Navigation and Control for Darpa Grand Challenge



The GPS and Vehicle Dynamics Lab at Auburn University teamed with two entries in the DARPA Grand Challenge 2005. Auburn provided the vehicle control (steering and throttle) along with the navigation routine for RASCAL. For TerraMax, Auburn provided a vehicle model for navigation and controller development. The GPS and Vehicle Dynamics Lab also intends to field an entry in the DARPA Urban Challenge in 2007.

The navigation algorithm combines the Global Positioning System (GPS) with an Internal Measurement Unit (IMU) and additional sensors to provide a highly accurate position and pose estimate at a high update rate. In particular, the various measurements are combined with a Kalman filter algorithm that accounts for the uncertainties associated with each sensor (such as noise and bias). The state estimates are then used in the control algorithms to keep the vehicle traveling along the correct path at the desired speed. In order to develop high performance controllers, accurate model for the vehicles must be determined. The lab has experience both in using experiments or data in specification documents to develop vehicle models. Both methods yield accurate models for use in the controller and/or position estimator.

Figure 1:


SciAutonics - RASCAL

Figure 2:


Rockwell Collins - Oshkosh - TerraMax