SLIP PREDICTION OF SKID-STEER MOBILE ROBOTS IN MANUFACTURING ENVIRONMENTS

Skid steer mobile robots (SSMR) are platforms with simplistic mechanical drives readily adapted for a variety of applications. Skid Steer robots require slipping when navigating general paths. This slipping behavior is a function of surface conditions (friction) as well as robot motion and forces (system dynamics). Slipping affects motion along the path as well as drive torques and power consumption and is therefore an important consideration in robot design. Slipping can be characterized through the instant centers of rotation of the contact patches of the left and right tracks, and it has been shown that these are functions of the system dynamics. Therefore, prediction of system dynamics is needed to better evaluate robot slipping behavior during run time. In this paper, these instant center locations are called the slip parameters. In particular, the paper looked at two alternative models for the slip parameters, one a kinematic estimate assuming constant slip behavior for a class of tasks and the other a dynamic estimate which predicts slip parameters to lie on continuous, closed curves that vary in size relative to payload conditions. Each of these models for slip parameters have been presented in the previous literature, but this paper experimentally tests the slipping for track-based SSMR's and compares these results to the kinematic and dynamic estimates. This paper will evaluate the validity of each model through real-time tracking and will improve understanding of slip behavior during typical manufacturing tasks. The paper will then present guidelines for the design of SSMR systems based on knowledge of ICR behavior.