A Closed-loop, State-centric, Multi-agent Framework for Passenger Load Estimation from Heterogeneous Data Streams

2026-05-19

AI Summary

This study addresses the challenge of inaccurate passenger load estimation in public transit systems arising from heterogeneous data streams—such as automated passenger counting systems—due to incremental errors, conflicting evidence, and fluctuating sensor reliability. To tackle this, the authors propose a closed-loop, state-centric multi-agent fusion framework that treats stop-level events as a unified backbone. The approach tightly couples perception, physical modeling, and fusion loops to infer passenger loads incrementally at each stop. Crucially, it integrates physical feasibility constraints throughout the inference process, dynamically allocates trust among multiple evidence sources, and employs residuals from physical violations for closed-loop calibration and adaptive learning. Experimental results demonstrate that the proposed method significantly enhances both accuracy and robustness of passenger load trajectory estimation under complex operational conditions.

Abstract

To support operations and passenger-facing services, transit agencies need reliable passenger load trajectories. Currently, load estimates are typically inferred from imperfect sensing systems rather than fully observed, and the accuracy of modern automatic passenger counting (APC) systems still varies with station layout, flow intensity, and operating conditions. To address the challenges of robust passenger load estimation from heterogeneous data streams, including incremental count errors, evidence conflicts, and context-dependent sensor reliability, we propose a closed-loop, state-centric, multi-agent framework. This method enforces physical feasibility at every step, allocates trust dynamically among evidence sources, and feeds physics-derived violation residuals back into training for robustness improvement. The architecture consists of a unified stop-event backbone, a coupled Perception--Physical--Fusion loop for stop-by-stop inference, and optional trip-level macro-correction and closed-loop calibration modules.

Problem

• passenger load estimation
• heterogeneous data streams
• automatic passenger counting
• sensor reliability
• multi-agent framework

Innovation

• closed-loop estimation
• state-centric framework
• multi-agent fusion
• heterogeneous data streams
• physical feasibility constraints