Self-tuning PID control of a brushless DC motor by adaptive interaction

Abstract

In this paper, a self-tuning algorithm for proportional integral derivative (PID) control based on the adaptive interaction (AI) approach theory efficiently used in artificial neural networks (ANNs) is proposed. In this approach, a system is decomposed into interconnected subsystems, and adaptation occurs in the interaction weights among these subsystems. The principle behind the adaptation algorithm is mathematically equivalent to a gradient descent algorithm. The same adaptation as the well-known backpropagation algorithm (BPA) can be achieved without the need of a feedback network, which would propagate the errors, by applying adaptive interaction. Thereby, the ANN controller can be adapted directly without wasting calculation time in order to increase the frequency response of the controller. The velocity control of a brushless DC motor (BLDCM) under slowly and rapidly changing load conditions is simulated to demonstrate the effectiveness of the algorithm. The AI tuning algorithm was used to tune up the PID gains, and the simulation results with PID adaptation process are presented by comparing the obtained results with the adaptive PID controller based on BPNN and a conventional PID controller.

A PI controller optimized with modified differential evolution algorithm for speed control of BLDC motor

“…A suitable fitness function is the basis of modified DE algorithm applied to control systems. However, it is difficult to meet the requirements that the response, stability, and robustness of the speed control for BLDC motor according to the speed error [36]. To optimize the dynamic and static characteristics of BLDC motor, a fitness function that contains multiple system parameters needs to be established.…”

Design and Simulation of Adaptive PID Controller Based on Fuzzy Q-Learning Algorithm for a BLDC Motor

“…During the use of the PID controller, determining the appropriate PID coefficients is probably challenging in the case of the existence of various nonlinearities and uncertainties, including hysteresis, friction, payload variations, etc. The desired speed tracking and control system performance could be affected by all these factors [2]. This issue can be addressed through applying intelligent control approaches, including fuzzy control, adaptive control, and neural networks.…”

“…Many tuning approaches and discussions were presented in [3]. Moreover, it is possible to discover countless research articles focusing on auto-tuning PID control [4], APID control [2,5], auto-tuning PID-type fuzzy control [6], etc. in the existing literature.…”

“…Approaches based on neural networks (NNs) for motor control system does have a powerful ability to overcome system uncertainty and disturbance, while more computational capacity and data storage space is required. These methods usually do not include transforming the continuous time domain plant into the equivalent of its neural network [2]. Fuzzy logic control is a powerful, intelligent control approach, primarily according to control of rules and the knowledge of professional staff and specialized skills is the key source of the rule.…”

“…However, the choice of the appropriate value of adaptation rate gain can affect the speed response. In [2], an adaptive interaction method theory-based self-tuning algorithm for PID control is presented, which is properly applied in artificial neural networks. This modern technique would not need the continuous time domain plant to transform into its corresponding neural network.…”