Self-Driving Cars Gear Up to Predict Foot Traffic

Self-Driving Cars Gear Up to Predict Foot Traffic

autonomous_car_cockpit.594bdcfa4b22b

Self-driving cars and AI technology are making headlines around the world. However, the technology is yet to fulfill its promise to millions at large. Soon, researchers from University of Michigan (UMich) could change that.

UMich researchers are studying human body symmetry, gait, and foot movement. They are using GPS, onboard cameras in vehicles, LiDAR to gather data. The data is then fed into a 3-D computer stimulation program to recreate a virtual reality. Currently, a bio-mechanically aided virtual environment helps them catalog foot traffic meticulously.

This environment has made it possible now for self-driving cars to predict foot movement within 50 yards, in time. A city intersection often scales up to 50 yards.

The research is a much-needed advancement from previous work in the area. Earlier, researchers were focussed on studying still images, according to Ram Vasudevan, assistant professor at UMich. Ram believes, it is essential for self-driving cars to experience the world in 3 dimensions.

Learning to predict the hard way

In order to do so, it is important for the next-gen vehicles to learn the pace of a pedestrian, understand symmetry of limbs and the impact of foot placement on stability.

According to Matthew Johnson-Roberson, associate professor at UMich, the researchers are training these vehicles to predict movement. The research helps cars map out pedestrian’s movement with ease.

The research will be bolstered in the near future. The team has parked numerous level-4 autonomous vehicle at several intersections in Ann Harbor.

Another U-M research engineer Xiaoxiao Du believes, the UMich team is contributing to a healthier, safer, and more efficient living environment.