Autonomous dynamic driving control of wheeled mobile robots

Abstract

We propose a novel control framework to enable nonholonomic wheeled mobile robots (WMRs) to autonomously drive in an environment with the speed fast enough so that the dynamics effect (e.g., Coriolis effect) is not negligible, yet, still less than a certain threshold to prevent slippage at the wheels. For this, instead of the Newtonian vehicle modeling, we adopt Lagrange-D'Alembert formulation, which then allows us to explicitly relate the system's state/control with the constraint force, so that we can predict/detect possibility of a given motion's violating the no-slip condition. We present a scheme to generate a no-slip/collision-free timed-trajectory for the WMRs using this Lagrange-D'Alembert formulation. We also propose a backstepping-based control law, which enables the WMR to track the generated trajectory while respecting its nonholonomic constraints. Experiment, using a modified commercial radio-controlled car, is performed to verify the theory.