Ohio State University at the 2004 DARPA Grand Challenge: Developing a Completely Autonomous Vehicle

I n t e l l i g e n t T r a n s p o r t a t i o n S y s t e m s accelerate research and development activities in autonomous ground vehicles, increase collaboration among different organizations and interest groups, and draw widespread attention to the technological issues and benefits of developing completely autonomous off-road ground vehicles. At Ohio State University, we have extensive experience in developing autonomous ground vehicles. Our team developed three vision-and radar-directed cars to participate in the US Automated Highway demonstration in 1997 and a car that followed a series of predefined routes using GPS in 1999. However, an off-road race such as GC 04 provided a series of new and different challenges. To meet these challenges , we formed Team TerraMax in collaboration with Oshkosh Truck Corporation (www.oshkoshtruckcorporation. com), which provided an MTVR (Medium Tactical Vehicle Replacement), an 8 × 28 ft., 32,000 lb. truck manufactured for off-road military applications. Our goal was to create a completely autonomous vehicle that could negotiate most of the GC 04 route and perhaps even finish the race. To achieve our goal, we needed to conquer three main technical challenges. The first was to install drive-by-wire technology in the vehicle and to modify the vehicle to fulfill DARPA safety requirements. In addition, we added equipment to provide electrical power for computers, actuators, and sensors and a second fuel tank to allow the vehicle to run for over 10 hours and over 300 miles. The second challenge was to provide sensing and information fusion algorithms. The autonomous vehicle needed to be able to sense both vehicle states such as position, speed, and direction and environment information such as the existence and relative position of obstacles. The autonomous-driving system would collect the information from different kinds of sensors installed on the vehicle. So, to provide accuracy, robustness, and usability, we needed to filter and fuse the information before the vehicle could use it. The third challenge was control. On the basis of the fused information, the autonomous-driving system would need to make the correct decision, apply the navigation algorithm , and properly control the vehicle through the drive-by-wire capability. Visual road detection, map-based path finding, and real-time communication between computers were also important issues. The hardware installed on TerraMax (see Figure 1) consists of the drive-by-wire electronic actuators, local network , computers, …