Trajectory Optimization with Complex Obstacle Avoidance Constraints via Homotopy Network Sequential Convex Programming

Space vehicles’ real-time trajectory optimization is the key to future automatic guidance. Still, current sequential convex programming (SCP) method suffers from a low convergence rate and poor performance when dealing with complex obstacle avoidance constraints (OACs). Given above challenges, this work combines homotopy neural network techniques SCP propose an innovative algorithm. Firstly, was used fit minimum signed distance field at obstacles’ different “growth” states represent OACs. Then, embedded framework, thus smoothly transforming OACs simple complex. Numerical simulations showed that proposed algorithm can efficiently deal under such as “maze”, has high flexible extensibility.