Robust vehicle detection algorithm with magnetic sensor

This paper presents a robust real-time vehicle detection algorithm. The proposed algorithm is performed on wireless magnetic sensor node to instrument roadways. The sensor node measures the Earth´s magnetic field disturbed by vehicle body moving within the neighborhood. This is performed in real-time and can be useful for detecting the vehicles consisting of 88.9% yield of motorcycles (small vehicles) and 99.3% yield of road vehicles (cars, pickups, vans, trucks, buses and SUVs) under various places and environment. Especially, the temperature effect resulting in performance can be resolved by the automatic adaptive baseline method. In addition, the algorithm can detect both the passing and the stationary vehicles. Moreover, the automatic check process is combined for checking the errors; the double detection due to low speed and the instant baseline level changing.