Nonlinear Control Analysis on Nonholonomic Dynamic Systems with Affine Constraints

There would always be some unknown geometric, inertial or any other kinds of parameters in governing differential equations of dynamic systems. These parameters are needed to be numerically specified in order to make these dynamic equations usable for dynamic and control analysis. In this study, two powerful techniques in the field of Artificial Intelligence (AI), namely Artificial Neural Network (ANN) and Adaptive Neuro-Fuzzy Inference System (ANFIS) are utilized to explain how unknown parameters in differential equations of dynamic systems can be identified. The data required for training and testing the ANN and the ANFIS are obtained by solving the direct problem i.e. solving the dynamic equations with different known parameters and input stimulations. The governing ordinary differential equations of the system is numerically solved and the output values in different time steps are obtained. The output values of the system and their derivatives, the time and the inputs are given to the ANN and the ANFIS as their inputs and the unknown parameters in the dynamic equations are estimated as the outputs. Finally, the performances of the ANN and the ANFIS for identifying parameters of the system are compared based on the test data Percent Root Mean Square Error (% RMSE) values.

Download Full-text

Nonlinear Control of Dynamic Systems Using Single Multiplicative Neuron Models

Volume 4: Dynamics, Control and Uncertainty, Parts A and B ◽

10.1115/imece2012-87440 ◽

2012 ◽

Author(s):

Jonathan G. Turner ◽

Biswanath Samanta

Keyword(s):

Nonlinear Control ◽

Dynamic Systems ◽

Optimization Technique ◽

Real Life ◽

Population Based ◽

Motor Speed ◽

Neuron Models ◽

On Line ◽

On Line Learning ◽

Better Than

The paper presents an approach to nonlinear control of dynamic systems using artificial neural networks (ANN). A novel form of ANN, namely, single multiplicative neuron (SMN) model is proposed in place of more traditional multi-layer perceptron (MLP). SMN derives its inspiration from the single neuron computation model in neuroscience. SMN model is trained off-line, to estimate the network weights and biases, using a population based stochastic optimization technique, namely, particle swarm optimization (PSO). Both off-line training and on-line learning of SMN have been considered. The development of the control algorithm is illustrated through the hardware-in-the-loop (HIL) implementation of DC motor speed control in LabVIEW environment. The controller based on SMN performs better than MLP. The simple structure and faster computation of SMN have the potential to make it a preferred candidate for implementation of real-life complex control systems.

Download Full-text