Resource: Robot Race

Most of us take for granted our ability to sense details in our environment and to identify and traverse safe routes through a world riddled with hazards. What's more, we perform this amazing process instantly and automatically. Robotics engineers know that for a robot to analyze and respond to even a fraction of the information about the environment that our brains process, it must have some of the most advanced computer hardware and software available. Take, for example, the twin exploratory rovers that landed on Mars early in 2004. After more than two years on the Red Planet, each rover had traveled only about three kilometers (1.8 miles) without human assistance. Yet it took engineers years to develop the computer software and hardware for these complex robots.

The 2005 DARPA Grand Challenge demanded a much faster pace than that held by the Mars rovers. The engineering teams that took part in the challenge had to develop autonomous vehicles that could cover the 210-kilometer (130-mile) course in the fastest time possible. Despite the distinct advantage of operating their vehicles on Earth, no team had come even close to completing the racecourse in the previous year. Nevertheless, twelve teams put their designs to the test in the 2005 challenge. Five would actually finish.

The teams that completed the 2005 DARPA Grand Challenge course all used a combination of sensing technologies that measured the surrounding environment. Their equipment included stereo-vision cameras, laser range finders, and radar. However, perhaps more important than the equipment each used to sense the environment was the computer hardware and software that compiled and analyzed incoming data to create a picture of the road and the hazards that lay ahead of the vehicle.

The team that won the challenge by a margin of 11 minutes over the second place robot used adaptive vision software to combine data from five laser range finders, one video camera, and radar. This program created what the team called a "cost map" in which hazards were color-coded relative to their risk or potential cost to the robot. In addition, the winning team's software enabled the robot to learn over time so that it could replicate its successes and avoid previous missteps.

No one yet knows how the innovations developed during the 2005 DARPA Grand Challenge will impact society, but we can all speculate. Potentially, autonomous vehicles may be more reliable than those piloted by drowsy, overworked, or distracted human drivers. Such vehicles may also allow commuters to be more productive as they travel to and from work. In addition, efforts to develop autonomous vehicles that operate reliably on Earth will surely speed efforts to improve autonomous navigation on unmanned spacecraft sent to explore other planets. Such advances will enable these missions to cover far more ground than ever before.

To learn more about remote sensing technologies, check out Sound Waves Underwater: Experiment with Sonar, Collecting Data Below the Earth's Surface, and Sound Waves Underwater: The Loch Ness Monster.

To learn more about the Mars robotic rovers, check out Anatomy of a Rover, Mars Dead or Alive: A Hostile Environment, and Mars Dead or Alive: Where to Land?.

To learn more about the history of robots in space, check out Robotic Exploration of Space Timeline.

• Describe a situation in which a driverless vehicle would be useful.
• Would you feel safe riding in a vehicle that was a robot and not operated by a person? Why or why not?
• Any engineering challenge usually has more than one solution. What were different ways teams solved this challenge?
• What was the role of planning in this process? Did trial and error play a role? Give an example.
• How would you define "robot"?