A Parking Occupancy Detection Algorithm Based on AMR Sensor

Abstract

Recently, with the explosive increase of automobiles in cities, parking problems are serious and even worsen in many cities. This paper proposes a new algorithm for parking occupancy detection based on the use of anisotropic magnetoresistive sensors. Parking occupancy detection is abstracted as binary pattern recognition problem. According to the status of the parking space, the recognition result contains two categories: 1) vacant and 2) occupied. The feature extraction method of the parking magnetic signal is proposed. In addition, the classification criteria are derived based on the distance discriminate analysis method. Eighty-two sensor nodes are deployed on the roadside parking spaces. By running the system for six months, we observed that the accuracy rate of the proposed parking occupancy detection algorithm is better than 98%.