Adaptive traffic signal control for developing countries using fused parameters derived from crowd-source data

Mishra, Sumit; Singh, Vishal; Gupta, Ankit; Bhattacharya, Devanjan; Mudgal, Abhisek · 2022 · OpenAlex-citations

DOI: 10.1080/19427867.2022.2050493

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the challenge of traffic congestion in developing countries, where unplanned road networks, heterogeneous traffic, and a lack of lane discipline render traditional fixed-time signal controls ineffective. The authors argue that fully adaptive systems requiring expensive sensor infrastructure are often unfeasible in these regions. Consequently, the study proposes an inexpensive, adaptive traffic signal control method that utilizes crowdsourced data from Google Maps APIs to manage congestion in real-time without disrupting the established green-to-red time ratios. The methodology leverages the "Additive Increase Multiplicative Decrease" (AIMD) principle, a congestion control algorithm originally designed for TCP/IP networking. The system computes a Congestion Score (CS) for each intersection by fusing quantitative data (Estimated Time of Arrival, or ETA) and qualitative data (color-coded congestion levels) obtained via Google’s Distance Matrix and JavaScript APIs. The CS is derived by multiplying a color-based congestion measure, calculated from the pixel fractions of green, orange, red, and dark brown on map tiles, with an ETA-based measure weighted by historical travel times. This score determines the current congestion level (no, low, mild, or severe). Based on this level, the algorithm adjusts the signal cycle length: it additively increases the cycle time up to a maximum threshold as congestion worsens and multiplicatively decreases it when conditions improve, while maintaining the fixed split ratio of green and red times. The proposed framework was validated through field observations at three intersections in Delhi, India, during peak hours, and via simulation studies. The field data demonstrated that the adaptive congestion levels exhibited significant variation, proving sensitive enough for the control algorithm to act efficiently. The simulation results confirmed that the proposed AIMD-based control algorithm effectively reduces vehicle waiting times and overall congestion compared to static timing methods. The system is designed to operate with minimal processing power, allowing for implementation on low-cost microcontrollers or single-board computers, thereby avoiding the need for extensive roadside sensor deployment. The significance of this work lies in providing a scalable, low-cost alternative for traffic management in developing nations. By utilizing existing crowdsourced data and simple computational logic, the method offers a practical solution for intersections lacking sophisticated infrastructure. It balances efficiency and fairness in resource allocation (green time) among intersection approaches, mitigating the risks of traffic collapse and driver anxiety associated with poorly managed signals. This approach enables real-time congestion mitigation at a resolution of cycle length rather than hourly adjustments, offering a viable path toward smart mobility in resource-constrained environments.

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.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• traffic density