Complete Street design has long relied on qualitative principles, even as urban traffic grows more complex. A new arXiv preprint proposes a microscopic simulation framework that explicitly represents heterogeneous road users—autonomous vehicles, connected vehicles, human-driven vehicles, pick-up/drop-off services, and micromobility—to evaluate mobility and safety together. The framework captures lateral movements, parking maneuvers, pedestrian crossings and jaywalking, and shared-lane bicycle and scooter interactions.
Applied to a segment of Belmont Avenue in Chicago, the model shows that dedicated bicycle lanes reduce conflict severity at the expense of vehicular throughput. Increasing AV penetration helps mitigate congestion by reducing traffic shockwaves and improves interactions with pedestrians compared with CVs and HDVs. A genetic algorithm then optimizes candidate street layouts, identifying Pareto-efficient designs that balance mobility and safety.
The paper's contribution is integration: heterogeneous behavioral modeling, physics-based conflict assessment, stochastic performance evaluation, and design optimization within a single quantitative testbed. This gives transportation planners a predictive virtual environment for weighing tradeoffs before implementation. The work is under review for publication in the Transportation Research Record.