Wireless Weigh-In-Motion Vehicle Axle Detection Method Using a Single Accelerometer Sensor

This study investigates a wireless Weigh-In-Motion (WIM) system using a single MEMS accelerometer (LIS331DLH) for vehicle axle detection. A prototype, integrating a Libelium Waspmote microcontroller and Xbee 3 radio module, was embedded in concrete pavement to capture vibration data from two trucks traversing at 10, 15, and 20 km/h. The signal processing pipeline included axis-specific normalization (min-max for Z-axis, max-absolute for X/Y), a moving average filter, a 3Hz Butterworth low-pass filter, and peak detection with thresholds and minimum separation criteria based on axle spacing and speed. Data from 99 trials were categorized into four types (Tire-On-Top, Mixed, Non-Tire-on-Top, Anomaly) using quality metrics like peak prominence and noise levels. Results showed that Tire-On-Top datasets achieved the highest accuracy, with clearly distinguishable dual peaks when axles aligned with the sensor. However, validation yielded only 17.2% accuracy overall, attributed to vehicle misalignment, sensor sensitivity, and noise interference. Non-Ideal conditions (e.g., uneven tire contact) produced irregular peaks, leading to false detections. While confirming the feasibility of single-sensor WIM systems for axle detection under controlled alignment, the study highlights limitations in dynamic environments. Future enhancements should incorporate multi-sensor arrays, advanced noise reduction, machine learning models, and real-world testing to improve robustness for structural health monitoring and traffic management applications.