Analyzing Fundamental Diagrams of Mixed Traffic Control at Unsignalized Intersections

TL;DR

The paper tackles how increasing robot-vehicle (RV) penetration reshapes fundamental diagrams (flow–density–speed) at unsignalized intersections. It uses SUMO simulations across four intersections with RV penetration ranging from 0% to 100% in 25% steps, analyzing speed $V$, density $k$, and flow $Q$, and fits FD curves such as $Q = a k^2 + b k + c$ to describe the relationships; an autonomous decentralized reinforcement-learning controller guides RVs at entrances. Findings show the typical flow–density pattern up to a capacity peak $Q_{max}$ at $k_{crit}$, but the peak and its location vary with RV rate and across intersections, indicating non-linear, locale-dependent effects. The results imply that intersection-specific traffic management is needed during the transition to higher RV presence, and future work should extend to city-scale networks with reconstruction and estimation techniques.

Abstract

This report examines the effect of mixed traffic, specifically the variation in robot vehicle (RV) penetration rates, on the fundamental diagrams at unsignalized intersections. Through a series of simulations across four distinct intersections, the relationship between traffic flow characteristics were analyzed. The RV penetration rates were varied from 0% to 100% in increments of 25%. The study reveals that while the presence of RVs influences traffic dynamics, the impact on flow and speed is not uniform across different levels of RV penetration. The fundamental diagrams indicate that intersections may experience an increase in capacity with varying levels of RVs, but this trend does not consistently hold as RV penetration approaches 100%. The variability observed across intersections suggests that local factors possibly influence the traffic flow characteristics. These findings highlight the complexity of integrating RVs into the existing traffic system and underscore the need for intersection-specific traffic management strategies to accommodate the transition towards increased RV presence.

Figures (2)

• Figure 1: The four intersections in our study.
• Figure 2: Fundamental diagrams for the four intersections with different RV penetration rates.