Stanford engineers to demonstrate software for controlling fleets of robo-cars

Engineers have spent the past decade developing and improving autonomous vehicles that use sensors and software systems to replace human drivers.

Now one Stanford professor is working on the next challenge: developing software systems to manage fleets of autonomous taxis, buses or shuttles.

“The work that we do can be viewed essentially as a giant dispatch system, whereby we are dispatching autonomous vehicles throughout an entire transportation network,” said Marco Pavone an assistant professor of aeronautics and astronautics at Stanford.

Pavone is part of a government-industry-academic consortium that is collaborating to create a test bed for an autonomous shuttle system.

That effort is called ARIBO, short for Applied Robotics for Installations and Base Operations, and its goals are to study how autonomous vehicles perform in real environments – moving passengers, avoiding pedestrians and generally serving as a shuttle system in a controlled environment, such as a military installation or college campus.

On Wednesday, Pavone will join his ARIBO teammates in Washington to take part in the SmartAmerica Challenge Summit, an event organized by the Presidential Innovation Fellows project. SmartAmerica is highlighting key advances in Cyber-Physical Systems, more popularly know as the Internet of Things – applications where networked technologies are revolutionizing everyday activities such as, in this case, riding in a robo-shuttle controlled entirely by software.

Professor Marco Pavone shows off the rug-size test bed where he is perfecting a system to dispatch fleets of autonomous shuttles, buses and taxis. (Video: Steve Schecter)

In advance of his visit to Washington, Pavone noted that Google recently announced plans to build and test a small fleet of self-driving cars.

“With Google … aggressively pursuing this technology, transportation networks with autonomous vehicles are becoming increasingly feasible,” he said.

Together with Rick Zhang, a graduate student in aeronautics and astronautics at Stanford, Pavone has developed systems to manage fleets of autonomous vehicles and prevent them, for instance, from bunching up at popular drop off points or the end of the line. At the moment, his “fleet” consists of fist-size models that scurry around the mock streets of a fake town in the basement of his Stanford lab.

“This test bed is a small-scale demonstration of what we hope to achieve on a real scale,” Zhang said.

After the SmartAmerica event in Washington, Pavone and Zhang will present their dispatch and control software at the Robotics: Science and Systems Conference in Berkeley, where their research has been nominated for best paper.