Fall Accident Detection System for Bicycle Riders using Support Vector Machines
Saurav Gupta, Ramanathan V, A. Sasithradevi · 2022 IEEE 11th International Conference on Communication Systems and Network Technologies (CSNT) · 2022
Bicycle riders have a very high impact on accident injuries when compared to motorists. Indian road accident report-2019 shows an increase in the fatal rate of 14% accidents to bicyclists when compared to 2018. Accidents occurring in urban areas at night may contribute to a delay in reporting the accidents to emergency centers. Therefore, a fast system development is essential to detect and notify the accidents to the nearest health care centers. This article proposes a compact design of hardware to fulfill the necessity. This hardware module is responsible to measure the features related to bicycle riding and fall accidents. For measuring the features, a magnetic, angular rate, gravity (MARG) sensor-based system is used. It consists of an accelerometer, gyroscope, and magnetometer for measuring the signals like acceleration, angular velocity, etc. Here, the popular machine learning technique called the support vector machines (SVM) algorithm is used for the detection of fall accidents. Simulation results are compared with the existing system to illustrate the improved accuracy of 95% of the fall accidents for bicycle riders.