Trajectory Tracking by Dgps-odometric Integration for Autonomous Vehicle with Industrial Application

This paper presents an industrial application concerning the real time trajectory tracking of an autonomous heavy vehicle (up to 7000 kg) for material handling in industrial environment. Material handling in indoor environment has been successfully developed with navigation based on LIDAR system. In outdoor environment this solution is not reliable due to fog and rain disturbances. In this work we have develop the following modules: 1) trajectory planning with minimization of centrifugal acceleration, 2) DGPSOdometric integration, 3) automatic calibration procedure. The trajectory planning is based on the minimization of centrifugal acceleration in curved trajectory, obtained by connecting straight line with portion of cubic spiral. DGPS is based on hw Trimble in RTK, in order to real time control of the vehicle trajectory we need the integration of odometric localization based on wheel and steering angle measurement. DGPS information is used to set-up the odometric incremental localization. In this way the localization frequency is 33 Hz that allows the real time control. Automatic calibration is used in order: 1) to identify the transformation matrix between LIDAR ad topocentric DGPS frame of references; 2) to identify the position of the DGPS antenna in the vehicle frame of references. The experimental measurements show an accuracy of ±2 cm comparable with the LIDAR accuracy for indoor applications.