Hybrid Kalman-MPC Strategy with Detailed Case Study on Speed Control of DC Motors

Effective speed control of direct current (DC) motors is often necessary for today's am industrial applications where many variables such as ambient interference, inaccurate or fluctuating parameters or changing environmental conditions may exist. Traditional control techniques such as proportional-integral-derivative (PID) control and fractional order proportional-integral-derivative (FOPID) allow for ease of implementation and low complexity. However, in dynamic and noisy environments, they may not provide optimal solution. This paper proposes a new hybrid control technique called the Kalman filter (KF) model predictive control (MPC) for speed regulation of dc motors. The Hybrid Kalman Model Predictive Controller (HKMPC) scheme uses Kalman filtering techniques to generate real-time estimated states that remove or reduce the effects of sensor noise, while the MPC will use these corrected estimated states to forecast future system behaviour and generate optimal control inputs subject to practical constraints. The HKMPC scheme is compared with traditional PID and stand-alone MPC controllers based on common performance characteristics in the time domain such as rise time, settling time, overshoot and steady state error. The simulation results, were achieved using MATLAB indicate that while the MPC improved the overall performance of the system when compared to the PID controller. The performance of the MPC degrading in a noisy environment. In contrast, the HKMPC could display improved robustness, improved speed of convergence, and improved disturbance rejection through the benefits associated with the combination of estimation and prediction of future behaviour of a system. The results show that our proposed hybrid approach strikes an optimal balance of speed, accuracy and robustness which makes it ideal for real-world applications in which measurement noise and uncertainties are present. As such, this paper highlights the potential of combining estimation and predictive control paradigms as effective next-generation control techniques for motor control applications.