Many researchers have studied robot system identification in recent years. Nonlinear system identification, whether, it be forward or inverse model identification, is becoming increasingly important in modern control applications. Nonlinear system identification can improve control performance significantly, especially when the system dynamic behaviors are unknown and exhibit great nonlinearity. Additionally, the network has been applied to control engineering. The concept of network simulates the concentration of a set of antibodies. The network system has the following features: self-organization, memory, recognition, adaptability and the ability of learning. Therefore, the network could be applied to nonlinear system identification and provide various feasible options for system models with robust and adaptive characteristics.
關聯:
Information Technology Journal , 10(3), pp.522-531
