Fast and Reliable Motion Model for Articulated Wheeled Mobile Robots on Extremely Rough and Rocky Terrains

Abstract

Articulated wheeled mobile robots (AWMRs) have reconfigurable body structures that permit them to traverse rough terrains. Predicting trajectories of these robots on a rough surface is vital to carry out path planning. Several model-based approaches have been proposed in the past to predict these trajectories, but the existing work has been mostly focused on soft soil, or hard terrains with a smoothly changing surface. In the future, AWMRs would likely be sent to challenging scenarios that are rocky. Hence, a novel model for AWMRs has been proposed that extends existing Kineto-static motion models to consider multiple points of contacts for each of the wheels of the robot, which is vital for rocky terrains. Simulations are carried out based upon the COLE VII robot to compare the proposed method with state-of-the-art Kineto-static and differential algebraic equation based constraint dynamics models. The accuracy of these techniques and the proposed method is benchmarked against the popular open dynamics engine. Two-way analysis of variance is performed and it is shown that accounting for wheel width and multiple points of contact does not have a significant impact on the simulation results on smoothly undulating terrain, whereas, it does on a rocky terrain. Moreover, accounting for wheel width and multiple points of contact does not significantly impact the run time of the underlying dynamics algorithm for fast collision checking.