Building a Smarter Car

Linda Cicero

You are driving down a slick two-lane highway when suddenly you see a truck sprawled across the lane ahead of you. You could veer into the oncoming lane, but another car is headed toward you. Whatever you do to avoid a collision will likely send you careening off the road. What's your best move?

In the future, your car may make that decision for you. Chris Gerdes, an assistant professor of mechanical engineering, and his students are busy working on a set of automobile control systems -- similar to antilock brakes -- that complement a car's mechanical systems. In that wet road scenario, for example, an average driver would almost certainly hit the oncoming car or spin off the road. But using mathematical analysis, Gerdes and his team are learning how to equip cars with sensors that counteract skids. Using imaginary walls on either side of the lane, the research also should help cars stay in the middle of the road, assisting them in steering clear of stationary objects and preventing loss of control on tight curves.

It may be serious study, but the student researchers could pass for kids playing with toys. They install a small computer and sensors in quarter-scale remote-control cars ($1,200 each) and run them in the alleys behind Old Union. Before they test-drive the system, they build elaborate computer simulations. The holy grail: a car that can drive itself.

Gerdes, who worked at Daimler-Benz before coming to Stanford, says the lab experience will help make the students better engineers once they enter the workforce. "They are learning to combine theory and practice," he says. And to prevent accidents.