Floating Car Observers in Intelligent Transportation Systems: Detection Modeling and Temporal Insights

TL;DR

This work tackles the challenge of enriching microscopic traffic data with Extended Floating Car Data by introducing Floating Car Observers (FCOs) that share sensor-derived detections. It compares multiple detection-modeling approaches—2D/3D raytracing, high-fidelity CARLA-SUMO co-simulation, and a neural emulation framework—to estimate FCO detections and integrates temporal insights to recover previously unseen vehicles. The neural emulation, especially an encoder-decoder transformer, delivers fast, scalable predictions while achieving high accuracy close to co-simulation benchmarks, and temporal recovery substantially enhances traffic state information at practical penetration rates (e.g., at $p=0.20$, LiDAR-based FCOs identify $0.65$ of vehicles). Overall, FCO-based xFCD markedly improves traffic state estimation and monitoring across varying penetration levels and traffic conditions, with temporal information enabling richer, more robust ITS applications and paving the way for future DRL-driven traffic control strategies.

Abstract

Floating Car Observers (FCOs) extend traditional Floating Car Data (FCD) by integrating onboard sensors to detect and localize other traffic participants, providing richer and more detailed traffic data. In this work, we explore various modeling approaches for FCO detections within microscopic traffic simulations to evaluate their potential for Intelligent Transportation System (ITS) applications. These approaches range from 2D raytracing to high-fidelity co-simulations that emulate real-world sensors and integrate 3D object detection algorithms to closely replicate FCO detections. Additionally, we introduce a neural network-based emulation technique that effectively approximates the results of high-fidelity co-simulations. This approach captures the unique characteristics of FCO detections while offering a fast and scalable solution for modeling. Using this emulation method, we investigate the impact of FCO data in a digital twin of a traffic network modeled in SUMO. Results demonstrate that even at a 20% penetration rate, FCOs using LiDAR-based detections can identify 65% of vehicles across various intersections and traffic demand scenarios. Further potential emerges when temporal insights are integrated, enabling the recovery of previously detected but currently unseen vehicles. By employing data-driven methods, we recover over 80% of these vehicles with minimal positional deviations. These findings underscore the potential of FCOs for ITS, particularly in enhancing traffic state estimation and monitoring under varying penetration rates and traffic conditions.

Figures (4)

• Figure 1: Visualizations of the 2D and 3D raytracing modeling approaches
• Figure 2: Comparison of the CARLA high-fidelity and microscopic SUMO traffic simulations, along with sensor perspectives including RGB camera and LiDAR outputs from a simulated vehicle.
• Figure 3: Illustration of the interaction between CARLA (blue), SUMO (green), and the network architecture of the emulation approach (orange) for training the emulation method. Further, the components for the loss calculation to train the emulation network are visualized (pink).
• Figure 4: Excerpt of the SUMO simulation visualizing the investigated intersections for the FCO and temporal insights potential.