Wang Dejiao, Zhang Tingting

Chongqing Industry & Trade Polytechnic

Abstract：

Addressing the critical need for modeling the driver’s neuromuscular dynamic characteristics in human-machine shared driving systems, this paper introduces a neuromuscular parameter identification method utilizing the Gauss-Newton algorithm. By constructing a single-inertia dynamic model of the driver – steering wheel interaction system and combining electromyographic signals with steering torque data collected from real vehicle tests under six different hand grasp positions, the method achieves precise identification of the inertia, damping, and stiffness parameters of the driver’s neuromuscular system. Experimental results show a phasor error of 0.09 and a variance ratio of 86%, with the dynamic characteristics of the neuromuscular feedback loop closely matching the measured data. This study provides a reliable parameter benchmark for the design of haptic takeover control systems in autonomous vehicles and lays a theoretical foundation for predicting the dynamic behavior of human-machine cooperative steering.

Key Words：

autonomous driving; human-machine shared driving system; neuromuscular; parameter identification