An oscillation-based algorithm for reliable vehicle detection with magnetic sensors

Abstract

Vehicle monitoring using a wireless sensor network is considered in this paper, where a new algorithm is proposed for vehicle detection with magnetic sensors. The proposed algorithm is based on processing the magnetic signal and thoroughly analyzing the number/direction of its oscillations. The main feature of the proposed algorithm over the state-of-the-art ones is its capability to detect vehicles with different shapes of signatures, while most state-of-the-art algorithms assume regular shapes of signatures. This makes the algorithm effective with all types of magnetic sensors. The proposed algorithm has been implemented on MICAz sensor motes and tested in real world scenarios. Results show reliability beyond 93% in all samples, and more than 95% in most of them.