A calculation method for line-of-sight stable attitude of networked optical transceiver based on depth feedforward neural network

Networking is the future development direction of space laser communication. According to the structural characteristics of optical transceiver in laser communication network, a numerical solution model for the line-of-sight stability of multi-directional mirrors based on Picard algorithm is established. The model can be converted into a numerical iterative program, which can be applied to the optical transceiver coarse tracking intelligent control system. Aiming at the numerical solution error caused by multi-axis joint motion coupling, a method of error compensation based on depth feedforward neural network is designed. The computer simulation results show that the Picard algorithm has high accuracy in the process of single-axis motion calculation. In the multi-axis joint movement, the application of depth feedforward neural network improves the accuracy of Picard algorithm.