スピーカとニューラルネットワークを使用した自動車盗難用セキュリティシステムに関する研究(機械力学,計測,自動制御)

概要

論文の詳細を見る
Various sensors have been introduced as theft prevention of the automobile as well as their commodities. Improvement in the effectiveness of crime prevention can be expected when sensor fusion techniques are used to achieve an intellectual and integrated judgment system. In this paper, we propose a security system that can detect two or more theft events of a particular automobile using a single sensor. We investigated how a loudspeaker can be used as the sensor. Since there are various events involved in the automobile theft, we target generated sounds from (A) opening a lock, (B) opening a hood, (C) opening a trunk, (D) picking, (E) opening and closing a door and (F) switching on an ignition switch of the engine. The generated sounds from the five theft events ((A), (B), (C), (D) and (E)) were measured with the loudspeaker. Next, two peak values were requested by signal conditioning. Neural network was used to provide relation between these values and five events. As a result, this method we suggested was able to distinguish the five events.
2009-06-25