Pedestrian Intent Detection using Skeleton-based Prediction for Road Safety

To predict the intent of human beings walking on the street or crossing the road can be an important safety feature for intelligent cars. It has drawn the attention of the automotive industry in the past few years. Estimating whether the pedestrian is going to cross the road or not is a very challenging task because there are multiple factors involved such as the speed at which the pedestrian is walking, whether the person is walking alone or walking with someone, are they aware of the vehicle etc. The Pedestrians Intent Detection system explores the different factors that affect the pedestrian's intent while walking or crossing the road as well as how different techniques are used to predict the intent of the pedestrian considering those factors. This research work will help design the system which predicts pedestrian's intent much like the human ability to interpret the behaviour of the pedestrian.