In urban road transportation, intersections are traffic bottlenecks with increased waiting delays and associated adverse effects. A recently proposed intelligent intersection management (IIM) approach, the Synchronous Intersection Management Protocol (SIMP), synchronizes the vehicles access to simple single-lane isolated intersections, outperforming competing approaches in various performance metrics. In this paper, we apply SIMP to multi-lane intersections, increasing significantly the applicability of the protocol while dealing with the additional complexity emerging from the multiple crossing conflicts. Using the SUMO simulator, we compare the performance of SIMP with two conventional (Round-Robin - RR and Trivial Traffic Light Control - TTLC) and two IIM approaches (Intelligent Traffic Light Control - ITLC and Q-learning based Traffic Light Control - QTLC) under continuous and interrupted upstream traffic flows scenarios in urban settings. The results using a maximum speed of 30 km/h confirm the superiority of SIMP, improving traffic throughput (~ 14.4%) and reducing travel delays (~ 64.4%) and associated fuel consumption (~ 25.5%) when compared to the best of the other approaches.

Operating autonomous vehicles (AVs) and human-driven vehicles (HVs) at urban intersections while observing requirements of safety and service level is complex due not only to the existence of multiple inflow and outflow lanes, conflicting crossing zones, and low-speed conditions but also due to differences between control mechanisms of HVs and AVs. Intelligent intersection management (IIM) strategies can tackle the coordination of mixed AV/HV intersections while improving intersection throughput and reducing travel delays and fuel wastage in the average case. An endeavor relevant to traffic planning and safety is assessing whether given worst-case service levels can be met. Given a specific arrival pattern, this can be done via the worst-case response time (WCRT) that any vehicle experiences when crossing intersections. In this research line, this paper estimates WCRT upper bounds and discusses the analytical characterization of arrival and service curves, including estimating maximum queue length and associated worst-case waiting time for various traffic arrival patterns. This analysis is then used to compare six state-of-the-art intersection management approaches from conventional to intelligent and synchronous. The analytical results show the advantage of employing a synchronous management approach and are validated with the vehicles floating car data (timestamped location and speed) and simulations carried out using SUMO.