Robust Control of Skid Steered Robotic Vehicles on High Friction Surfaces

Abstract

The autonomous control of unmanned ground vehicles (UGVs) is a growing research area. Skid steered UGVs are desired because of their simple control inputs, however the control algorithm requires complex dynamic analysis. The dynamic model is required to properly implement the control algorithm and this paper presents a linearized model for use in optimal and robust linear control methods. For localization of the robot sensors are required and for many applications low cost sensors are desired. This study used low cost sensors which require proper handling because noise is often increased in lower cost sensors. This study investigated the use of Kalman filtering and fusion on low cost sensors along with a novel approach of satellite selection for improved GPS precision. The sensor information from the Kalman filter was then used in a robust control algorithm and the vehicle’s path tracking ability was tested.