Google certainly has pockets deep enough to trick out self-driving cars with any kind of pricey gear, but researchers at the University of Oxford have begun testing a solution that aims to keep things affordable. Currently, the system leverages an array of low-profile stereo cameras and lasers that rings up at about £5,000 (approximately $7,750), but the next goal is to knock the price down to £500, and eventually to a cool £100 (roughly $150). "Really, we do need to solve the engineering challenges of not relying on expensive sensors, but relying on cheap sensors," Professor Paul Newman told the Telegraph. "But doing some really smart things with those cheap sensor feeds."
Rather than a vehicle that acts as a chauffeur at all times, Newman's vision for the modified Nissan Leaf, dubbed RobotCar, is for it to take control on select occasions. While drivers go about their commute, the system composes a 3D map of the car's environs and commits it to memory. When the auto identifies a familiar setting and feels confident about its ability to take the reigns, it could let the driver know it's ready to assume control. Right now, the automobile's been tested on private roads, but the team behind it is working with the UK's Department of Transportation to roll it onto public streets. Head past the jump for a glimpse of RobotCar in action.