Intelligent Weigh-in-Motion System Using Embedded MEMS Sensors for Pavement Monitoring

Monitoring of traffic volume and loads is critical to maintaining the structural integrity of asphalt pavements. However, the random nature of vehicles traveling on pavement, including the speed and location of traffic loads, is rarely discussed, which causes errors in weigh-in-motion (WIM). This study introduces an innovative approach to identify traffic loads, considering vehicle motion by employing several techniques. First, cost-effective micro-electromechanical system (MEMS) sensors, including accelerometers and magnetometers, are used to collect traffic data on asphalt pavement. The sensors are housed in a protective nylon case and are solar powered to ensure continuous operation. In addition, temperature and gravity compensation techniques are used to significantly reduce accelerometer errors. Multiple sensor nodes are installed, each equipped with accelerometers and magnetometers to monitor vehicle position, speed, and axle spacing. The speed measurement radar is used to calculate speed errors within a 5% margin. A finite element method (FEM) is used to analyze the impact of vehicle speed and traffic load position on measured pavement vibration, providing a data source for the machine learning model. Finally, a neural network model is developed to estimate traffic loads based on pavement acceleration, vehicle speed, and load position. The results indicate that the single-axis error is below the permissible limit for the long-term pavement performance (LTPP) at a 95% confidence level. In summary, the proposed technique utilizes intelligent sensor nodes to enhance traffic data collection and enable accurate WIM for pavement monitoring applications.