ANN based self tuned PID like adaptive controller design for high performance PMSM position control

• Published in: Expert systems with applications Vol. 41; no. 17; pp. 7995 - 8002
• Main Authors: Kumar, Vikas, Gaur, Prerna, Mittal, A.P.
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier Ltd 01.12.2014; Elsevier
• Subjects: Algorithmics. Computability. Computer arithmetics; Algorithms; Applied sciences; Artificial intelligence; Computer science; control theory; systems; Connectionism. Neural networks; Cuckoo search; Exact sciences and technology; Gain; Learning theory; Least squares method; Mixed local recurrent neural network; Neural network; Neural networks; On-line systems; Proportional integral derivative; Self tuning control; Sequential learning and servo drives; Synchronous motors; Theoretical computing; Training; Cuckoo search; Neural network; Sequential learning and servo drives; Mixed local recurrent neural network; Self tuning control; Initialization; Artificial life; Variable conditions; On line; Position control; Adaptability; Operating mode; Control synthesis; Updating; Adaptive control; Dynamic conditions; Optimization; Local network; Synchronous motor; Uncertain system; Unmodeled dynamics; Least squares method; Sequential method; Self control; Robustness; Dc motor; Learning algorithm; Capability index; High performance; Recurrent neural nets; Recursive algorithm; Search algorithm; Variable load; Heuristic method; Differential integral proportional control; Dynamic load
• ISSN: 0957-4174; 1873-6793
• DOI: 10.1016/j.eswa.2014.06.040

•A novel training algorithm for NN based PID controller design is proposed.•NN based PID controller offers better performance as compared to PID controller.•The performance of proposed controller is robust to the parametric variations.•Training algorithm is at least 10 times faster than other training methods of ANN. Proportional-integral-derivative (PID) being the most simple and the widely deployed controller in the industrial drives is not quite amenable to the solution for high performance drives as these drives are subjected to the parametric uncertainty, unmodeled dynamics and variable load conditions during operation. In order to expand the robustness and adaptive capabilities of conventional PID controller, a neural network based PID (NNPID) like controller which is tuned when the controller is operating in an on line mode for high performance permanent magnet synchronous motor (PMSM) position control is proposed in this paper. The NN based PID like controller is composed of a mixed locally recurrent neural network and contains at most three hidden nodes which form a PID like structure. A novel training algorithm for the PID controller gain initialization based upon the minimum norm least square solution is proposed. An on line sequential training algorithm based on recursive least square is then derived to update controller gains in an on line manner. The proposed controller is not only easy to implement but also requires least number of parameters to be tuned prior to the implementation. The performance of the proposed controller is evaluated in the presence of parametric uncertainties and load disturbances also the result outcomes are compared with the conventional PID controller, optimized using Cuckoo search based optimization method.