Federated Learning for Accurate Detection of Speed Breakers on The Road

摘要

Vehicle drivers get distracted because of the in-temperate use of the speed-breakers on roads. Furthermore, most of the time drivers can not notice the presence of the unmarked speed breaker and loose control of the vehicle. Sometimes it leads to a serious accident and take their lives. There are some solutions regarding this problem those are highly time-consuming as well as error-prone. Besides none of them pay any attention to renovate the data about speed- breakers within a short amount of time. In this paper, we come forward with a system offering highly accurate detection of speed-breakers by applying federated learning. Our system detects speed-breakers and generates a warning to the driver. After the detection of any speed-breaker, feedback is sent to the cloud server. The server aggregates all the pieces of information and gets updated each time a vehicle is passing a speed-breaker. The proposed system outclasses the latest works with which it is compared to in the matter of accuracy.