PSO-based single multiplicative neuron model for time series prediction
Expert Systems with Applications: An International Journal
About stability of mechatronic systems driven by asynchronous motors
WSEAS Transactions on Circuits and Systems
Neural design procedure for an ATTR system based on video imagery usage
WSEAS Transactions on Circuits and Systems
Refined binary particle swarm optimization and application in power system
WSEAS TRANSACTIONS on SYSTEMS
Adaptive control using neural networks and approximate models
IEEE Transactions on Neural Networks
Particle swarm optimization models applied to neural networks using the R language
WSEAS TRANSACTIONS on SYSTEMS
Hi-index | 0.00 |
This paper focuses on the design of robustness controller for asynchronous motor. a new PID neural network-integral (PIDNN-I) synthesis control strategy is proposed for the controller design, in which NARMA-L2, an approximated model of nonlinear auto regressive moving average(NARMA) model, is employed to represent the input-output behavior of the motor and gives out the expected control input. PID neurons network (PIDNN), as a kind of novel neural network model with dynamic characteristics, is adopted in NARMA-L2 to identify the motor. PIDNN integrates the advantages of PID with those of artificial neuron network. However, the conventional back-propagation (BP) algorithm, which easily gets trapped in local minimum and is being adopted in the current model, constrains the identifying ability of PIDNN so as to harm to the completion of the controller design. Particle swarm optimization (PSO) algorithm, a new population-based evolutionary global optimization method, is proposed to replace the BP algorithm to train the neurons model. Cooperative particle swarm optimization (CPSO), an improved version of cooperative random learning particle swarm optimization (CRPSO), is put forward to enhance the performances of the conventional PSO in the design. Due to the existence of the tracking error caused by approximate error between identifying and real system, integral (I) control is introduced into the design, namely adopting PIDNN control in large tracking error scale and PIDNN-I control in small tracking error scale. Compared with conventional PID control strategy, simulation results demonstrate that the CPSO-based PIDNN-I synthesis control strategy has improved the control performances of asynchronous motor in robustness and accuracy efficiently.