Aiming at the conflict between right turning vehicles and straight bicycles at signalized intersections, the behavior characteristics of vehicles and bicycles were analyzed. According to the initial release characteristics of green light of non-motor vehicles, an expansion model was established, and the expansion degrees of different signal types were compared. From the perspective of driving psychology, the driving style of vehicles and bicycles was analyzed. Through video capturing and image processing, the average speed of right turning vehicles passing through the intersection under the influence of bicycles at the intersection of Shanghai Kongjiang Road and Longchang Road was extracted. K-means clustering analysis was used to classify drivers into cautious, robust and radical type for the speed of right turning vehicles passing through the conflict zone, and the speed and trajectory of right turning vehicles under the influence of straight bicycles were analyzed.