Neural Network/PID Adaptive Compound Control Based on RBFNN Identification Modeling for an Aerial Inertially Stabilized Platform

To reduce the influences of multisource disturbances on the stability accuracy of an aerial remote sensing inertially stabilized platform (ISP), a neural network/PID (NN/PID) compound control method based on radial basis function neural network (RBFNN) is proposed. First, an accurate identification modeling method based on RBFNN is proposed, which solves the problem of difficulty in accurately describing the characteristics of the ISP system under multisource disturbances. The offline/online compound identification method is designed to ensure the real-time performance in the dynamic adjustment of the model. Then, on the basis of the RBFNN system identification modeling, a NN/PID adaptive compound control method is proposed to realize the adaptive adjustment of system parameters, thereby reducing overshoot and steady-state errors of the ISP, and improving the control performance of the system. Finally, the effectiveness of the method is verified by simulations and experiments. Compared with the PID control method, the stability accuracies of the ISP with this compound control method under the moving base and dynamic car experiments are improved by 55% and 41%. These results demonstrate that the proposed adaptive compound control method can significantly enhance the disturbance suppression ability of ISP and improve the stability accuracy of ISP control system.