Achieving Realistic Cyclist Behavior in SUMO using the SimRa Dataset

TL;DR

This work tackles the mismatch between SUMO's default cyclist behavior and real-world cycling dynamics by leveraging the SimRa crowdsourced dataset to derive distributions of acceleration, deceleration, velocity, and left-turn choices. It introduces three velocity-based cyclist types (slow, medium, fast) plus a general class, reparameterizes longitudinal kinematics with distribution-based parameters, and implements an intersection-specific left-turn model, all integrated into SUMO. The approach is evaluated on Berlin, Munich, and Hanover scenarios, demonstrating substantial realism gains over the default SUMO bicycle model in acceleration, deceleration, velocity, and crossing durations, while also highlighting limitations from the dataset and e-bikes. The practical impact lies in enabling more accurate urban planning analyses and V2X safety studies that rely on realistic cyclist dynamics within SUMO-based simulations.

Abstract

Increasing the modal share of bicycle traffic to reduce carbon emissions, reduce urban car traffic, and to improve the health of citizens, requires a shift away from car-centric city planning. For this, traffic planners often rely on simulation tools such as SUMO which allow them to study the effects of construction changes before implementing them. Similarly, studies of vulnerable road users, here cyclists, also use such models to assess the performance of communication-based road traffic safety systems. The cyclist model in SUMO, however, is very imprecise as SUMO cyclists behave either like slow cars or fast pedestrians, thus, casting doubt on simulation results for bicycle traffic. In this paper, we analyze acceleration, deceleration, velocity, and intersection left-turn behavior of cyclists in a large dataset of real world cycle tracks. We use the results to improve the existing cyclist model in SUMO and add three more detailed cyclist models and implement them in SUMO.

Figures (14)

• Figure 1: Histogram of the empirical average velocity capabilities of cyclists found in the SimRa dataset. The average velocity of all bicycle rides after the preprocessing is 4.38 m/s with a standard deviation of 0.89 and a median of 4.42 m/s.
• Figure 2: Histogram of maximum acceleration capabilities found in the empirical SimRa dataset and their respective distribution functions. The red scalar represents the default value in SUMO.
• Figure 3: Histogram of maximum deceleration capabilities found in the empirical SimRa dataset and their respective distribution functions. The red scalar represents the default value in SUMO.
• Figure 4: Histogram of maximum velocity capabilities found in the empirical SimRa dataset and their respective distribution functions. The red scalar represents the default value in SUMO.
• Figure 5: Qualitative comparison between the SUMO default intersection model and real world data given by SimRa for the intersection between Alexanderstraße and Karl-Marx-Allee in Berlin. SimRa shows two distinct left-turn paths (i.e., a direct and an indirect one) whereas SUMO default only models the direct path.