control problem. After simulating MOOCPs, the results show even Pareto Front and prove the validity and efficiency of it.This paper studies multi-objective and multi-constraints trajectory optimization problem, improves its models and gives a new idea of solving method. The result serves as a salutary reference for estimating performance of the vehicle and total designing at the moment of research and manufacture both in theory and application.