Point-to-Point (PTP) motion control systems play an important role in industrial engineering applications such as advanced manufacturing systems, semiconductor manufacturing systems and robot systems. Until now, Proportional-Integral-Derivative (PID) controllers are still the most popular controller used in industrial control systems including PTP motion control systems due to their simplicity and also satisfactory performances. However, since the PID controller is developed based on the linear control theory, the controller gives inconsistent performance for different condition due to system nonlinearities. In order to overcome this problem, Neural-tuned PID control using Model Reference Adaptive Control (MRAC) is proposed. By using Extended Minimal Resource Allocation Algorithm (EMRAN) to train the Radial Basis Funciton (RBF) Network, the PID controller can learn, adapt and change its parameters based on the condition of the controlled-object in real-time. The effectiveness of the proposed method is evaluated experimentally in real time using an experimental rotary positioning system. The experimental results show that the proposed system is better than classical PID controller in terms of not only positioning performance but also robustness to inertia variations.
