Controlling the Position of the Carriage in Real-Time Using the RBF Neural Network Based PID Controller

摘要

This paper proposes an algorithm for estimating parameters of a proportional integral derivative (PID) controller in real-time using the radial basis function (RBF) neural network. The proposed algorithm is then applied to precisely control the position of a carriage. An Arduino control board is utilized to connect between the proposed approach using Matlab Simulink and the actual position control model (i.e., carriage model). The control signal of the PID controller is sent to the Arduino control board to control the position of the carriage through a serial protocol. The actual position of the carriage is read by a displacement infrared sensor and sent to the developed algorithm using the same serial transceiver protocol. The experimental results show that the RBF based PID controller gives better performance than the conventional PID by reducing the overshoot. It is noteworthy that the proposed approach could precisely control the position of the carriage with its uncertain mathematical model.