Roadside Visual Sensor Deployment in Urban Networks Considering Vehicle Control Boundary

Reasonable deployment of roadside perception sensors is important for intelligent and connected vehicles (ICVs). This article proposes a novel roadside camera deployment method that considers control security boundaries. We first establish a roadside camera geometry model to determine an object’s presence within the perception domain. Then, a novel camera localization error model is developed, relating allowable localization errors to the maximum camera perception range. Then, by integrating vehicle lateral and longitudinal control models, an analysis framework for allowable error is introduced to define the control error boundaries for assisting in determining the maximal perception range. Subsequently, based on the camera geometry model and considering allowable localization errors, an equivalent camera group model is constructed using the greedy search method (GSM) for multiple identical roadside cameras, which aims to achieve a wider longitudinal field of view. Lastly, leveraging the camera group model, a dynamic programming (DP) optimization method is employed to deploy roadside cameras in a flexible decision pattern along the road centerline. The effectiveness of our work is validated through a joint simulation experiment on a real-world road network. The results show that the proposed camera group model can achieve optimal sensor deployment under the premise of ensuring the safety of the vehicle’s control performance.