Robust Multi-Task Framework for Motorcycle Rider Safety in Unstructured Traffic
INTERNATIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT · 11.04.2026 · 10 (04) · 1-9
DOI: 10.55041/ijsrem59861 ↗Kurz gefasst
Fachartikel in der Zeitschrift INTERNATIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT, online erschienen am 11.04.2026. Verfasst von Ganesh S. Gupta und 1 weiteren Autor:innen.
Abstract (Originalsprache)
– The increasing prevalence of two-wheelers in urban centers has made enforcing rules regarding safety a major issue, due to the presence of uncontrolled and chaotic traffic conditions. Traditional surveillance systems were not designed to function effectively with cameras pointing at varying angles, or due to camera motion blur, or other unpredictable rider behaviors. In this paper, we present a Robust Multi-Task Framework (RMTF) that detects compliance with helmet use law and identifies instances of triple riding simultaneously in a real-time environment. To achieve this, we utilize a single YOLOv8-based architecture to process handheld (mobile) recorded video data under dynamic viewing angles with noise to extract metrics about safety. Our system uses both spatio-temporal logic techniques to recognize multiple-rider scenarios, and also includes custom pre-processing technique (e.g., de-blur algorithms) to improve License Plate Recognition (LPR) under severe motion-blur images and relatively low contrast conditions that can be found in the real world. Furthermore, our validation test datasets were obtained from custom roadside video recordings that reflect the non-standardized characteristic of real-world traffic, contrasting with most current models trained on static pre-recorded datasets. The results of our experiments show a high mAP (>92%) and low latency, which shows the utility of our framework in providing a portable and low-cost solution for law enforcement efforts aimed at enforcing public safety. Finally, this solution represents a scalable method for improving the safety of the public and improving the efficiency of law enforcement. Key Words: Intelligent Traffic Surveillance, YOLOv8, Helmet Compliance, Triple-Riding Detection, Unstructured Traffic, Multi-Task Learning, Mobile-Based Surveillance
Zitieren
APA
Gupta, G. S., & Shetye, S. N. (2026). Robust Multi-Task Framework for Motorcycle Rider Safety in Unstructured Traffic. INTERNATIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT, 10(04), 1-9. https://doi.org/10.55041/ijsrem59861
Vancouver
Gupta GS, Shetye SN. Robust Multi-Task Framework for Motorcycle Rider Safety in Unstructured Traffic. INTERNATIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT. 2026;10(04):1-9. doi:10.55041/ijsrem59861
BibTeX
@article{Gupta2026,
author = {Gupta, Ganesh S. and Shetye, Sanika N.},
title = {Robust Multi-Task Framework for Motorcycle Rider Safety in Unstructured Traffic},
journal = {INTERNATIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT},
year = {2026},
volume = {10},
number = {04},
pages = {1--9},
doi = {10.55041/ijsrem59861}
}
EUVIMED