An inverse neural network prediction control method for low-frequency compensation

Low-speed rotation systems are widely used in many fields, such as precision measuring instruments, aerospace equipment, and tracking and navigation systems. However, external disturbances, such as low-frequency vibrations, can cause position errors. In this paper, to address this issue, an inverse neural network prediction control method is proposed for the position control of low-speed rotation systems. First, the dynamic model and transfer function for the system are established. Next, the inverse neural network prediction method, based on extreme learning machine (ELM), is developed. This method continuously updates historical data to predict the position for the next time step. The predicted value is then used to adjust the PID controller, improving its compensation capability. Finally, experiments show that the inverse neural network prediction method, when combined with the PID controller, can significantly improve both the stability and position tracking accuracy of the system under different frequency disturbances.