Title:Optimal Drive-by Sensing in Urban Road Networks with Large-scale Ridesourcing Vehicles

Abstract:The sensing and monitoring of the urban road network contribute to the efficient operation of the urban transportation system and the functionality of urban systems. However, traditional sensing methods, such as inductive loop sensors, roadside cameras, and crowdsourcing data from massive urban travelers (e.g., Google Maps), are often hindered by high costs, limited coverage, and low reliability. This study explores the potential of drive-by sensing, an innovative approach that employs large-scale ridesourcing vehicles (RVs) for urban road network monitoring. We first evaluate RV sensing performance by coverage and reliability through historical road segment visits. Next, we propose an optimal trip-based RV rerouting model to maximize the sensing coverage and reliability while preserving the same level of service for the RVs' mobility service. Furthermore, a scalable column generation-based heuristic is designed to guide the cruising trajectory of RVs, assuming trip independence. The effectiveness of the proposed model is validated through experiments and sensitivity analyses using real-world RV trajectory data of over 20,000 vehicles in New York City. The optimized rerouting strategy has yielded significantly improved results, elevating explicit sensing coverage of the road network by 15.0\% to 17.3\% (varies by time of day) and achieving an impressive enhancement in sensing reliability by at least 24.6\% compared to historical records. Expanding the path-searching space further improved sensing coverage of up to 4.5\% and reliability of over 4.2\%. Moreover, considering incentives for RV drivers, the enhanced sensing performance comes at a remarkably low cost of \$0.10 per RV driver, highlighting its cost-effectiveness.