Sliding mode control strategy of 3-UPS/S shipborne stable platform with LSTM neural network prediction

Abstract

In order to isolate the influence of wind and waves on shipborne equipment, the shipborne stable platform has been widely used to compensate ship orientation fluctuation and ensure the normal operation of shipborne equipment. Efficient and accurate control strategy is the basis of its normal work. In this paper, a motion control strategy for 3-UPS/S parallel stable platform is proposed. Firstly, a sliding mode controller (SMC) is designed based on the kinematic model of the mechanism, and a new variable power reaching law is proposed, which takes into account both the initial speed and the stability of the final approach stage. Secondly, long and short-term memory (LSTM) neural network is used to predict ship orientation. On this basis, the LSTM ship orientation prediction module is embedded into the SMC to realize the predictive control of the stable platform. The experimental results show that the control accuracy of the SMC with LSTM prediction is better than 0.07°, and the stability accuracy of the controller is improved by about 25% and 60%, respectively, compared with SMC and proportional integral derivative (PID) controller.