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

arXiv: https://arxiv.org/abs/2605.19834 · HTML: https://mlsanity.com/paper/2605.19834/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 7 of its 23 references that are in the map.
- First posted on arXiv 19 May 2026.

## Builds on (most cited first, 7 of 7)

- [On the Limitations of Physics-informed Deep Learning: Illustrations Using First Order Hyperbolic Conservation Law-based Traffic Flow Models](https://mlsanity.com/paper/2302.12337/) (2023)
- [Engineering the Neural Automatic Passenger Counter](https://mlsanity.com/paper/2203.01156/) (2022)
- [Individual Path Recommendation Under Public Transit Service Disruptions Considering Behavior Uncertainty](https://mlsanity.com/paper/2301.00916/) (2023)
- [Unified Occupancy on a Public Transport Network through Combination of AFC and APC Data](https://mlsanity.com/paper/2403.05546/) (2024)
- [Context-Aware Automated Passenger Counting Data Denoising](https://mlsanity.com/paper/2402.08688/) (2024)
- [Physics Enhanced Residual Policy Learning (PERPL) for safety cruising in mixed traffic platooning under actuator and communication delay](https://mlsanity.com/paper/2409.15595/) (2024)
- [Logic-informed reinforcement learning for cross-domain optimization of large-scale cyber-physical systems](https://mlsanity.com/paper/2511.00806/) (2025)

## Abstract (arXiv)

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.