Single Neuron PID Control of Agricultural Robot Steering System Based on Online Identification

Abstract

Aiming at the complexity, nonlinearity, time-variability of the agricultural robot steering system and soil variety, a self-adaptive single neuron PID control is designed based on RBF neural network on-line identification. The RBF neural network identifier achieves the online identification of the Jacobian matrix of steering system, which also acquires on-line tuning information of PID control parameters. The self-turning of PID controller parameters is completed by the single neuron to improve the traditional PID control performance, and the intelligent control of the agricultural robot steering system is achieved. Experiment results show that compared with conventional PID control method, the proposed single neuron PID control of online identification-based steering system, which not only makes full use of the advantages of the optimal approximation feature of RBF neuron network, but also has strong adaptability of single neuron, is characterized for its high precision, great adaptability and robustness, making it feasible for agricultural robot steering system.