Improve Path Following Efficiency for 4WD4WS Robot by Using Exponential Functions in Virtual Target Guidance Algorithm

Most of the path following research for 4WD4WS mobile robots is conducted with two control loops, the kinematic loop and the dynamic loop.This article proposes a method using exponential functions in virtual target guidance algorithm to decrease the convergence time in the kinematic control loop and improve robustness over disturbances.The process of synthesizing the control law is strictly mathematically guaranteed.Simulations in Matlab visually represent the research results. IndexTerms-mobile robot, path following, four-wheel-drive/four-wheel-steer, virtual target guidance algorithm, kinematic control. I. INTRODUCTIONThe four-wheel-independent-drive and four-wheel-independent-steer robot or 4WD4WS is a redundant drive system with high flexibility and mobility with 8 motors (4 drive motors and 4 steering motors) is being used widely.In addition to its high road holding ability and performance, 4WD4WS is also used in harsh conditions such as military tasks.Path following is a common problem in mobile robot control and many re-searches have been done for many years.Most of the cases, authors built two control loops: kinematic and dynamic.With 4WD4WS, some authors built the controller on a dynamic model with unrealistic assumptions about the system's response [6], when the robot, in fact, is a non-holonomic system despite the residual drive.Research in [7] is one of the very detailed studies with the introduction of two separate control loops, in which the steering loop also provides independent forward and rear steering angles.However, by imposing the same steering angle on the front and rear wheels [10, 11], the author has created a side slip phenomenon at the drive wheels in all operating cases.Geometric steering methods are also widely used [1-5, 12] for path following problems.One of those is virtual target guidance method [1,2, 12] which was applied to the 4WD4WS robot, however, when results in [1, 2] have large