Genetic Algorithm–Based Approach for Optimal Deployment of Roadside Units in Connected and Automated Vehicle Environments

Abstract

Traffic management strategies play a critical role in connected and automated vehicle (CV/AV) environments in which vehicles will be the primary source of traffic information. In CV/AV environments, traffic data are collected from vehicles, and decisions are made at the transportation management centers (TMCs) and communicated back to vehicles via deployed roadside units (RSUs). This process poses a challenge at the early deployment stages of the technology because of the anticipated low market penetration. Under such conditions, RSUs must be optimally located throughout the network to provide continuity and expand the coverage of vehicle-to-vehicle communication systems. This study presents a genetic algorithm–based approach for determining the optimal locations of RSUs by maximizing the connectivity robustness measure, by taking into consideration vehicle clustering (groups of vehicles in transmission range of each other), network size, and other factors. Traffic simulation data were generated from a microscopic simulation platform (VISSIM) and used to test the proposed approach for different penetration rates. The results show that the proposed approach identified locations where more vehicles can communicate. The optimized RSU locations enabled communication between more vehicles in the network, which was identified by the increased robustness of connectivity. This aspect can maximize the amount of exchanged information between vehicles and the RSUs. Consequently, better traffic monitoring can be achieved by collecting more representative data of the traffic conditions in the network. Thus, optimal decisions, such as vehicle rerouting, are made at the TMCs and disseminated to as many vehicles as possible, which helps in achieving better management.