Enhancing Driver Safety through Real-Time Feedback on Driving Behavior: A Deep Learning Approach

This paper presents a driving safety application designed to improve driving behaviors and adherence to road regulations. The application utilizes a combination of features, including Lane and Object Detection, Driveable Area Segmentation, and dB-Level Tracking, to offer recommendations for safer and more fuel-efficient driving. Object Detection is utilized to identify traffic lights, signs, and surrounding vehicles, among other things. Lane Detection tracks the number of lane cuts and changes made by a driver within a specific time frame. dB-Level Tracking is used to monitor the noise levels inside the vehicle, while Driveable Area Segmentation identifies the driveable areas on the road. A driver score will aggregate the data collected from the various features to assess a driver’s safety, behavior, and passenger safety. By providing drivers with a comprehensive evaluation of their driving habits, the application encourages better driving practices and reduces the number of risks posed by motor vehicles. The implementation of this driving safety application has the potential to create a safer driving environment and promote responsible driving practices. The results of this study demonstrate that the application’s combination of features can provide valuable insights into driving behavior, which can be utilized to improve road safety and reduce the number of accidents caused by driver error.