Collision risk assessment at road intersection for autonomous vehicle

• Main Author: Li, Pingan
• Other Authors: Wang Dan Wei
• Format: Theses and Dissertations
• Language: English
• Published: 2019
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/78414
• Institution: Nanyang Technological University

With the development and application of autonomous vehicle technology, ensuring traffic safety has become a top priority. It is expected that autonomous vehicles can estimate the risk of collision with pedestrians at road intersection so that an accident can be avoided. This dissertation proposes a collision risk assessment approach based on human behavior analysis. The velocity risk model is obtained by recognizing and tracking the posture of the pedestrian. A position risk model is built by calculating the average distance from pedestrian to vehicle. The two models are combined to establish a collision risk model that can estimate the probability of pedestrian collision risk based on the perceived data of the visual camera and 3D LiDAR. The experiment was conducted on the basis of five different behaviors of pedestrians walking along the road, quickly passing through the intersection, slowly passing through the intersection, making a quick turn and then passing through the intersection, slow turning and then passing through the intersection test results of these five cases. The results show that the proposed collision risk model generates reasonable test results in all five data sets.