The first Stanley won the DARPA Grand Challenge:
http://www.grandchallenge.org/
That URL is likely to change, by the way, from the 2005 win, around November of this year... google it for better hits if so.
In any case, that was about getting a car to drive autonomously for over 100 miles through the desert. Obstacles were few although when they did arise they often lead vehicles to disaster. Stanford won the race by a significant margin by using a Neural Nets approach, having the car tail a professional driver and learn what "good driving" looks like in terms of data. It innovated in the Neural Nets space by adding criticism - so data it had recorded as "good driving" that later proves wrong frequently would now be removed, rather than just always growing the data and hoping to statistically drown wrong results.
So, no not a solar car.
This Stanley 2.0, aka Junior, needs to drive in a much more constrained environment. It's being called the Urban challenge but will probably be done in an abandoned town, and vehicles will need to avoid other vehicles at places like intersections.
The beauty of this is that it may someday lead us to cars that get us from A to B without our help. Most likely the US's lawsuit-happy, sue-everything-rewarding climate will never allow for that however. The really sad part is that this is the DARPA Grand Challenge - as in the military's research arm, the only research arm in the US with a lot of money. So add a big gun to the top of all these self-driving vehicles and picture them in countries filled with poor people, and you get the more likely result of success here. Bittersweet.
http://www.darpa.mil/grandchallenge/index.asp