TY  - JOUR
T1  - Implementation of a Deep Learning Platform for Post-Accident Response in Motorcycle Accidents
AU  - Park, Jaehan 
AU  - Choi, Mun Kyu 
AU  - Kwon, Dae Hyeon 
AU  - Shin, Soo-young 
JO  - The Journal of Korean Institute of Communications and Information Sciences
PY  - 2024
DA  - 2024/1/1
DO  - 10.7840/kics.2024.49.4.645
KW  - Motorcycle
KW  - Safety Platform
KW  - Deep Learning
KW  - Accident safety
AB  - This paper presents the implementation of a rapid accident detection and response platform for motorcycle rider safety in the event of an accident. The platform aims to utilize deep learning for forward vision analysis, IMU sensors for accident detection, and GPS for location tracking to enable swift accident response.
Enhancements in image recognition speed were achieved through deep learning-based license plate recognition of front vehicles, and the IMU's tilt values were used to detect whether the motorcycle is operating normally.
Additionally, the GPS tracks the motorcycle's location every second, allowing for rapid determination of the rider's location in the event of an accident. The collected location and front vision information are merged and transmitted to a server via a wireless communication module, facilitating initial accident response. Overall, the platform is developed with a focus on motorcycle accident detection and response, and its reliability has been validated through performance testing. The paper also proposes potential uses of the platform and future service development.