MIT's ridesharing network is learning to dodge pedestrians

The collaboration with Ford gives students a lift while improving autonomous systems.

Lyft and Uber already operate in and around Boston, but students on MIT's campus in Cambridge, MA also have a new, research-oriented option for on-demand ridesharing. Ford and MIT announced today a new project that will shuttle students around on both campus walkways and city streets in a fleet of electric shuttles -- but the project isn't just for helping lazy college kids get to class, it will also be conducting research on pedestrian traffic patterns that will help it optimize the service, as well as future generations of autonomous vehicles.

In the first phase of the Ford-MIT project, which launches publicly in September, students can hail rides in a electric vehicle small enough to navigate sidewalks without pushing out the rest of the foot traffic. Over the past five months, the research team has actually been gathering data on how pedestrians move around campus by using vehicles equipped with LiDAR sensors similar to those used in most current autonomous cars. By combining that data with outside information like class schedules, current weather conditions and even "the dynamic habits of students and professors across different semesters," the team is able to anticipate where to route and position the vehicles during the day.

While this is all very helpful for creating an efficient ridesharing system, according to TechCrunch a secondary goal for Ford is to improve pedestrian detection in autonomous vehicle systems. With better LiDAR performance, self-driving cars should be able to rely on fewer cameras and potentially drive down costs while also getting better at navigating dense pedestrian zones. Eventually, the goal is to support Ford's Smart Mobility program and replace human-driven employee or campus shuttles with fully autonomous ones. They'd better get their homework done soon though: in Amsterdam, the Mercedes FutureBus is already making driverless trips to the airport.