실내 음향 채널의 상관 대역폭 변화를 기반으로 한 침입자 탐지

Abstract

Traditional security systems include which is video camera continuously recode and transmit for surveillance and system requiring password for preventing entry to unauthorized persons. Actually those systems have weakness such as blind spot or danger of steal. Using of acoustic processing for detection applications has been researched in recent years. Numerous acoustic detection systems are available but it has not been completely probed. Because It depend on the sound that the intruder. It is vulnerable in noisy environments or that intruder make little noise. So this thesis focus on intruder detection in indoor environment, which is based on acoustic channel characteristics. Basic concept of this paper is to compute coherence bandwidth from measured room impulse response. Coherence bandwidth is influenced to channel delay profiles. Intruders give rise to room channel impulse response change inside room. This variation causes change of channel delay profile. Experiments are designed to measure mutation of room impulse response according to some cases using only a single microphone data. One is existence or nonexistence of intruder and another is distance which changing between transmitter and microphone intruder in room and the other is background noise. Experiments have been performed to present the effectiveness of the proposed method.