Data- and Model-Driven Adaptive Integral Terminal Sliding Mode Control for PMSM Speed Control

Simultaneously achieving minimal overshoot, fast response, and low oscillations under various operating conditions remains a major challenge in the speed control of permanent magnet synchronous motors (PMSMs). To address this, a data- and model-driven adaptive integral terminal sliding mode control (AITSMC) method is proposed, which uses a structurally adaptive sliding surface to achieve simultaneous optimization. Furthermore, a radial basis function neural network (RBFNN) is employed to adjust the AITSMC parameters online. The RBFNN model is trained offline to map load torque and target speed to optimal AITSMC parameters. Optimal datasets under varying operating conditions are obtained using an online automatic parameter calibration method. System stability is verified using Lyapunov analysis. Experimental results demonstrate the effectiveness and feasibility of the RBFNN-AITSMC approach.