Lateral control of autonomous vehicles based on fuzzy logic

• A new description method of the local reference path has been proposed.
• Use all path points with multiple look-ahead distances to reduce tracking errors.
• Multiple fuzzy rule bases are designed to fit forward and backward driving.
• System stability condition is proposed to guide designing fuzzy inference engines.
• A new strategy of drawing the given driving speed is introduced.

Autonomous vehicles have attracted considerable attention in the research community and industry. This paper addresses a problem in designing lateral control law and develops a strategy to determine the given speed of autonomous vehicles. An improved method for calculating the lateral offset and heading angle error is proposed to reduce the impact of reference path data noise. Multiple fuzzy inference engines are used to design the steering controller and determine the given driving speed, including the forward and backward directions. The stability condition is given to guide the design of fuzzy inference engines. Satisfactory simulation and experimental results have been obtained from different reference paths.