Optimization Algorithm-Based Artificial Neural Network Control of Nonlinear Systems
For control of various nonlinear systems, many controllers have been developed over a period of time. Most of the plants are either nonlinear or their parameters vary with time. PID controller has limitations as it is unable to handle complex nonlineariti
- PDF / 480,442 Bytes
- 9 Pages / 439.37 x 666.142 pts Page_size
- 38 Downloads / 224 Views
Abstract For control of various nonlinear systems, many controllers have been developed over a period of time. Most of the plants are either nonlinear or their parameters vary with time. PID controller has limitations as it is unable to handle complex nonlinearities as well as parameter variations in plants. Therefore, a suitable controller needs to be designed for these plants. Artificial neural network (ANN) has capability to handle these complex issues. In this paper, teacher learning-based algorithm (TLBO) has been used with ANN which optimizes the controller parameters and adapts the nonlinearities present in the plants. To analyze the performance, TLBO with ANN controller is used on various systems, i.e., inverted pendulum and a plant discussed in [1]. The simulation results show the plant output versus mean square error (MSE). Keywords Artificial neural network (ANN) · Teacher learning-based optimization (TLBO) · Inverted pendulum · Nonlinear plant · Mean square error (MSE)
1 Introduction Artificial neural network (ANN) has characteristics of self-adapting and learning, which are used to control nonlinear systems [2, 3]. ANN is a powerful technique based on learning and neurological functions of brain [1]. Artificial neurons are trained to perform complex calculations based on information [4]. ANN has ability of adaption and can adapt new patterns without prior information [5, 6], which made it superior to conventional methods. Rumelhart, Hinton and Williams [7] proposed a back-propagation algorithm, which is popular supervised learning method. Optimization algorithms are used to find best possible solution [8] which can improve the V. Srivastava (B) · S. Srivastava Department of Instrumentation and Control Engineering, Netaji Subhas Institute of Technology Sector-3, Dwarka, New Delhi, India e-mail: [email protected] S. Srivastava e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2021 D. Gupta et al. (eds.), International Conference on Innovative Computing and Communications, Advances in Intelligent Systems and Computing 1165, https://doi.org/10.1007/978-981-15-5113-0_85
1007
1008
V. Srivastava and S. Srivastava
system performance. Teacher learning-based optimization (TLBO) [9, 10] is powerful algorithm which can deal with system nonlinearities and uncertainty with less computational effort [11]. In this paper, artificial neural network (ANN) controller is implemented with TLBO algorithm on inverted pendulum and nonlinear plant [2]. The simulation results show the effectiveness of ANN controller in terms of plant outputs and mean square error (MSE). Comparative study in terms of time series parameters has also been done for both the plants.
2 Mathematical Equations of Nonlinear Systems To implement artificial neural network (ANN) on well-known benchmark problem, inverted pendulum and nonlinear plant model [2] are taken. The description and governing equations of system are given below:
2.1 Inverted Pendulum Inverted pendulum is a well-known problem in control system [12
Data Loading...
