Time-Delay System Identification Using Genetic Algorithm – Part One: Precise FOPDT Model Estimation

Due to the unknown dead-time coefficient, the time-delay system identification turns to be a non-convex optimization problem. This paper investigates the identification of a simple time-delay system, named First-Order-Plus-Dead-Time (FOPDT), by using the Genetic Algorithm (GA) technique. The quality and performance of the GA-based identification are compared with those based on extended Least-Mean-Square (LMS) methods, subject to the consideration of different types of time-delay systems, excitation signals, Signal-to-Noise Ratios, and different evaluation criteria. The obtained results exhibit that the GA technique has a very promising capability in handling this type of non-convex system identification problem.