Auto-GeRo: An IOT based Geo Spatial Model for Real-Time Road Condition Detection

This study proposes a novel IoT model, the Smart Auto IoT-based Geo-Spatial Rover (Auto-GeRo), which integrates a multi-sensor fusion approach for detection of road anomalies in real-time. The model proposed here utilizes machine learning techniques to achieve a high precision rate of 0.43548 with Random Forest in pothole detection, thereby leading from existing models in both accuracy and cost-effectiveness. The work presented here follows a multi-sensor fusion approach that integrate a motion processing unit (IMU-MPU 9250) sensor with other sensors to capture 3-axial XYZ data for real-time road surface analysis. The primary objective is to identify and detect different road conditions, specifically distinguishing between flat and rough patches. By applying various machine learning algorithms, the research presented here tries to evaluate and optimize the performance of road surface classification based on sensor data collected during the simulation of proposed Auto-GeRo simulation. While Logistic Regression exhibits high accuracy (0.96038), the focus shifts towards precision for effective pothole detection, leading to the selection of the Random Forest model due to its superior precision (0.43548) in identifying roads with potholes. Thus the research highlights the potential of IoT-enabled multi sensor fusion techniques in road transportation network for better road safety and early maintenance thereby reducing the maintenance cost and vehicle damage.