Optimizing the Deployment of Automated Speed Camera at the Intersections Using GPS Trajectories

Abstract

The economical and rational deployment of automated speed camera is a critical issue for traffic police department to implement speed management efficiently. Based on taxi GPS trajectories collected from Chengdu, 2016, this study optimizes the deployment interval and number of ASCs at the intersections by using K-means clustering and kernel density estimation according to the critical mixed distance halo effect and the delta speed distribution, respectively. Results illustrate that speeding is more likely to happen within the speed limit of 40 km/h rather than 60 km/h. From the whole perspective, with the growing deployment number of ASCs, the upstream distance halo effects gradually increase, while the downstream distance halo effects gradually decrease within the range of about 4500 m. Given that the interaction between two adjacent ASCs, the critical mixed distance halo effect of ASCs is about 215 m and 529 m corresponding to the smaller and larger values of critical delta speed in the northeast direction respectively, and about 315 m and 585 m in the southwest direction. Generally, one ASC should be deployed every 500 m and 600 m within the speed limit of 60 km/h, and every 200 m and 300 m within the speed limit of 40 km/h in the northeast and southwest directions, respectively.