A new research collaboration between Ford and MIT studies how pedestrians move in densely populated areas in order to improve transportation such as point-to-point shuttle services—and the data will also help Ford beef up its autonomous and driver-assist technologies.
The MIT project is designed to enhance models that predict demand for mobility-on-demand services, such as the dynamic shuttle that Ford’s employees use at its Michigan headquarters. During the past five months, researchers have used sensors and cameras mounted to the vehicles to record foot traffic between different points on campus. (LiDAR is considered to be far more accurate than GPS when it comes to detecting and localizing objects from the surrounding environment.)
Ford and MIT then use the data to study the overall pattern of how pedestrians move across campus, which helps them predict where the most demand for the shuttles will be at any given moment. Based on that information, the shuttles can be pre-positioned and routed to serve the campus as efficiently as possible. Other factors researchers take into account include weather conditions, class schedules, and even the habits of individual professors.
“Researchers are using LiDAR sensors, cameras, machine learning, and algorithms to measure pedestrian flow and accurately predict demand for shuttles,” said Bryan Goodman, manager and technical leader in Ford’s division of autonomous vehicle analytics. “Autonomous vehicles will use 3D LiDAR, and that will be the future, but MIT’s goal was to see how far they could push lower-cost 2D LiDAR and still get good results.”
Over the summer, a group of faculty and students were granted access to three small electric shuttle vehicles (see photo) that resemble futuristic golf carts and are small enough to traverse sidewalks and other areas typically off-limits to cars. Starting in September, the vehicles will be available to the entire campus; riders summon the vehicles by using an app on their smartphones. Goodman said MIT’s campus, with its abundance of pedestrian-filled quads, made it an ideal place to do the research.
“MIT’s security service already used these small vehicles, and the Aerospace Controls Lab specializes in predicting movement and detecting presence,” Goodman explained. “How do you optimize the position of the shuttle?”
The MIT researchers Ford is working with study topics related to autonomous systems and control design for aircraft, spacecraft, and ground vehicles, particularly estimation and navigation, planning and learning under uncertainty, and vehicle autonomy.
Ford plans to take the findings from the MIT collaboration and incorporate them into its own in-house mobility services. The company hopes to reduce the wait times for its dynamic shuttle, which provides point-to-point rides to employees via mobile app on its Dearborn, MI, campus. The MIT project is one of more than 30 mobility research initiatives currently underway between Ford and universities in the U.S., Germany, and China.
