Real-Time Terrain Aware Drive Mode Suggestor using Deep Learning

Modern cars are becoming more technologically advanced, yet many drivers are not aware to fully utilize their capabilities. It is identified that the underutilization of drive modes is a significant factor impacting vehicle efficiency and performance. To address this issue, the "Terrain Aware Drive Mode Suggestor" system has been developed. This solution utilizes a combination of computer vision algorithms, including YOLOv5, YOLOv7, Simple CNN, and MobileNetV2. By analyzing real-time video captured by a camera and processed on a Raspberry Pi, the system accurately detects various terrain types, such as dry mud, wet mud, and normal road conditions, as well as the frequency of potholes. This information allows to suggest appropriate drive modes tailored to the detected terrain, ultimately enhancing the driving experience and reducing fuel consumption. Furthermore, the system aims to bridge the gap between advanced automotive technology and its effective utilization by providing drivers with intuitive suggestions for drive modes based on real-time terrain analysis. Through the integration of machine learning models trained on diverse datasets, ensure the robustness and reliability of the terrain detection and drive mode suggestion functionalities. Consequently, the solution prioritizes user comfort and safety by facilitating smooth transitions between drive modes and alerting drivers to potential road hazards, such as potholes, in advance.