Smart Robot Design and Implementation to Assist Pedestrian Road Crossing

At some point each day every one of us is a pedestrian. Unfortunately, pedestrian fatalities remain to be high. In 2020 alone, 6,516 pedestrians were killed, and an estimated 55,000 pedestrians were injured nationwide. According to the National Highway Traffic Safety Administration, children ages 16 and under are the most at risk of getting hit while crossing the street. In California alone, six million students attend 10,453 public schools on any given school day. Of these students, more than 3.1 million are of elementary school age. 

There has been only a handful of studies on developing systems for pedestrian safety most of which are reply on the assumption that there already exists a traffic light infrastructure and we could add an additional framework onto it. In this project, it is crucial to accurately detect the vehicles in the street as well as pedestrians at the crossroads in order to make an intelligent decision. We will use advanced Machine Learning algorithms for vehicle as well as pedestrian and cyclist detection in real-time. To do that we will initially be gathering large data sets from the four different sensors and training the advanced machine learning models. The machine learning algorithms will be programmed using Python running on a minicomputer. By combining the information from the four sensors, a LIDAR, thermal infrared cameras, RADAR sensors, and video cameras, and by integrating the data to an efficient and accurate assessment of the road conditions the robot will make an intelligent decision on when to walk the pedestrians/cyclists through the crossroad intersections.