Accident Detection System for Bicycle Athletes Using GPS/IMU Integration and Kalman Filtered AHRS Method

With a view to reducing unforeseen accidents, several studies have been carried out involving sensors embedded in bicycles and computing performed on Internet of Things (IoT) platforms. However, the sensor’s accuracy in determining the bicycle’s position is low and as a result, the system can send false alarms at high speed. The purpose of this research is to implement the Madgwick AHRS algorithm and Kalman Filter to increase the performance of accidents detection for bicycle athletes. A web server hosting is deployed to store the GPS position results in a map that is provided by Google Maps API. The track of the bicycle race and the position of the bicycle can be determined in this web server. The results of this study show that with the implementation of Madgwick AHRS and Kalman Filter, the measurement of angle estimation is less noisy, with a MAPE value of 15.84%. As an effect, the false alarm rate of the system in detecting accidents can decrease from 100% to 42.86%.