Automotive technology is rapidly advancing to the point where we eventually won't need to actually drive a car ourselves in most real world scenarios. At the forefront of this development work is Volkswagen and its new electronics lab in Palo Alto, California that's operated in conjunction with Stanford University. VW and Stanford have a long relationship in autonomous vehicle development and are continuing to develop "Junior," an autonomous Passat wagon that was built for the last DARPA Urban Challenge a couple of years ago. We recently got to play passenger in the backseat of Junior for a demonstration and were quite impressed with its parking prowess.
The basic hardware technology that makes Junior work is fairly straightforward: a mix of GPS, radar, laser, optical and inertial sensors. The hard part is creating the software algorithms that allow the system to detect and work through the more esoteric scenarios that arise in the real world.
The videos after the jump show just far those algorithms have come: Junior can now reverse power-sliding into a parallel parking spot. While this particular scenario is probably not something you want your autonomous car to do without your permission, learning how to do it is important. In the real world there are many variables that cannot be programmed into the controls. The Stanford engineers are developing methods to allow the system to blend instructions for basic scenarios with learned responses from situations like this. Check out the videos after the jump. A tip of the hat to James!
[Source: Crunchgear, Youtube]