Dynamic extreme learning machine identification for nonlinear system with long time delay

• Published in: 2017 36th Chinese Control Conference (CCC) pp. 2094 - 2098
• Main Authors: Dazi Li, Tianjiao Kang, Jinglin Zhou, Qibing Jin
• Format: Conference Proceeding
• Language: English
• Published: Technical Committee on Control Theory, CAA 01-07-2017
• Subjects: Delay effects; Delays; Extreme learning machine; Heuristic algorithms; Long time delay; Mathematical model; Nonlinear dynamical systems; Particle swarm optimization; pH neutralization; Prediction algorithms; System identification
• ISSN: 2161-2927
• DOI: 10.23919/ChiCC.2017.8027664

In this paper, a new method is presented for identification of single-input single-output (SISO) nonlinear systems with long time delay. The proposed dynamic extreme learning machine (DELM) is different from extreme learning machine in that it adds adaptive delay parameters in the output layer of the network, and the delay parameter is optimized by use of particle swarm algorithm. The dynamic extreme learning machine has two main advantages: First, for any system, it needs only two input nodes. Second, it does not need to know the specific delay time. A pH neutralization process and a nonlinear dynamic system process are used to evaluate the performance of the proposed method. Simulation results demonstrate the effectiveness of the proposed identification algorithm for a class of nonlinear system with long time delay.