Evaluating and Validating Technology Options for Estimating Transit Vehicle Occupancy in Real Time
archive: archived pipeline: cataloged verified
Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)
Summary
This study addresses the challenge of accurately estimating real-time transit vehicle occupancy, a critical metric for evaluating service efficiency and enhancing passenger experience. While real-time vehicle location data is widely available, occupancy data remains scarce due to the high costs and labor intensity of traditional methods like manual surveys, automatic passenger counters (APCs), and automatic fare collection (AFC) systems. The research, conducted by Florida State University for the Florida Department of Transportation, aimed to evaluate and validate emerging technology options for real-time occupancy estimation. The project involved a comprehensive literature review to identify potential technologies, followed by technical and non-technical evaluations covering factors such as accuracy, latency, cost, privacy, and implementation ease. Based on these evaluations, Wi-Fi probing was identified as a promising technology due to its low hardware requirements, real-time data availability, and minimal privacy concerns. The researchers conducted pilot studies in three Florida transit systems: StarMetro in Tallahassee, Lynx in Orlando, and Miami-Dade Transit. Data collection involved capturing Wi-Fi frames using onboard hardware while simultaneously recording ground-truth passenger counts through manual surveys. A significant technical challenge emerged during the study: MAC (Media Access Control) address randomization in modern mobile devices prevents the unique identification of passengers, rendering traditional counting methods ineffective. To overcome this, the team developed a data-driven learning algorithm that utilizes various Wi-Fi frame features rather than relying solely on unique MAC addresses. The results demonstrated that the data-driven approach could effectively estimate vehicle occupancy despite MAC address randomization. In the StarMetro case study on the Evergreen route, the model achieved an R² value of 0.84, indicating strong predictive performance. Comparable results were observed in the Orlando and Miami pilot studies. The analysis highlighted that feature engineering and hyperparameter tuning were crucial for success, with signal strength and packet counts serving as important predictors. The study noted that the two-week data collection period was a limitation; continuous, long-term data collection could further improve model accuracy by allowing better filtering of noise and more robust training. The significance of this research lies in demonstrating that Wi-Fi probing, combined with machine learning, offers a viable, cost-effective alternative to expensive APC systems for real-time occupancy estimation. The findings provide transit agencies with a validated method to integrate occupancy data into passenger information systems, potentially improving rider decision-making and operational efficiency. The report concludes that while MAC randomization poses challenges, it does not preclude the use of Wi-Fi data for occupancy estimation, provided that appropriate data-driven algorithms are employed. This work advances the field by offering a scalable solution for agencies seeking to enhance real-time transit information without significant capital investment.
Key finding
A data-driven machine learning algorithm using Wi-Fi frame data achieved an R2 value of 0.84 in estimating transit vehicle occupancy, successfully overcoming the challenges posed by MAC address randomization.
Methodology
mixed_methods
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | rosap | — | — | 2 | 2026-05-23 |
| archive | success | — | — | — | 1 | 2026-05-23 |
| extract | success | cached | — | — | 2 | 2026-06-10 |
| clean | success | — | — | — | 1 | 2026-06-01 |
| chunk | success | — | — | — | 1 | 2026-06-01 |
| embed | success | — | — | — | 1 | 2026-06-02 |
| enrich | success | — | — | — | 1 | 2026-05-23 |
| promote | success | — | — | — | 1 | 2026-05-23 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 3 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 24 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Empirical Findings: observational prevalence