Machine Learning-Based Speed Breaker and Potholes Detection for On-Road Drivers

Good condition road systems are essential for a nation's economic well-being. However, the presence of speed breakers and potholes presents considerable safety hazards due to insufficient signage, limited visibility at night, and driver irresponsibility. Furthermore, these problems can cause significant harm to vehicle suspensions, particularly in automobiles with a lower clearance from the ground. This paper presents a device that utilizes a recurrent neural network (RNN) to identify and classify potholes and speed bumps. The device, equipped with an ESP32, GPS module, monitors and sends the precise position of these dangers to a cloud-based database. The neural network, which has undergone training, makes predictions based on real-time data gathered by the ESP32. A web server, accessible through both desktop and mobile browsers, then displays the resulting output. The device also has a voice module to alert the system with the help of DF player mini which is connected to the speaker. This voice module plays a major role in visually impaired persons' safety. The primary objective of this system is to optimize road safety, boost the condition of vehicle suspension, minimize driver exhaustion, and deliver prompt notifications. Consequently, it offers advantages to both drivers and society as a whole.