An unmanned vehicle completed a 12.42-mile race to the top of Colorado’s Pikes Peak last week. The Jeep Grand Cherokee, which operated without human interaction or remote control, finished the course in 47 minutes and 10.3 seconds, navigating 156 corners and 2000-ft cliffs with no guardrails (see the figure).
Axion Racing’s autonomous vehicle, known as Spirit, uses onboard computers to analyze data from GPS signals, lasers, stereo cameras, sound systems, and radar units to operate itself. Among the car’s components:
Mechanical Tools:
A reactive sensory enclosure (RSE) helps Spirit’s sensors “see” through sun glare, rain, and mud. Separate inputs for exist for gas, brakes, and emergency systems. Enhanced machine controller (EMC) devices, similar to those used by disabled drivers to control their cars, also are used to control the vehicle.
GPS Tools:
Spirit’s human team uses a high-end GPS+IMU+DMI unit to help battle a myriad of raceday reception issues. Spirit can drive autonomously without GPS. Multiple GPS units enable the vehicle’s human operators to check the signal strength and accuracy of each unit. Spirit provides an indicator when GPS is unavailable.
Forward-Looking Infrared (FLIR) Cameras:
FLIR cameras, long-wave fix-mounted cameras that help identify various objects, are used to determine temperature differences. The FLIR video camera system can see water, dirt, grass, people, and other temperature-producing obstacles.
Bumblebee Cameras:
Five stereo vision cameras, which feature two high-quality 1/3-in. progressive-scan CCD sensors each, are mounted around the race vehicle. These Bumblebee cameras are useful for the visualization of obstacles in off-road driving.
RGB Cameras:
RGB cameras are CCD visual tools that help Spirit’s team find a variety of obstacles in the vehicle’s path. They’re particularly useful in finding the road in front of the vehicle, as well as for visualizing obstacles for off-road driving.
Computing Tools:
Four Dell 2650 servers that run on a Linux operating system are used for rack computing power. Axion Racing uses LabVIEW, MySQL, Linux, and the team’s patent-pending AI code to command the Spirit.
Light Detection and Ranging (Ladar) Sensors:
Spirit’s ladar sensors can see grass, water, and rocks, as well as identify other obstacles in its path.
Path-Planning Tools:
Satellite imagery, USGS/NASA maps, 3D mapping, and route measurements are used to plan a path for autonomous operation. AI uses information to internally determine the quickest computed path from the first to last waypoint.
Sound Sensors:
Soundwave processing systems are used to determine terrain, outside environment, and inside operations. They’re also helpful for providing inputs for future blind drivers of autonomous vehicles. And, they’re an integral part of the vehicle’s “black-box” system.
For more detailed information, go to www.axionracing.com.
Reprints
