Increasing cyclists safety using intelligent Bicycle Light Based On Artificial Intelligence

Nowadays, there is a lot of effort to increase safety on roads. The main focus of research is on vehicles and their associated functionality. This paper focuses on increasing safety for other cyclist road transport participants. The research is focused on intelligent light for which propose a solution based on a connection between the mobile device and light device. The solution is aimed at brake light and turn signals. We use data gathered from motion sensors to achieve a simple device that can only be mounted on a bicycle and provide lighting functionality. The data was evaluated by artificial intelligence, namely LSTM (Long Short-Term Memory). The model was trained in the prototype phase, where labeled braking data was created. For labeling data hall sensor was used located on the bicycle brake. Additionally, the solution introduces a turn signal light controlled by voice. Through the utilization of an offline voice recognition library on a mobile phone, the directional lights are governed by verbal commands. This paper describes the hardware, communication, and software parts required to build the intelligent light. Experimental analysis has been conducted to collect data in different situations that occur while cycling, such as braking with different intensities or braking when the bicycle is tilted differently. During the testing phase, we evaluate the effectiveness of the trained LSTM model in detecting bicycle braking during the real ride.