Autonomous vehicles could someday take the driver out of the driver's seat. Now the debate centers on which sensors offer researchers the most performance for self-navigation. For years man has envisioned a world of smart cars traveling on highways, self-navigating planes soaring overhead and unmanned underwater vehicles lurking in the ocean depths. But with the exception of a few promising prototypes fresh from the research lab, the dream is still just a dream. There have been significant strides along the way, however, driven by advances in cameras, laser scanners and IR sensors. The unmanned helicopter, which has flown successful missions over the Houghton crater north of the Arctic Circle, represents one of the most advanced flying prototypes. The latest generation of autonomous ground vehicle, embodied in Carnegie Mellon University's Navlab series, incorporates an onboard vision system that relies on charge-coupled device cameras, laser rangefinders and radar. These tools enable the vehicles to drive themselves at high speeds, while following the correct lane and avoiding obstacles.