Trajectory optimization design method of hopping on surface of small bodies based on machine learning

The exploration of small bodies has received more and more attention, and the surface exploration of small bodies is the hotspot of current research. Nowadays, the hopping solution for small probes is an important form of surface exploration for small bodies, because the hopping probe is small and simple, and can better adapt to the microgravity field environment on the surface of small bodies. The optimal design of hopping trajectory is the key technology for surface exploration, but there are problems such as complicated dynamic model, long computation time and low efficiency when solving the optimal trajectory with traditional optimization methods. To address this problem, this paper proposes an efficient method for solving the optimal trajectory based on the idea of machine learning. By establishing a training model the mapping between the control variable and the initial value is found. The method avoids the complex modelling and calculation process, so the value of the control variable can be quickly obtain. To ensure the quality of the required data samples, a high-precision convex optimization method is used to obtain the training data, and then the training model is built and simulated. The results show that the average relative error of the proposed method is within 4% and the calculation time is less than 1s, while the convex optimization method requires more than 200s. This shows that the trajectory optimization design method based on machine learning is fast and efficient.