Modified Dual-Observer Structure and Radial Basis Function Neural Networks for Linear Active Disturbance Rejection Control

A method combining external structures with radial basis function neural networks (RBFNN) is proposed to improve the performance of linear active disturbance rejection controller (LADRC). To complement this method, the layout of the LADRC state observer is optimized into a dual-observer structure. Through trajectory control simulation experiments on a single-degree-of-freedom manipulator, it was verified that the neural network controller with the new structure is more effective than LADRC in terms of performance.