Deployment Optimization of Roadside Sensing Units Based on NSGA-II for Vehicle Infrastructure Cooperated Autonomous Driving

Currently, the Vehicle Infrastructure Cooperated Autonomous Driving has been rapidly developing, and the deployment of road-side perception devices for vehicles has also become a hot research topic. However, the traditional deployment methods based on human experience suffer from problems such as high cost, subjective factors, uneven quantification of coverage area, and lack of the performance models of sensors. Therefore, there still lacks a scientific and systematic deployment optimization schemes and evaluation system based on sensor performance models for road-side device deployment. To address these issues, this paper, which is based on research of current perception devices, proposes a genetic algorithm and NSGA-II algorithm by using a multi-objective optimization method for coverage, redundancy, and cost on two-dimensional grid maps of typical road surfaces, such as intersections, using sensor performance models. Optimize coverage, redundancy, and cost while meeting the constraints of coverage and redundancy. To further verify the model algorithm, real road scenario experiments are conducted. The input of the experiment includes the perception sensor type, two-dimensional grid map, and basic algorithm parameters while the output is the Pareto-optimal solution for perception devices deployment. The experimental results show that the coverage rate converges to 100%, the redundancy rate converges to 1%, and the cost is optimized by 73.1%, which is much faster and accurate than traditional GA. In a nutshell, the innovation of this paper lies in the modeling of roadside sensing units and applying NSGA-II algorithm for global deployment optimization on road grid maps.