CSP 2017 Online Program

A Hierarchical Clustering Analysis (HCA) in Automatic Driving Regarding Vehicle-to-Vehicle Pedestrian Position Identification (303417)

Keywords: Statistics learning, Clustering methods, big data analysis, hierarchical clustering analysis, K-means algorithm, dynamic data

To correctly identify the pedestrian position is a challenging but crucial topic in the autonomous driving area, also is an arising area in need of using latest big data and data science techniques. In this poster, we present an application of a hierarchical clustering (HCA) statistics learning algorithm in determining the location and amount of pedestrians detected by different vehicles equipped with a Pedestrian Autonomous Emergency Braking (PAEB) system, taking into consideration the inherent inaccuracy of the pedestrian sensing from these vehicles. In order to compare with the hierarchical clustering analysis, we also applied K-mean clustering algorithm to recognize and reconstruct actual pedestrians. The HCA method showed better and much robust result compare to K-means since the HCA structure takes the block between subject into the calculation. The HCA method determines the possible number of actual pedestrians by grouping the nearby pedestrians sensed and broadcasted together by different vehicles and considers them as one pedestrian. The computer simulations illustrate the effectiveness and applicability of the proposed method. The results provide more realistic information for vehicle Autonomous Emergency Braking (AEB) systems to make better decisions to avoid crashing into pedestrians.