RF & Microwave Technology
Technology to Precisely Predict Pedestrian Movement for Self-Driving Cars
University of Michigan engineers have developed a "biomechanically inspired recurrent neural network" that interprets pedestrian gait to predict future positions in 3D. The research has immediate applications in driverless cars. Data collected by vehicles through cameras, LIDAR, and GPS allow the researchers to capture video of people in motion and then recreate them in 3D computer simulation. By focusing on peoples' gait, foot placement, and body symmetry, the researchers are teaching autonomous cars to recognize and predict pedestrian movements with greater precision than current technologies.