JP2011232812A - Intruder detecting apparatus, intruder detecting method and intruder detecting program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To detect a subject with higher accuracy by means of an independent component analysis method.SOLUTION: The intruder detecting apparatus 20 comprises: plural signal receiving units 21 to 2N for receiving signals; a signal separating unit 31 for generating plural separated signals by removing noises from the plural received signals by means of an independent component analysis method; a similarity calculating unit 33 for calculating similarities between the received signals and the separated signals and for associating the received signals with the separated signals; a storage unit 32 for storing associations between the received signals and the separated signals; a comparison unit (a variation calculating unit) 34 for calculating a variation of a signal for each separated signal; a determining unit 35 for comparing each variation to a threshold value and for determining that a person is detected in the vicinity of the signal receiving unit corresponding to the separated signal whose variation exceeds a threshold value; and a detected signals transmitting unit (an informing unit) 36 for informing a detection result.

Description

本発明は、人の侵入により変動した受信信号を用いて、人の侵入を検知するとともに、どのアンテナでの変動であるかを検知する侵入者検知技術に関する。   The present invention relates to an intruder detection technique for detecting a person's intrusion using a received signal that has fluctuated due to a person's intrusion and detecting at which antenna the fluctuation is.

近年、住宅やオフィス等の屋内への侵入犯罪に対する防犯意識の高まりや、少子高齢化社会に伴う見守りサービスの需要が増えている。これら安心・安全を目的としたサービスでは、人を検知するためのセンサがあり、現在は赤外線や画像による人の検知が主流となっている。赤外線センサは主に２種類あり、１つは送信機と受信機間で赤外線を送受信し、物体によって赤外線がさえぎられて送受信が出来ない場合に人がいることを検知するものである。もう１つは、人や動物が発する赤外線を検出し、人を検知するものである。   In recent years, there has been an increase in crime prevention awareness against intrusion crimes inside houses and offices, and there has been an increasing demand for watching services accompanying an aging society with a declining birthrate. These services for the purpose of safety and security include sensors for detecting people. At present, detection of people by infrared rays or images is the mainstream. There are mainly two types of infrared sensors. One is to transmit / receive infrared rays between a transmitter and a receiver, and to detect that there is a person when the infrared rays are blocked by an object and cannot be transmitted / received. The other is to detect people by detecting infrared rays emitted by people and animals.

画像センサは、カメラが撮影した画像（現画像と前画像）の差分を解析し、その差分値の変化から人を検知する。画像センサは、近年の画像処理や画像認識技術の向上により、人を精度良く検知する手段として注目を集めている。   The image sensor analyzes a difference between images captured by the camera (current image and previous image), and detects a person from a change in the difference value. Image sensors are attracting attention as means for accurately detecting humans due to recent improvements in image processing and image recognition techniques.

また、最近では、電磁波を利用した人の検知方法が研究されている。電磁波センサは、信号源を備えるアクティブ型（信号源と受信装置のセットとなっているもの）と、携帯電話基地局から放出される電磁波など、空間中に既に存在する電磁波を利用したパッシブ型（受信装置のみ）がある。いずれの場合も、アンテナで受信される電圧の変動を観測するものである（特許文献１参照）。   Recently, methods for detecting humans using electromagnetic waves have been studied. The electromagnetic wave sensor is either an active type with a signal source (a set consisting of a signal source and a receiving device) or a passive type using electromagnetic waves that already exist in space, such as electromagnetic waves emitted from mobile phone base stations ( Receiving device only). In either case, the fluctuation of the voltage received by the antenna is observed (see Patent Document 1).

ところで、このような電磁波を利用したセンサの場合、アンテナで受信される電圧の変動を効率的に抽出することが求められる。人の侵入や移動に伴う電圧の変動抽出方として、複数の信号が混在した環境下において、混合前の元信号変動を推定する手法である独立成分分析手法（非特許文献１、２参照）があり、実環境への適用を目的とした装置やプログラムが提案されている（特許文献２〜８参照）。独立成分分析手法の実環境への適用については、混在する音声信号や生体信号の分離実現を目的に検討が進められている。   By the way, in the case of a sensor using such an electromagnetic wave, it is required to efficiently extract fluctuations in the voltage received by the antenna. An independent component analysis method (see Non-Patent Documents 1 and 2), which is a method for estimating the original signal fluctuation before mixing in an environment where a plurality of signals are mixed, is used as a method for extracting voltage fluctuations due to intrusion or movement of a person. There have been proposed devices and programs intended for application to real environments (see Patent Documents 2 to 8). The application of the independent component analysis method to the real environment is being studied for the purpose of realizing separation of mixed audio signals and biological signals.

特開２００９−２２９３１８号公報JP 2009-229318 A 特許第３８８７２４７号公報Japanese Patent No. 3887247 特開２００４−１１０４０４号公報JP 2004-110404 A 特開２００５−９１５６０号公報JP 2005-91560 A 特開２００５−２５８３６３号公報JP 2005-258363 A 特開２００６−３３７８５１号公報JP 2006-337851 A 特開２００６−２４３６６４号公報JP 2006-243664 A 特開２００７‐２０６０３７号公報JP 2007-206037 A

A. Hyvarinen, “Survey on Independent Component Analysis” Neural Computing Surveys, 2, pp.94-128, 1999.A. Hyvarinen, “Survey on Independent Component Analysis” Neural Computing Surveys, 2, pp.94-128, 1999. A. Hyvarinen and E. Oja, “Independent Component Analysis: Algorithms and Applications”, Neural Networks, 13(4-5), pp.411-430, 2000.A. Hyvarinen and E. Oja, “Independent Component Analysis: Algorithms and Applications”, Neural Networks, 13 (4-5), pp.411-430, 2000.

赤外線センサは、検知感度が高いため多用されているが、直進性が強いという特性を有するため、屋内の障害物（壁、柱など）によって死角が生じる場合がある。このため、複数の赤外線センサを設置しなければならず、コストが高くなってしまう。また、夏など外気温が高くなった場合に誤動作する恐れがある。   Infrared sensors are widely used because of their high detection sensitivity, but because they have the property of high straightness, blind spots may be caused by indoor obstacles (walls, pillars, etc.). For this reason, it is necessary to install a plurality of infrared sensors, which increases the cost. In addition, malfunction may occur when the outside air temperature becomes high, such as in summer.

また、画像センサは、屋内を撮影することから、一般住居などのプライバシーにかかわる場所では適用に心理的な抵抗感がある。また、画像センサは、赤外線センサと同様に、壁などに遮られた死角については画像を撮影できない。このため、複数の画像センサを設置しなければならず、コストが高くなってしまう。   In addition, since the image sensor shoots indoors, there is a psychological resistance to application in a place related to privacy such as a general residence. Further, like the infrared sensor, the image sensor cannot capture an image of a blind spot blocked by a wall or the like. For this reason, it is necessary to install a plurality of image sensors, which increases the cost.

電磁波センサは、電磁波を利用するため、利用する周波数によっては、センサの見通し外であっても検知できる利点があるものの、人の移動に伴う受信電圧の変動を効率的に抽出する必要がある。   Since the electromagnetic wave sensor uses electromagnetic waves, depending on the frequency used, there is an advantage that it can be detected even when the sensor is out of line of sight, but it is necessary to efficiently extract fluctuations in the received voltage due to movement of the person.

また、複数信号が混合した信号を観測し、混合前の元信号変動の推定手法である独立成分分析手法について、これまでの検討内容は、推定精度の改善や計算処理の効率化を目的とするものが主であった。独立成分分析手法により得られる推定信号に関しては、観測した複数信号との関係性（あるいは、観測順序と推定順序）が必ずしも保証されるわけではない。   In addition, the independent component analysis method, which is a method for estimating the original signal fluctuation before mixing by observing a mixed signal of multiple signals, is aimed at improving the estimation accuracy and improving the efficiency of the calculation process. The thing was the main. Regarding the estimated signal obtained by the independent component analysis method, the relationship (or the observation order and the estimation order) with the observed multiple signals is not necessarily guaranteed.

本発明は上記を鑑みてなされたものであり、本発明の目的は、独立成分分析手法を利用して、より信頼性の高い侵入者検知装置、侵入者検知方法および侵入者検知プログラムを提供することにある。   The present invention has been made in view of the above, and an object of the present invention is to provide a more reliable intruder detection device, intruder detection method, and intruder detection program using an independent component analysis technique. There is.

上記目的を達成するため、本発明は、侵入者検知装置であって、任意の周波数の信号を受信する複数の信号受信部と、前記複数の信号受信部がそれぞれ受信した複数の受信信号から、独立成分分析手法を用いて、ノイズを除去した複数の分離信号を生成する信号分離部と、各受信信号と各分離信号との間の類似度を算出し、算出した類似度に基づいて受信信号と分離信号とをそれぞれ対応付ける類似度算出部と、前記類似度算出部が算出した受信信号と分離信号との対応付けを記憶する記憶部と、分離信号毎に、対象物の移動に伴う信号の変動量を算出する変動量算出部と、算出した各変動量と前記記憶部に記憶された閾値とを比較し、変動量が閾値を超える分離信号については、当該分離信号に対応する受信信号を受信した信号受信部の付近で人が検知されたと判別する判定部と、前記判定部の検知結果を通知する通知部と、を有する。   In order to achieve the above object, the present invention is an intruder detection device, which includes a plurality of signal reception units that receive signals of an arbitrary frequency, and a plurality of reception signals received by the plurality of signal reception units, respectively. Using an independent component analysis technique, a signal separation unit that generates a plurality of separated signals from which noise has been removed, and a similarity between each received signal and each separated signal are calculated, and the received signal is calculated based on the calculated similarity. A similarity calculation unit that associates the separated signal with each other, a storage unit that stores the correspondence between the received signal calculated by the similarity calculation unit and the separated signal, and a signal associated with the movement of the object for each separated signal The fluctuation amount calculation unit for calculating the fluctuation amount is compared with each calculated fluctuation amount and the threshold value stored in the storage unit, and for a separated signal whose fluctuation amount exceeds the threshold value, a reception signal corresponding to the separation signal is obtained. Attaching the received signal receiver It has in a determination unit that determines that a person has been detected, and a notification unit for notifying the detection result of the determination unit.

また、本発明は、侵入者検知装置であって、任意の周波数の信号を受信する複数の信号受信部と、前記複数の信号受信部がそれぞれ受信した複数の受信信号から、独立成分分析手法を用いて、ノイズを除去した複数の分離信号を生成する信号分離部と、各受信信号と各分離信号との間の類似度を算出し、算出した類似度に基づいて受信信号と分離信号とをそれぞれ対応付ける類似度算出部と、前記類似度算出部が算出した受信信号と分離信号との対応付けを記憶する記憶部と、複数の分離信号の中から、最も信号の変動が少ない分離信号を基準信号として特定し、他の分離信号の各々について前記基準信号との差分である検知信号を生成する検出部と、生成した検知信号毎に、対象物の移動に伴う信号の変動量を算出する変動量算出部と、算出した各変動量と前記記憶部に記憶された閾値とを比較し、変動量が閾値を超える検知信号については、当該検知信号に対応する受信信号を受信した信号受信部の付近で人が検知されたと判別する判定部と、前記判定部の検知結果を通知する通知部と、を有する。   In addition, the present invention is an intruder detection device, comprising: a plurality of signal receiving units that receive signals of an arbitrary frequency; and a plurality of received signals that are received by the plurality of signal receiving units, respectively. A signal separation unit that generates a plurality of separated signals from which noise has been removed, and calculates a similarity between each received signal and each separated signal, and the received signal and the separated signal are calculated based on the calculated similarity. A similarity calculation unit to be associated with each other, a storage unit for storing the correspondence between the received signal calculated by the similarity calculation unit and the separated signal, and a separated signal with the least signal variation among a plurality of separated signals A detection unit that generates a detection signal that is a difference from the reference signal for each of the other separated signals, and a variation that calculates a variation amount of the signal accompanying the movement of the object for each generated detection signal Amount calculator and calculation Each variation amount is compared with the threshold value stored in the storage unit, and a detection signal whose variation amount exceeds the threshold value is detected in the vicinity of the signal receiving unit that has received the reception signal corresponding to the detection signal. A determination unit that determines that the determination has been made; and a notification unit that notifies a detection result of the determination unit.

また、本発明は、侵入者検知装置が行う侵入者検知方法であって、複数の信号受信部で、任意の周波数の信号を受信する受信ステップと、前記受信ステップでそれぞれ受信した複数の受信信号から、独立成分分析手法を用いてノイズを除去した複数の分離信号を生成する信号分離ステップと、各受信信号と各分離信号との間の類似度を算出し、算出した類似度に基づいて受信信号と分離信号とをそれぞれ対応付け、記憶部に記憶する類似度算出ステップと、分離信号毎に、対象物の移動に伴う信号の変動量を算出する変動量算出ステップと、算出した各変動量と前記記憶部に記憶された閾値とを比較し、変動量が閾値を超える分離信号については、当該分離信号に対応する受信信号を受信した信号受信部の付近で人が検知されたと判別する判定ステップと、前記判定ステップの検知結果を通知する通知ステップと、を行う。   Further, the present invention is an intruder detection method performed by an intruder detection device, wherein a plurality of signal reception units receive a signal of an arbitrary frequency, and a plurality of reception signals respectively received in the reception step. From the signal separation step of generating a plurality of separated signals from which noise has been removed using an independent component analysis method, the similarity between each received signal and each separated signal is calculated, and reception is performed based on the calculated similarity. The similarity calculation step for associating the signal with the separation signal and storing them in the storage unit, the variation amount calculation step for calculating the variation amount of the signal accompanying the movement of the object for each separation signal, and the calculated variation amounts And a threshold value stored in the storage unit, and for a separated signal whose fluctuation amount exceeds the threshold value, it is determined that a person has been detected in the vicinity of the signal receiving unit that has received the received signal corresponding to the separated signal. A step, a notification step of notifying the detection result of the determination step, is carried out.

また、本発明は、侵入者検知装置が行う侵入者検知方法であって、複数の信号受信部で、任意の周波数の信号を受信する受信ステップと、前記受信ステップでそれぞれ受信した複数の受信信号から、独立成分分析手法を用いてノイズを除去した複数の分離信号を生成する信号分離ステップと、各受信信号と各分離信号との間の類似度を算出し、算出した類似度に基づいて受信信号と分離信号とをそれぞれ対応付け、記憶部に記憶する類似度算出ステップと、複数の分離信号の中から、最も信号の変動が少ない分離信号を基準信号として特定し、他の分離信号の各々について前記基準信号との差分である検知信号を生成する検知信号生成ステップと、生成した検知信号毎に、対象物の移動に伴う信号の変動量を算出する変動量算出ステップと、算出した各変動量と前記記憶部に記憶された閾値とを比較し、変動量が閾値を超える検知信号については、当該検知信号に対応する受信信号を受信した信号受信部の付近で人が検知されたと判別する判定ステップと、前記判定ステップの検知結果を通知する通知ステップと、を行う。   Further, the present invention is an intruder detection method performed by an intruder detection device, wherein a plurality of signal reception units receive a signal of an arbitrary frequency, and a plurality of reception signals respectively received in the reception step. From the signal separation step of generating a plurality of separated signals from which noise has been removed using an independent component analysis method, the similarity between each received signal and each separated signal is calculated, and reception is performed based on the calculated similarity. The similarity calculation step of associating each signal with the separated signal and storing it in the storage unit, and identifying the separated signal with the least signal fluctuation as a reference signal among the plurality of separated signals, and each of the other separated signals A detection signal generation step of generating a detection signal that is a difference from the reference signal for, a fluctuation amount calculation step of calculating a fluctuation amount of the signal accompanying the movement of the object for each generated detection signal, Each detected fluctuation amount is compared with the threshold value stored in the storage unit, and a detection signal whose fluctuation amount exceeds the threshold value is detected by a person in the vicinity of the signal receiving unit that has received the reception signal corresponding to the detection signal. A determination step of determining that the determination has been made, and a notification step of notifying the detection result of the determination step.

また、本発明は、侵入者検知プログラムであって、コンピュータに、任意の周波数の信号を受信する複数の信号受信手段、前記複数の信号受信手段がそれぞれ受信した複数の受信信号から、独立成分分析手法を用いて、ノイズを除去した複数の分離信号を生成する信号分離手段、各受信信号と各分離信号との間の類似度を算出し、算出した類似度に基づいて受信信号と分離信号とをそれぞれ対応付ける類似度算出手段、前記類似度算出手段が算出した受信信号と分離信号との対応付けを記憶する記憶手段、分離信号毎に、対象物の移動に伴う信号の変動量を算出する変動量算出手段、算出した各変動量と前記記憶手段に記憶された閾値とを比較し、変動量が閾値を超える分離信号については、当該分離信号に対応する受信信号を受信した信号受信手段の付近で人が検知されたと判別する判定手段、および、前記判定手段の検知結果を通知する通知手段、として機能させる。   The present invention is also an intruder detection program, wherein a plurality of signal receiving means for receiving a signal of an arbitrary frequency are received by a computer, and independent component analysis is performed from a plurality of received signals respectively received by the plurality of signal receiving means. Signal separation means for generating a plurality of separated signals from which noise has been removed using a method, calculating a similarity between each received signal and each separated signal, and receiving the received signal and the separated signal based on the calculated similarity Similarity calculating means for associating each of them, storage means for storing the correspondence between the received signal calculated by the similarity calculating means and the separated signal, and fluctuation for calculating the fluctuation amount of the signal accompanying the movement of the object for each separated signal The amount calculation unit compares each calculated variation amount with the threshold value stored in the storage unit, and for a separated signal whose variation amount exceeds the threshold value, a signal that has received a reception signal corresponding to the separation signal. Judging means for judging a person in the vicinity of the signal means is detected, and notification means for notifying the detection result of said determination means, to function as a.

また、本発明は、侵入者検知プログラムであって、コンピュータに、任意の周波数の信号を受信する複数の信号受信手段、前記複数の信号受信手段がそれぞれ受信した複数の受信信号から、独立成分分析手法を用いて、ノイズを除去した複数の分離信号を生成する信号分離手段、各受信信号と各分離信号との間の類似度を算出し、算出した類似度に基づいて受信信号と分離信号とをそれぞれ対応付ける類似度算出手段、前記類似度算出手段が算出した受信信号と分離信号との対応付けを記憶する記憶手段、複数の分離信号の中から、最も信号の変動が少ない分離信号を基準信号として特定し、他の分離信号の各々について前記基準信号との差分である検知信号を生成する検出手段、生成した検知信号毎に、対象物の移動に伴う信号の変動量を算出する変動量算出手段、算出した各変動量と前記記憶手段に記憶された閾値とを比較し、変動量が閾値を超える検知信号については、当該検知信号に対応する受信信号を受信した信号受信手段の付近で人が検知されたと判別する判定手段、および、前記判定手段の検知結果を通知する通知手段として機能させる。   The present invention is also an intruder detection program, wherein a plurality of signal receiving means for receiving a signal of an arbitrary frequency are received by a computer, and independent component analysis is performed from a plurality of received signals respectively received by the plurality of signal receiving means. Signal separation means for generating a plurality of separated signals from which noise has been removed using a method, calculating a similarity between each received signal and each separated signal, and receiving the received signal and the separated signal based on the calculated similarity A similarity calculation means for associating each of them, a storage means for storing the correspondence between the received signal calculated by the similarity calculation means and the separated signal, and a separated signal with the least signal fluctuation among a plurality of separated signals as a reference signal Detecting means for generating a detection signal that is a difference from the reference signal for each of the other separated signals, and for each of the generated detection signals, the fluctuation amount of the signal accompanying the movement of the object The fluctuation amount calculating means to be output, each calculated fluctuation amount and the threshold value stored in the storage means are compared, and for a detection signal whose fluctuation amount exceeds the threshold value, the signal reception that has received the reception signal corresponding to the detection signal A determination unit that determines that a person is detected in the vicinity of the unit, and a notification unit that notifies the detection result of the determination unit.

本発明によれば、独立成分分析手法を利用して、より信頼性の高い侵入者検知装置、侵入者検知方法および侵入者検知プログラムを提供することができる。   According to the present invention, an intruder detection device, an intruder detection method, and an intruder detection program with higher reliability can be provided using an independent component analysis method.

第１の実施形態の侵入者検知装置の構成図である。It is a lineblock diagram of the intruder detection device of a 1st embodiment. 第１の実施形態の処理の流れを示すフロー図である。It is a flowchart which shows the flow of a process of 1st Embodiment. 信号波形の一例を示す図である。It is a figure which shows an example of a signal waveform. 第２の実施形態の侵入者検知装置の構成図である。It is a block diagram of the intruder detection apparatus of 2nd Embodiment. 第２の実施形態の処理の流れを示すフロー図である。It is a flowchart which shows the flow of a process of 2nd Embodiment. 信号波形の一例を示す図である。It is a figure which shows an example of a signal waveform.

＜第１の実施形態＞
図１は、本発明の第１の実施の形態による侵入者検知装置の構成例を示す。図１において、信号源１０と、侵入者検知装置２０と、インターネット等のネットワーク４０と、が示されている。侵入者検知装置２０は、アンテナ（またはセンサ）を備える複数の信号受信部２１〜２Ｎと、信号分離部３１と、記憶部３２と、類似度算出部３３と、比較部（変動量算出部）３４と、判定部３５と、検出信号送出部（通知部）３６と、を有する。 <First Embodiment>
FIG. 1 shows a configuration example of an intruder detection device according to a first embodiment of the present invention. In FIG. 1, a signal source 10, an intruder detection device 20, and a network 40 such as the Internet are shown. The intruder detection device 20 includes a plurality of signal reception units 21 to 2N including an antenna (or sensor), a signal separation unit 31, a storage unit 32, a similarity calculation unit 33, and a comparison unit (variation amount calculation unit). 34, a determination unit 35, and a detection signal transmission unit (notification unit) 36.

図示する侵入者検知装置２０は、Ｎ個（Ｎは２以上の自然数）の信号受信部２１〜２Ｎを備え、各信号受信部は、侵入者検知装置２０が設置された住宅やオフォスの所定の場所にそれぞれ設置されている。   The intruder detection device 20 shown in the figure includes N (N is a natural number of 2 or more) signal reception units 21 to 2N, and each signal reception unit is a predetermined house or office in which the intruder detection device 20 is installed. It is installed in each place.

図１に示した構成例では、信号源１０からの送信信号が人、動物などの対象物の侵入により変動する系を想定している。このとき、アンテナ等で検出された受信信号は、信号受信部２１〜２Ｎを介して信号分離部３１へ送られる。   In the configuration example shown in FIG. 1, a system is assumed in which the transmission signal from the signal source 10 fluctuates due to the intrusion of an object such as a person or an animal. At this time, the reception signal detected by the antenna or the like is sent to the signal separation unit 31 via the signal reception units 21 to 2N.

図示する侵入者検知装置２０（信号分離部３１、記憶部３２、類似度算出部３３、比較部３４、判定部３５、検出信号送出部３６）は、例えば、ＣＰＵと、メモリと、ＨＤＤ等の外部記憶装置と、入力装置と、出力装置とを備えた汎用的なコンピュータシステムを用いることができる。このコンピュータシステムにおいて、ＣＰＵがメモリ上にロードされた侵入者検知装置２０用のプログラムを実行することにより、侵入者検知装置２０の各機能が実現される。また、侵入者検知装置２０用のプログラムは、ハードディスク、フレキシブルディスク、ＣＤ−ＲＯＭ、ＭＯ、ＤＶＤ−ＲＯＭなどのコンピュータ読取り可能な記録媒体に記憶することも、ネットワークを介して配信することもできる。   The intruder detection device 20 (the signal separation unit 31, the storage unit 32, the similarity calculation unit 33, the comparison unit 34, the determination unit 35, and the detection signal transmission unit 36) illustrated in the figure includes, for example, a CPU, a memory, an HDD, and the like. A general-purpose computer system including an external storage device, an input device, and an output device can be used. In this computer system, each function of the intruder detection device 20 is realized by the CPU executing a program for the intruder detection device 20 loaded on the memory. The program for the intruder detection device 20 can be stored in a computer-readable recording medium such as a hard disk, a flexible disk, a CD-ROM, an MO, a DVD-ROM, or can be distributed via a network.

図２は、侵入者検知装置２０の処理の流れを示したフロー図である。図示する処理は、常時、所定の時間間隔で繰り返し行われるものとする。   FIG. 2 is a flowchart showing the processing flow of the intruder detection device 20. It is assumed that the illustrated process is always repeated at a predetermined time interval.

信号受信部２１〜２Ｎの各々は、信号源１０から発出される周波数に同調して信号を受信し、当該受信信号を所定のタイミング（時間間隔）で信号分離部３１へ出力する（Ｓ１１）。信号受信部２１〜２Ｎは、例えば、信号源１０からの周波数が高い場合には、ダウンコンバートするなど取り扱いが容易な周波数に変換したのち、任意のサンプリング間隔でデジタルデータ（時間領域の信号）へ変換し、信号分離部３１へ送信（送出）する。各信号受信部２１〜２Ｎで受信される受信信号は、人などの移動による信号の変動だけでなく、ノイズ（マルチパス干渉や外来ノイズ等）の影響により、歪をともなっていることが想定される。   Each of the signal receiving units 21 to 2N receives a signal in synchronization with the frequency emitted from the signal source 10, and outputs the received signal to the signal separation unit 31 at a predetermined timing (time interval) (S11). For example, when the frequency from the signal source 10 is high, the signal receiving units 21 to 2N convert the frequency to a frequency that is easy to handle, such as down-conversion, and then to digital data (time domain signal) at an arbitrary sampling interval. The signal is converted and transmitted (sent) to the signal separation unit 31. The received signals received by the signal receiving units 21 to 2N are assumed to be distorted due to the influence of noise (multipath interference, external noise, etc.) as well as signal fluctuations due to movement of people and the like. The

なお、各信号受信部２１〜２Ｎは、信号分離部３１へ受信信号を送信する際に、当該信号受信部２１〜２Ｎを識別するための識別情報（いずれの信号受信部から受信されたものかを判別するための情報であり、例えばｐ１〜ｐＮなどのＩＤ番号を付与することが考えられる。）を受信信号に付加して送信するものとする。   Each of the signal receiving units 21 to 2N, when transmitting a reception signal to the signal separation unit 31, identifies identification information for identifying the signal receiving units 21 to 2N (from which signal receiving unit is received) For example, ID numbers such as p1 to pN are conceivable.) Are added to the received signal and transmitted.

信号分離部３１は、各信号受信部２１〜２Ｎから送信されたＮ個の受信信号（ノイズによる影響を受けた信号）を入力する。そして、信号分離部３１は、独立成分分析手法を用いて、Ｎ個の受信信号からノイズの影響を除去・分離した、Ｎ個の分離信号（推定信号）を生成する（Ｓ１２）。なお、独立成分分析手法については後述する。   The signal separation unit 31 inputs N reception signals (signals affected by noise) transmitted from the signal reception units 21 to 2N. Then, the signal separation unit 31 generates N separated signals (estimated signals) obtained by removing and separating the influence of noise from the N received signals using an independent component analysis method (S12). The independent component analysis method will be described later.

また、信号分離部３１は、Ｎ個の受信信号と、生成したＮ個の分離信号とを記憶部３２に記憶する。これにより、記憶部３２には、Ｎ個の受信信号と、Ｎ個の分離信号とが記憶される。   Further, the signal separation unit 31 stores the N reception signals and the generated N separation signals in the storage unit 32. As a result, the storage unit 32 stores N received signals and N separated signals.

類似度算出部３３は、Ｎ個の受信信号およびＮ個の分離信号を記憶部３２から読み出して、各受信信号と各分離信号との間の類似度をそれぞれ算出する。そして、類似度算出部３３は、算出した類似度に基づいて、どの分離信号がどの受信信号に対応するものであるかを１対１に対応付ける（Ｓ１３）。そして、類似度算出部３３は、算出した類似度および分離信号と受信信号との対応付け情報を記憶部３２に記憶する。分離信号と受信信号との対応付け情報には、受信信号に付加された識別情報も含まれる。なお、類似度の算出手順については後述する。   The similarity calculation unit 33 reads N received signals and N separated signals from the storage unit 32, and calculates the similarity between each received signal and each separated signal. Then, the similarity calculation unit 33 associates one-to-one with which received signal corresponds to which received signal based on the calculated similarity (S13). Then, the similarity calculation unit 33 stores the calculated similarity and association information between the separated signal and the received signal in the storage unit 32. The association information between the separated signal and the received signal includes identification information added to the received signal. The similarity calculation procedure will be described later.

比較部３４は、Ｓ１２で信号分離部３１により生成されたＮ個の分離信号の各々について、変動量を算出し、算出した変動量に基づいて分離信号の順位付けを行う（Ｓ１４）。変動量の算出（評価）は、例えば分離信号の振幅の変動幅や、任意の時間単位での積分値、また、微分値等の算出が考えられる。また、比較部３４は、変動量が大きい順に、１番、２番、・・・Ｎ番の順位を設定する。なお、比較部３４は、設定した順位を分離信号と対応付けて記憶部３２に記憶してもよい。   The comparison unit 34 calculates a variation amount for each of the N separated signals generated by the signal separation unit 31 in S12, and ranks the separated signals based on the calculated variation amount (S14). The calculation (evaluation) of the fluctuation amount may be, for example, calculation of the amplitude fluctuation width of the separated signal, an integral value in an arbitrary time unit, a differential value, or the like. Further, the comparison unit 34 sets the order of No. 1, No. 2,... N in descending order of variation. The comparison unit 34 may store the set order in the storage unit 32 in association with the separation signal.

判定部３５は、Ｓ１４で比較部３４が分離信号毎に算出した各変動量と、記憶部３２に記憶された閾値とをそれぞれ比較し、閾値を超える変動量が存在するか否かを判別する（Ｓ１５）。記憶部３２には、人などの侵入・移動による信号の変動であることを判別するための閾値があらかじめ記憶されている。   The determination unit 35 compares each variation amount calculated for each separated signal by the comparison unit 34 in S14 with the threshold value stored in the storage unit 32, and determines whether there is a variation amount exceeding the threshold value. (S15). The storage unit 32 stores in advance a threshold value for determining that the signal changes due to intrusion / movement of a person or the like.

そして、閾値を超える変動量の分離信号については（Ｓ１５：ＹＥＳ）、判定部３５は、人などの侵入または移動があったと判別する。そして、判定部３５は、変動量が閾値を超える全ての分離信号の各々について、当該分離信号に対応する受信信号の識別情報（例えば、信号受信部２１〜２Ｎの番号等）を、Ｓ１３で記憶部３２に記憶した対応付け情報から検索する。そして、判定部３５は、人の侵入等があったと判定された各分離信号に対応する受信信号の識別情報と、Ｓ１４で比較部３４が決定した変動量の順位とを含む検知情報を、検出信号送出部３６へ送信・送出するとともに、記憶部３２にログ情報として記憶する。   And about the separation signal of the variation | change_quantity exceeding a threshold value (S15: YES), the determination part 35 discriminate | determines that there existed invasion or movement of a person etc. And the determination part 35 memorize | stores the identification information (for example, the number of the signal receiving parts 21-2N etc.) (for example, the number of the signal receiving parts 21-2N) corresponding to the said separation signal about each of all the separation signals whose fluctuation amount exceeds a threshold value by S13. Search is performed from the association information stored in the unit 32. Then, the determination unit 35 detects detection information including the identification information of the received signal corresponding to each separated signal determined to have entered a person and the rank order of the fluctuation amount determined by the comparison unit 34 in S14. The information is transmitted / transmitted to the signal transmission unit 36 and stored as log information in the storage unit 32.

なお、判定部３５は、検知情報に変動量の順位を含めることなく、受信信号の識別情報を変動量の順位に従って（例えば、１番から順に）検知情報に設定することとしてもよい。   The determination unit 35 may set the identification information of the received signal as the detection information according to the order of the variation amount (for example, in order from the first) without including the variation amount rank in the detection information.

検出信号送出部３６は、判定部３５から送信された検知情報を、例えばインターネットなどのネットワーク４０を介して、警備会社、または侵入者検知装置２０が設置された住宅やオフォスの所有者（例えば携帯電話等）へ通知する（Ｓ１６）。そして、Ｓ１１に戻り、所定の時間間隔で繰り返し図２の処理を行う。   The detection signal transmission unit 36 uses the detection information transmitted from the determination unit 35 via a network 40 such as the Internet, for example, a security company or an owner of an office (for example, a mobile phone) where the intruder detection device 20 is installed. (S16). Then, the process returns to S11, and the process of FIG. 2 is repeatedly performed at predetermined time intervals.

一方、すべての分離信号の変動量が閾値を超えない場合（Ｓ１５：ＮＯ）、判定部３５は、人の侵入がないと判別し、Ｓ１６の処理を行うことなく、Ｓ１１に戻り、所定の時間間隔で繰り返し図２の処理を行う。   On the other hand, when the fluctuation amount of all the separated signals does not exceed the threshold value (S15: NO), the determination unit 35 determines that there is no human intrusion, returns to S11 without performing the process of S16, and performs a predetermined time. The process of FIG. 2 is repeatedly performed at intervals.

図３は、各信号の波形の例を示したものである。図３（０）は、信号源１０が送信する送信信号の波形を表わしている。図３（ａ）は、ある信号受信部の付近において、人の移動により図３（０）の送信信号が変動した元信号の波形を表わしている。図３（ｂ）は、ある信号受信部において受信される受信信号を表わしている。受信信号は、人の移動による変動・影響だけでなく、マルチパス干渉や外来ノイズなどにより図３（ａ）の元信号が歪められている。図３（ｃ）は、図３（ｂ）の受信信号からノイズ成分が除去された分離信号を表わしている。   FIG. 3 shows an example of the waveform of each signal. FIG. 3 (0) represents the waveform of the transmission signal transmitted by the signal source 10. FIG. 3A shows the waveform of the original signal in which the transmission signal in FIG. 3O has fluctuated due to the movement of a person near a certain signal receiving unit. FIG. 3B shows a received signal received by a certain signal receiving unit. In the received signal, the original signal in FIG. 3A is distorted due to multipath interference, external noise, and the like as well as fluctuations and influences due to movement of people. FIG. 3C shows a separated signal obtained by removing noise components from the received signal of FIG.

図３（ｂ）の受信信号では、外来ノイズ等の影響で発生した突発的な変動や、マルチパス干渉等による歪みが発生している。図３（ｃ）の分離信号では、受信信号にある外来ノイズ等の影響による変動成分等が除去され、人の移動による信号の変動成分が明確に現れている。   In the received signal of FIG. 3B, sudden fluctuations caused by the influence of external noise or the like, or distortion due to multipath interference or the like occurs. In the separated signal of FIG. 3C, the fluctuation component due to the influence of the external noise or the like in the reception signal is removed, and the fluctuation component of the signal due to the movement of the person clearly appears.

本実施形態では、ノイズやマルチパス干渉の影響がなく、人の移動による変動のみが表わされた図３（ａ）の元信号を、独立成分分析法を用いて推定し、図３（ｃ）の分離信号として生成する。図２のＳ１４において、受信信号ではなく、外来ノイズ等の影響による変動成分等が除去された分離信号を用いて変動量を算出することで、より高い精度で人などを検知することができる。   In the present embodiment, the original signal in FIG. 3 (a), which is not affected by noise or multipath interference and represents only the fluctuation due to the movement of a person, is estimated using an independent component analysis method, and FIG. ) As a separated signal. In S14 of FIG. 2, a person or the like can be detected with higher accuracy by calculating the fluctuation amount using the separated signal from which the fluctuation component due to the influence of external noise or the like is removed instead of the received signal.

[独立成分分析手法]
次に、Ｓ１２の分離信号の生成に用いられる独立成分分析手法について説明する。独立成分分析手法は、受信した複数の受信信号から外来ノイズ等の影響を受けていない分離信号を推定する手法であって、例えば、非ガウス性の最大化に基づく手法、最尤推定に基づく手法、相互情報量に基づく手法、非線形関数無相関に基づく手法、テンソルに基づく手法など（非特許文献１、２）のうち、少なくとも１つの手法を用いることができる。 [Independent component analysis method]
Next, the independent component analysis method used for generating the separation signal in S12 will be described. The independent component analysis method is a method for estimating a separated signal that is not affected by external noise or the like from a plurality of received signals. For example, a method based on maximization of non-Gaussianity or a method based on maximum likelihood estimation. At least one of the methods based on the mutual information amount, the method based on the non-correlation of the nonlinear function, the method based on the tensor (Non-Patent Documents 1 and 2) can be used.

送信信号に人の移動に伴う変動を受けたＮ個の元信号ｓ（図３（ａ））を式（１）とし、Ｎ個の受信信号ｘ（信号受信部２１〜２Ｎで受信された信号：図３（ｂ））を式（２）とし、Ｎ個の分離信号ｙ（信号分離部３１でノイズ成分を除去した信号：図３（ｃ））を式（３）とする。なお、Ｔは行列の転置を表している。   The N original signals s (FIG. 3 (a)) subjected to fluctuations due to the movement of the person in the transmission signal are expressed by equation (1), and N received signals x (signals received by the signal receiving units 21 to 2N). 3 (b)) is represented by equation (2), and N separated signals y (signals from which noise components have been removed by the signal separation unit 31: FIG. 3 (c)) are represented by equation (3). Note that T represents transposition of the matrix.

ｓ＝（ｓ１、ｓ２、・・・、ｓＮ）^T ・・・・式（１）
ただし、ｓ１＝｛ｓ１−１、ｓ１−２、・・・、ｓ１−ｍ｝、ｓ２＝｛ｓ２−１、ｓ２−２、・・・、ｓ２−ｍ｝、・・・、ｓＮ＝｛ｓＮ−１、ｓＮ−２、・・・、ｓＮ−ｍ｝であり、ｍは元信号を構成する各データのサンプル数を表わす。 s = (s1, s2,..., sN) ^T ... Equation (1)
However, s1 = {s1-1, s1-2, ..., s1-m}, s2 = {s2-1, s2-2, ..., s2-m}, ..., sN = {sN −1, sN−2,..., SN−m}, where m represents the number of samples of each data constituting the original signal.

ｘ＝（ｘ１、ｘ２、・・・、ｘＮ）^T ・・・・式（２）
ただし、ｘ１＝｛ｘ１−１、ｘ１−２、・・・、ｘ１−ｍ｝、ｘ２＝｛ｘ２−１、ｘ２−２、・・・、ｘ２−ｍ｝、・・・、ｘＮ＝｛ｘＮ−１、ｘＮ−２、・・・、ｘＮ−ｍ｝であり、ｍは受信信号を構成する各データのサンプル数を表わす。 x = (x1, x2,..., xN) ^T ... Equation (2)
However, x1 = {x1-1, x1-2, ..., x1-m}, x2 = {x2-1, x2-2, ..., x2-m}, ..., xN = {xN −1, xN−2,..., XN−m}, where m represents the number of samples of each data constituting the received signal.

ｙ＝（ｙ１、ｙ２、・・・、ｙＮ）^T ・・・・式（３）
ただし、ｙ１＝｛ｙ１−１、ｙ１−２、・・・、ｙ１−ｍ｝、ｙ２＝｛ｙ２−１、ｙ２−２、・・・、ｙ２−ｍ｝、・・・、ｙＮ＝｛ｙＮ−１、ｙＮ−２、・・・、ｙＮ−ｍ｝であり、ｍは分離信号を構成する各データのサンプル数を表わす。 y = (y1, y2,..., yN) ^T ... Formula (3)
However, y1 = {y1-1, y1-2, ..., y1-m}, y2 = {y2-1, y2-2, ..., y2-m}, ..., yN = {yN −1, yN−2,..., YN−m}, where m represents the number of samples of each data constituting the separated signal.

この場合に、受信信号ｘと元信号ｓとの関係については式（４）で、分離信号ｙと受信信号ｘとの関係については式（５）で表現することができる。ただし、Ａは未知の行列であって、Ｗは推定対象の行列を表わしている。   In this case, the relationship between the received signal x and the original signal s can be expressed by Equation (4), and the relationship between the separated signal y and the received signal x can be expressed by Equation (5). Here, A is an unknown matrix, and W represents a matrix to be estimated.

ｘ＝Ａｓ ・・・・式（４）
ｙ＝Ｗｘ ・・・・式（５）
なお、式（１）〜式（５）を構成する各変数のうち、ｓ、ｓ１、ｓ２、・・・、ｓＮ、ｘ、ｘ１、ｘ２、・・・、ｘＮ、ｙ、ｙ１、ｙ２、・・・、ｙＮについては、ベクトルである。以降も同様である。 x = As (4)
y = Wx Formula (5)
Of the variables constituting the equations (1) to (5), s, s1, s2,..., SN, x, x1, x2,..., XN, y, y1, y2,. .., yN is a vector. The same applies thereafter.

独立成分分析手法の具体的な処理は、信号分離部３１は、Ｎ個の信号受信部２１〜２Ｎが受信したＮ個の受信信号ｘを入力する。次に、式（５）を用いて、分離信号ｙの各成分が互いに統計的に独立となるよう、非ガウス性の最大化に基づく手法、最尤推定に基づく手法、相互情報量に基づく手法、非線形関数無相関に基づく手法、テンソルに基づく手法等を用いて、行列Ｗを算出する。そして、Ｗ≒Ａ^−１と仮定し、算出した行列Ｗを式（５）に代入して分離信号ｙを算出する。なお、分離信号ｙは元信号ｓに近似するため、ｓ＝Ａ^−１ｘ≒Ｗｘ＝ｙの関係が成立している。 In specific processing of the independent component analysis method, the signal separation unit 31 inputs N received signals x received by the N signal receiving units 21 to 2N. Next, using Equation (5), a method based on maximization of non-Gaussianity, a method based on maximum likelihood estimation, and a method based on mutual information so that each component of the separated signal y is statistically independent from each other The matrix W is calculated using a method based on non-correlation with a nonlinear function, a method based on a tensor, or the like. Then, assuming that W≈A− ¹ , the separation signal y is calculated by substituting the calculated matrix W into Equation (5). Since the separated signal y approximates the original signal s, the relationship of s = A ⁻¹ x≈Wx = y is established.

［類似度の算出手順について］
独立成分分析手法を適用した際には各受信信号と各分離信号との関連性（分離信号の生成順序）が保証されない。すなわち、どの分離信号がどの受信信号に対応するかという「分離信号の順序性」が保証されるわけではない。そこで、受信信号ｘと分離信号ｙとの間の類似度を算出し、算出した類似度を用いて、受信信号ｘに対応する分離信号ｙの順序性を決定する。すなわち、Ｓ１３で算出する類似度を用いて、Ｎ個の分離信号とＮ個の受信信号とを、１対１に対応付けする。 [Similarity calculation procedure]
When the independent component analysis method is applied, the relationship between each received signal and each separated signal (the generation order of the separated signals) is not guaranteed. That is, it is not guaranteed that “the order of the separated signals”, which separated signals correspond to which received signals. Therefore, the similarity between the received signal x and the separated signal y is calculated, and the order of the separated signal y corresponding to the received signal x is determined using the calculated similarity. That is, using the similarity calculated in S13, N separated signals and N received signals are associated one-to-one.

Ｓ１３の類似度の算出法としては、各受信信号と各分離信号との間の相互相関係数を用いることができる。ここで、Ｎ個の受信信号をｘ１＝｛ｘ１−１、ｘ１−２、・・・、ｘ１−ｍ｝、ｘ２＝｛ｘ２−１、ｘ２−２、・・・、ｘ２−ｍ｝、・・・、ｘＮ＝｛ｘＮ−１、ｘＮ−２、・・・、ｘＮ―ｍ｝（ただし、ｍは受信信号を構成する各データのサンプル数）とする。また、Ｎ個の分離信号をｙ１＝｛ｙ１−１、ｙ１−２、・・・、ｙ１−ｍ｝、ｙ２＝｛ｙ２−１、ｙ２−２、・・・、ｙ２−ｍ｝、・・・、ｙＮ＝｛ｙＮ−１、ｙＮ−２、・・・、ｙＮ―ｍ｝（ただし、ｍは分離信号を構成する各データのサンプル数）とする。この場合に、相互相関係数ρは式（６）で定義することができる。   As a method of calculating the similarity in S13, a cross-correlation coefficient between each received signal and each separated signal can be used. Here, N received signals are represented as x1 = {x1-1, x1-2,..., X1-m}, x2 = {x2-1, x2-2,..., X2-m},. .., XN = {xN−1, xN−2,..., XN−m} (where m is the number of samples of each data constituting the received signal). Further, N separated signals are represented by y1 = {y1-1, y1-2,..., Y1-m}, y2 = {y2-1, y2-2,..., Y2-m},. YN = {yN−1, yN−2,..., YN−m} (where m is the number of samples of each data constituting the separated signal). In this case, the cross-correlation coefficient ρ can be defined by Equation (6).

ρ（ｘ、ｙ）＝Ｅ[（ｘｉ―Ｅ[ｘｉ]）・（ｙｉ―Ｅ[ｙｉ]）]／σｘｉσｙｉ
・・・式（６）
ただし、ｉ＝１、２、・・・、Ｎであり、Ｅ[ ]は各データの平均値を示し、σは各データの標準偏差を示す。また、「・」はベクトルの内積演算を表す。式（６）を用いることにより、各受信信号ｘｉと各分離信号ｙｉとの間の類似度をそれぞれ算出することができる。 ρ (x, y) = E [(xi−E [xi]) · (yi−E [yi])] / σxiσyi
... Formula (6)
Here, i = 1, 2,..., N, E [] represents the average value of each data, and σ represents the standard deviation of each data. “·” Represents an inner product operation of vectors. By using Equation (6), the similarity between each received signal xi and each separated signal yi can be calculated.

なお、相互相関係数に代えて、受信信号と分離信号との間の相互相関係数の加工値（例えば、２乗値、逆数など）、受信信号と分離信号との間の距離（ユークリッド距離、ミンコフスキー距離、マハラノビス距離など）、クラスター型分析手法（階層的クラスタリング、ｃ平均法、ｋ平均法、ファジィｃ平均法など）（特許文献８参照）により算出される対象データ間の類似度（＝データ間の距離）、などを１つ以上利用することも考えられる。   Note that, instead of the cross-correlation coefficient, a processed value of the cross-correlation coefficient between the received signal and the separated signal (for example, a square value or an inverse number), a distance between the received signal and the separated signal (Euclidean distance) , Minkowski distance, Mahalanobis distance, etc.), cluster-type analysis methods (hierarchical clustering, c-average method, k-average method, fuzzy c-average method, etc.) (see Patent Document 8) It is also conceivable to use one or more data distances).

このように、Ｎ個の受信信号ｘを前述の式（２）とし、Ｎ個の分離信号ｙを前述の式（３）とした場合に、各受信信号ｘと各分離信号ｙとの間で算出された各類似度をそれぞれ比較し、比較の結果、受信信号ｘに対して最も類似度の高い分離信号を、この受信信号ｘに対応する分離信号として決定する。   As described above, when N received signals x are set to the above-described expression (2) and N pieces of separated signals y are set to the above-described expression (3), between each received signal x and each separated signal y, The calculated similarities are respectively compared, and as a result of the comparison, a separated signal having the highest similarity with respect to the received signal x is determined as a separated signal corresponding to the received signal x.

以上説明した本実施形態では、マルチパス干渉や外来ノイズなどにより歪められた複数の受信信号から、これらの影響を除去した複数の分離信号を生成することで、人などの移動に伴い変動した変動成分をより明確にすることができ、より高い精度で人などの対象物の侵入や移動を検知することができる。   In the present embodiment described above, fluctuations caused by movement of a person or the like are generated by generating a plurality of separated signals from which a plurality of received signals are removed from a plurality of received signals distorted by multipath interference or external noise. The components can be made clearer, and intrusion and movement of an object such as a person can be detected with higher accuracy.

また、本実施形態では、信号源からの送信信号を複数の信号受信部（アンテナ等）で受信し、人の侵入等があったと判定された各分離信号に対応する受信信号を受信した信号受信部の識別情報を取得する。これにより、本実施形態では、どの信号受信部の付近で侵入者がいるのかを、精度よく検知することができる。また、当該識別情報を検知情報として通知することにより、警備会社または住宅やオフォスの所有者は、どの信号受信部（アンテナ等）の付近で侵入者がいるのかを把握することができる。   In this embodiment, a signal reception is performed in which a transmission signal from a signal source is received by a plurality of signal reception units (antennas, etc.) and a reception signal corresponding to each separated signal determined to have been intruded by a person or the like. Get the part identification information. Thereby, in this embodiment, it is possible to accurately detect which signal receiving unit is near the intruder. In addition, by notifying the identification information as detection information, the security company or the owner of the house or office can grasp which signal receiving unit (antenna or the like) is near the intruder.

＜第２の実施形態＞
図４は、本発明の第２の実施の形態による侵入者検知装置の構成例を示す。図４において、信号源１０と、侵入者検知装置２０Ａと、インターネット等のネットワーク４０と、が示されている。侵入者検知装置２０Ａは、アンテナ（またはセンサ）を備える信号受信部２１〜２Ｎと、信号分離部３１と、記憶部３２と、類似度算出部３３と、比較部３４と、判定部３５と、検出信号送出部３６と、検出部３７と、を有する。 <Second Embodiment>
FIG. 4 shows a configuration example of an intruder detection device according to the second embodiment of the present invention. In FIG. 4, a signal source 10, an intruder detection device 20A, and a network 40 such as the Internet are shown. The intruder detection device 20A includes signal receiving units 21 to 2N each including an antenna (or sensor), a signal separation unit 31, a storage unit 32, a similarity calculation unit 33, a comparison unit 34, a determination unit 35, A detection signal transmission unit 36 and a detection unit 37 are provided.

図示する侵入者検知装置２０Ａは、Ｎ個（Ｎは２以上の自然数）の信号受信部２１〜２Ｎを備え、各信号受信部は、侵入者検知装置２０Ａが設置された住宅やオフォスの所定の場所にそれぞれ設置されている。   The intruder detection device 20A shown in the figure includes N (N is a natural number of 2 or more) signal reception units 21 to 2N, and each signal reception unit is a predetermined house or office where the intruder detection device 20A is installed. It is installed in each place.

図４に示した構成例では、信号源１０からの信号が人などの侵入により変動する系を想定している。このとき、アンテナ等で検出された受信信号は、信号受信部２１〜２Ｎを介して信号分離部３１へ送られる。   The configuration example shown in FIG. 4 assumes a system in which the signal from the signal source 10 fluctuates due to the intrusion of a person or the like. At this time, the reception signal detected by the antenna or the like is sent to the signal separation unit 31 via the signal reception units 21 to 2N.

図示する侵入者検知装置２０Ａ（信号分離部３１、記憶部３２、類似度算出部３３、比較部３４、判定部３５、検出信号送出部３６、検出部３７）は、例えば、ＣＰＵと、メモリと、ＨＤＤ等の外部記憶装置と、入力装置と、出力装置とを備えた汎用的なコンピュータシステムを用いることができる。このコンピュータシステムにおいて、ＣＰＵがメモリ上にロードされた侵入者検知装置２０Ａ用のプログラムを実行することにより、侵入者検知装置２０Ａの各機能が実現される。また、侵入者検知装置２０Ａ用のプログラムは、ハードディスク、フレキシブルディスク、ＣＤ−ＲＯＭ、ＭＯ、ＤＶＤ−ＲＯＭなどのコンピュータ読取り可能な記録媒体に記憶することも、ネットワークを介して配信することもできる。   The intruder detection device 20A (the signal separation unit 31, the storage unit 32, the similarity calculation unit 33, the comparison unit 34, the determination unit 35, the detection signal transmission unit 36, and the detection unit 37) illustrated in FIG. A general-purpose computer system including an external storage device such as an HDD, an input device, and an output device can be used. In this computer system, each function of the intruder detection device 20A is realized by the CPU executing a program for the intruder detection device 20A loaded on the memory. Further, the program for the intruder detection device 20A can be stored in a computer-readable recording medium such as a hard disk, a flexible disk, a CD-ROM, an MO, a DVD-ROM, or can be distributed via a network.

図５は、侵入者検知装置２０Ａの処理の流れを示したフロー図である。図示する処理は、常時、所定の時間間隔で繰り返し行われるものとする。   FIG. 5 is a flowchart showing the processing flow of the intruder detection device 20A. It is assumed that the illustrated process is always repeated at a predetermined time interval.

信号受信部２１〜２Ｎの各々は、第１の実施形態の信号受信部（図２：Ｓ１１）と同様に、信号源１０から発出される周波数に同調して信号を受信し、当該受信信号を所定のタイミング（時間間隔）で信号分離部３１へ出力する（Ｓ２１）。また、各信号受信部２１〜２Ｎは、第１の実施形態と同様に、当該信号受信部２１〜２Ｎを識別するための識別情報を受信信号に付加して送信するものとする。   Each of the signal receiving units 21 to 2N receives a signal in synchronization with the frequency emitted from the signal source 10 in the same manner as the signal receiving unit (FIG. 2: S11) of the first embodiment, and receives the received signal. The signal is output to the signal separation unit 31 at a predetermined timing (time interval) (S21). In addition, each signal receiving unit 21 to 2N adds the identification information for identifying the signal receiving unit 21 to 2N to the received signal and transmits the same as in the first embodiment.

信号分離部３１は、第１の実施形態の信号分離部（図２：Ｓ１２）と同様に、各信号受信部２１〜２Ｎから送信されたＮ個の受信信号（ノイズによる影響を受けた信号）を入力し、独立成分分析手法を用いて、Ｎ個の受信信号からノイズの影響を除去・分離した、Ｎ個の分離信号を生成する（Ｓ２２）。そして、信号分離部３１は、Ｎ個の受信信号と、生成したＮ個の分離信号とを記憶部３２に記憶する。これにより、記憶部３２には、Ｎ個の受信信号と、Ｎ個の分離信号とが記憶される。独立成分分析手法については、第１の実施形態と同様であるため、ここでは説明を省略する。   Similarly to the signal separation unit (FIG. 2: S12) of the first embodiment, the signal separation unit 31 includes N reception signals (signals affected by noise) transmitted from the signal reception units 21 to 2N. And using the independent component analysis method, N separated signals are generated by removing and separating the influence of noise from the N received signals (S22). Then, the signal separation unit 31 stores the N received signals and the generated N separated signals in the storage unit 32. As a result, the storage unit 32 stores N received signals and N separated signals. Since the independent component analysis method is the same as that of the first embodiment, the description thereof is omitted here.

類似度算出部３３は、第１の実施形態の類似度算出部（図２：Ｓ１３）と同様に、Ｎ個の受信信号およびＮ個の分離信号を記憶部３２から読み出して、各受信信号と各分離信号との間の類似度をそれぞれ算出する。そして、類似度算出部３３は、算出した類似度に基づいて、どの分離信号がどの受信信号に対応するものであるかを１対１に対応付ける（Ｓ２３）。そして、類似度算出部３３は、算出した類似度および分離信号と受信信号との対応付け情報を記憶部３２に記憶する。分離信号と受信信号との対応付け情報には、受信信号に付加された識別情報も含まれる。類似度の算出手順については、第１の実施形態と同様であるため、ここでは説明を省略する。   Similar to the similarity calculation unit (FIG. 2: S13) of the first embodiment, the similarity calculation unit 33 reads N received signals and N separated signals from the storage unit 32, The similarity between each separated signal is calculated. Then, the similarity calculating unit 33 associates which separated signal corresponds to which received signal on a one-to-one basis based on the calculated similarity (S23). Then, the similarity calculation unit 33 stores the calculated similarity and association information between the separated signal and the received signal in the storage unit 32. The association information between the separated signal and the received signal includes identification information added to the received signal. Since the similarity calculation procedure is the same as that in the first embodiment, the description thereof is omitted here.

検出部３７は、各分離信号から、人の移動に伴う変動分の信号である検知信号を抽出する。例えば、検出部３７は、Ｎ個の分離信号の中から最も信号の変動（振幅等）の少ない分離信号を基準信号として特定する。そして、他のＮ個−１個の分離信号の各々と、基準信号との差をとることによりＮ−１個の検知信号を抽出・生成する。   The detection part 37 extracts the detection signal which is a signal for the fluctuation accompanying a person's movement from each separation signal. For example, the detection unit 37 specifies a separated signal having the smallest signal fluctuation (such as amplitude) as the reference signal from among the N separated signals. Then, N-1 detection signals are extracted and generated by taking a difference between each of the other N-1 separated signals and the reference signal.

図６は、各信号の波形の例を示したものである。図６（０）は、信号源１０が送信する送信信号の波形を表わしている。図６（ａ）は、２つの信号受信部の付近において、人の移動により図６（０）の送信信号が変動した元信号６５と、人の移動の影響をほとんど受けていない元信号６６とを表わしている。図６（ｂ）は、２つの信号受信部において受信される、図６（ａ）の元信号６５、６６がノイズなどにより歪められた受信信号６１、６２を表わしている。図６（ｃ）は、図６（ｂ）の受信信号６１、６２からノイズ成分が除去された分離信号６３、６４を表わしている。左側の分離信号６３は、人が近くを通った信号受信部から入力された分離信号であり、右側の分離信号６４は、人の移動の影響を受けていない（人が近くで移動していない）信号受信部から入力された分離信号である。   FIG. 6 shows an example of the waveform of each signal. FIG. 6 (0) represents the waveform of the transmission signal transmitted from the signal source 10. FIG. 6A shows an original signal 65 in which the transmission signal in FIG. 6 (0) fluctuates due to movement of a person near the two signal receiving units, and an original signal 66 that is hardly affected by movement of the person. Represents. FIG. 6B shows received signals 61 and 62 obtained by distorting the original signals 65 and 66 of FIG. 6A received by two signal receiving units due to noise or the like. FIG. 6C shows separated signals 63 and 64 obtained by removing noise components from the received signals 61 and 62 shown in FIG. The separation signal 63 on the left side is a separation signal input from a signal receiving unit that a person has passed nearby, and the separation signal 64 on the right side is not affected by the movement of the person (the person is not moving nearby). ) Separated signal input from the signal receiver.

検出部３７は、分離信号６４のような信号の変動（振幅等）の最も少ない分離信号を基準信号として特定する。そして、検出部３７は、分離信号６３と分離信号６４（基準信号）との差をとることで、図６（ｄ）に示すような、分離信号６３から人の移動に伴う変動成分のみの検知信号を抽出する。   The detection unit 37 specifies a separated signal such as the separated signal 64 that has the least signal fluctuation (such as amplitude) as a reference signal. And the detection part 37 detects only the fluctuation component accompanying a movement of a person from the separated signal 63 as shown in FIG.6 (d) by taking the difference of the separated signal 63 and the separated signal 64 (reference signal). Extract the signal.

図６に示すように、図６（ｃ）の分離信号では外来ノイズ等の影響で発生した突発的な変動成分等が除去され、図６（ｄ）の検知信号では人の移動による信号の変動成分が明らかとなるとともに、人の動き成分のみを抽出することで、より人の動きの判定が容易となる。すなわち、検知信号を用いることで、より高い精度で人などを検知することができる。   As shown in FIG. 6, the separated signal in FIG. 6 (c) removes a sudden fluctuation component or the like generated due to the influence of external noise or the like, and the detection signal in FIG. 6 (d) changes the signal due to the movement of a person. The components become clear, and by extracting only the human movement component, it becomes easier to determine the human movement. That is, by using the detection signal, a person or the like can be detected with higher accuracy.

なお、外来ノイズ等の少ない環境においては、受信信号と対応する分離信号との差をとることで、検知信号を抽出することも可能である。   In an environment with little external noise or the like, the detection signal can be extracted by taking the difference between the received signal and the corresponding separated signal.

検出部３７は、このように抽出した検知信号を、比較部３４へ送信するとともに、記憶部３２に記憶する。なお、検知信号は、分離信号および受信信号と対応付けて記憶部に記憶されるものとする。   The detection unit 37 transmits the detection signal extracted in this way to the comparison unit 34 and stores it in the storage unit 32. It is assumed that the detection signal is stored in the storage unit in association with the separation signal and the reception signal.

比較部３４は、Ｓ２４で抽出・生成されたＮ個−１個の検出信号の各々について、変動量を算出し、算出した変動量に基づいて検出信号の順位付けを行う（Ｓ２５）。変動量の算出（評価）は、例えば分離信号の振幅の変動幅や、任意の時間単位（サンプリング期間）での積分値、また、微分値等の算出が考えられる。また、比較部３４は、変動量が大きい順に、１番、２番、・・・Ｎ番の順位を設定する。なお、比較部３４は、設定した順位を検出信号と対応付けて記憶部３２に記憶してもよい。   The comparison unit 34 calculates the fluctuation amount for each of the N-1 detection signals extracted and generated in S24, and ranks the detection signals based on the calculated fluctuation amount (S25). The calculation (evaluation) of the fluctuation amount may be, for example, calculation of an amplitude fluctuation width of the separated signal, an integral value in an arbitrary time unit (sampling period), a differential value, or the like. Further, the comparison unit 34 sets the order of No. 1, No. 2,... N in descending order of variation. The comparison unit 34 may store the set order in the storage unit 32 in association with the detection signal.

判定部３５は、Ｓ２５で比較部３４が検出信号毎に算出した各変動量と、記憶部３２に記憶された閾値とをそれぞれ比較し、閾値を超える変動量が存在するか否かを判別する（Ｓ２６）。記憶部３２には、人の侵入・移動による信号の変動であることを判別するための閾値があらかじめ記憶されている。   The determination unit 35 compares each variation amount calculated for each detection signal by the comparison unit 34 in S25 and the threshold value stored in the storage unit 32, and determines whether there is a variation amount exceeding the threshold value. (S26). The storage unit 32 stores in advance a threshold value for determining that the signal changes due to the intrusion / movement of a person.

そして、閾値を超える変動量の検出信号については（Ｓ２６：ＹＥＳ）、判定部３５は、人などの侵入または移動があったと判別する。そして、判定部３５は、変動量が閾値を超える全ての検出信号の各々について、当該検出信号に対応する受信信号の識別情報（例えば、信号受信部２１〜２Ｎの番号等）を、Ｓ２３およびＳ２４で記憶部３２に記憶した対応付け情報から検索する。そして、判定部３５は、人の侵入等があったと判定された各検出信号に対応する受信信号の識別情報と、Ｓ２４で比較部３４が決定した変動量の順位とを含む検知情報を、検出信号送出部３６へ送信・送出するとともに、記憶部３２にログ情報として記憶する。   And about the detection signal of the variation | change_quantity exceeding a threshold value (S26: YES), the determination part 35 discriminate | determines that there existed invasion or movement of a person. Then, the determination unit 35 uses, for each detection signal whose variation amount exceeds the threshold, identification information (for example, the numbers of the signal reception units 21 to 2N) of the reception signal corresponding to the detection signal as S23 and S24. The search is performed from the association information stored in the storage unit 32. Then, the determination unit 35 detects detection information including the identification information of the reception signal corresponding to each detection signal determined to have been intruded by the person and the rank order of the amount of variation determined by the comparison unit 34 in S24. The information is transmitted / transmitted to the signal transmission unit 36 and stored as log information in the storage unit 32.

なお、判定部３５は、検知情報に変動量の順位を含めることなく、受信信号の識別情報を変動量の順位に従って（例えば、１番から順に）検知情報に設定することとしてもよい。   The determination unit 35 may set the identification information of the received signal as the detection information according to the order of the variation amount (for example, in order from the first) without including the variation amount rank in the detection information.

検出信号送出部３６は、判定部３５から送信された検知情報を、例えばインターネットなどのネットワーク４０を介して、警備会社、または侵入者検知装置２０Ａが設置された住宅やオフォスの所有者の（例えば携帯電話等）へ通知する（Ｓ２７）。そして、Ｓ２１に戻り、所定の時間間隔で繰り返し図５の処理を行う。   The detection signal transmission unit 36 uses the detection information transmitted from the determination unit 35 via the network 40 such as the Internet, for example, the security company or the owner of the house or office where the intruder detection device 20A is installed (for example, (S27). Then, the process returns to S21, and the process of FIG. 5 is repeatedly performed at predetermined time intervals.

一方、すべての検知信号の変動量が閾値を超えない場合（Ｓ２６：ＮＯ）、判定部３５は、人の侵入がないと判別し、Ｓ２７の処理を行うことなく、Ｓ２１に戻り、所定の時間間隔で繰り返し図５の処理を行う。   On the other hand, when the fluctuation amount of all the detection signals does not exceed the threshold value (S26: NO), the determination unit 35 determines that no person has entered, returns to S21 without performing the process of S27, and performs a predetermined time. The processing of FIG. 5 is repeatedly performed at intervals.

以上説明した第２の実施形態では、最も信号の変動（振幅等）の少ない分離信号を基準信号として特定し、当該基準信号と他の分離信号との差をとることで、人の移動成分のみを抽出することが可能となり、より人の動きの判定が容易となる。   In the second embodiment described above, the separated signal with the smallest signal fluctuation (amplitude etc.) is specified as the reference signal, and the difference between the reference signal and other separated signals is taken, so that only the human movement component is obtained. Can be extracted, and the determination of the movement of the person becomes easier.

＜第３の実施形態＞
上述した第１および第２の実施形態の侵入者検知装置２０、２０Ａにおいて、記憶部３２に、信号受信部２１〜２Ｎが備える各アンテナ（またはセンサ）について、予め設定した周波数区間単位での周波数特性を予め記録しておくものとする。そして、侵入者検知装置２０、２０Ａは、受信信号の時間応答を、記憶部３２に記憶された周波数特性を用いて補正するデータ補正部を、さらに備える。 <Third Embodiment>
In the intruder detection devices 20 and 20A of the first and second embodiments described above, the frequency in units of frequency sections set in advance for each antenna (or sensor) included in the signal reception units 21 to 2N in the storage unit 32. The characteristics are recorded in advance. The intruder detection devices 20 and 20A further include a data correction unit that corrects the time response of the received signal using the frequency characteristics stored in the storage unit 32.

このように、本実施形態では、データ補正部を備えることにより、Ｎ個の各アンテナの周波数特性が同一でない場合であっても、精度よく人の動きの判定を行うことができる。   As described above, in the present embodiment, by providing the data correction unit, it is possible to accurately determine the movement of a person even when the frequency characteristics of the N antennas are not the same.

なお、本発明は上記実施形態に限定されるものではなく、その要旨の範囲内で数々の変形が可能である。例えば、上記実施形態では、信号受信部の識別情報を検知情報として通知することとしたが（図２：Ｓ１６、図５：Ｓ２７）、識別情報の代わりに信号受信部の設置場所を検知信号として通知することとしてもよい。この場合、侵入者検知装置の記憶部３２には、信号受信部の識別情報と設置場所とが対応付けられた対応テーブルが予め記憶されているものとする。   In addition, this invention is not limited to the said embodiment, Many deformation | transformation are possible within the range of the summary. For example, in the above embodiment, the identification information of the signal receiving unit is notified as detection information (FIG. 2: S16, FIG. 5: S27), but the installation location of the signal receiving unit is used as a detection signal instead of the identification information. It is good also as notifying. In this case, it is assumed that a correspondence table in which the identification information of the signal reception unit and the installation location are associated with each other is stored in advance in the storage unit 32 of the intruder detection device.

１０：信号源
２０、２０Ａ：侵入者検知装置
２１〜２Ｎ ：信号受信部
３１：信号分離部
３２：記憶部
３３：類似度算出部
３４：比較部
３５：判定部
３６：検出信号送出部
４０：ネットワーク DESCRIPTION OF SYMBOLS 10: Signal source 20, 20A: Intruder detection apparatus 21-2N: Signal receiving part 31: Signal separation part 32: Storage part 33: Similarity calculation part 34: Comparison part 35: Determination part 36: Detection signal transmission part 40: network

Claims (8)

任意の周波数の信号を受信する複数の信号受信部と、
前記複数の信号受信部がそれぞれ受信した複数の受信信号から、独立成分分析手法を用いて、ノイズを除去した複数の分離信号を生成する信号分離部と、
各受信信号と各分離信号との間の類似度を算出し、算出した類似度に基づいて受信信号と分離信号とをそれぞれ対応付ける類似度算出部と、
前記類似度算出部が算出した受信信号と分離信号との対応付けを記憶する記憶部と、
分離信号毎に、対象物の移動に伴う信号の変動量を算出する変動量算出部と、
算出した各変動量と前記記憶部に記憶された閾値とを比較し、変動量が閾値を超える分離信号については、当該分離信号に対応する受信信号を受信した信号受信部の付近で人が検知されたと判別する判定部と、
前記判定部の検知結果を通知する通知部と、を有すること
を特徴とする侵入者検知装置。 A plurality of signal receivers for receiving signals of an arbitrary frequency;
A signal separation unit that generates a plurality of separated signals from which noise has been removed using an independent component analysis technique from a plurality of received signals received by the plurality of signal receiving units, respectively,
A similarity calculation unit that calculates a similarity between each received signal and each separated signal, and associates the received signal with the separated signal based on the calculated similarity;
A storage unit for storing the correspondence between the received signal calculated by the similarity calculation unit and the separated signal;
For each separated signal, a fluctuation amount calculation unit that calculates the fluctuation amount of the signal accompanying the movement of the object,
Each calculated fluctuation amount is compared with the threshold value stored in the storage unit, and a separated signal whose fluctuation amount exceeds the threshold value is detected by a person in the vicinity of the signal receiving unit that has received the reception signal corresponding to the separation signal. A determination unit for determining that the
An intruder detection device comprising: a notification unit that notifies a detection result of the determination unit. 任意の周波数の信号を受信する複数の信号受信部と、
前記複数の信号受信部がそれぞれ受信した複数の受信信号から、独立成分分析手法を用いて、ノイズを除去した複数の分離信号を生成する信号分離部と、
各受信信号と各分離信号との間の類似度を算出し、算出した類似度に基づいて受信信号と分離信号とをそれぞれ対応付ける類似度算出部と、
前記類似度算出部が算出した受信信号と分離信号との対応付けを記憶する記憶部と、
複数の分離信号の中から、最も信号の変動が少ない分離信号を基準信号として特定し、他の分離信号の各々について前記基準信号との差分である検知信号を生成する検出部と、
生成した検知信号毎に、対象物の移動に伴う信号の変動量を算出する変動量算出部と、
算出した各変動量と前記記憶部に記憶された閾値とを比較し、変動量が閾値を超える検知信号については、当該検知信号に対応する受信信号を受信した信号受信部の付近で人が検知されたと判別する判定部と、
前記判定部の検知結果を通知する通知部と、を有すること
を特徴とする侵入者検知装置。 A plurality of signal receivers for receiving signals of an arbitrary frequency;
A signal separation unit that generates a plurality of separated signals from which noise has been removed using an independent component analysis technique from a plurality of received signals received by the plurality of signal receiving units, respectively,
A similarity calculation unit that calculates a similarity between each received signal and each separated signal, and associates the received signal with the separated signal based on the calculated similarity;
A storage unit for storing the correspondence between the received signal calculated by the similarity calculation unit and the separated signal;
A detection unit that identifies a separation signal with the least signal fluctuation among a plurality of separation signals as a reference signal, and generates a detection signal that is a difference from the reference signal for each of the other separation signals;
For each generated detection signal, a fluctuation amount calculation unit that calculates the fluctuation amount of the signal accompanying the movement of the object,
Each calculated fluctuation amount is compared with the threshold value stored in the storage unit, and a detection signal whose fluctuation amount exceeds the threshold value is detected by a person in the vicinity of the signal receiving unit that has received the reception signal corresponding to the detection signal. A determination unit for determining that the
An intruder detection device comprising: a notification unit that notifies a detection result of the determination unit. 請求項１または請求項２に記載の侵入者検知装置であって、
前記類似度算出部は、
前記受信信号と前記分離信号間の相互相関係数、
前記受信信号と前記分離信号間の相互相関係数の加工値、
前記受信信号と前記分離信号間のユークリッド距離、
前記受信信号と前記分離信号間のミンコフスキー距離、
前記受信信号と前記分離信号間のマハラノビス距離、
階層的クラスタリングにより算出される前記受信信号と前記分離信号間の類似度、
c平均法により算出される前記受信信号と前記分離信号間の類似度、
ｋ平均法により算出される前記受信信号と前記分離信号間の類似度、
の少なくとも１つを用いて類似度を算出すること
を特徴とする侵入者検知装置。 The intruder detection device according to claim 1 or 2,
The similarity calculation unit includes:
A cross-correlation coefficient between the received signal and the separated signal,
A processed value of a cross-correlation coefficient between the received signal and the separated signal;
Euclidean distance between the received signal and the separated signal,
Minkowski distance between the received signal and the separated signal,
Mahalanobis distance between the received signal and the separated signal,
The similarity between the received signal and the separated signal calculated by hierarchical clustering,
c The similarity between the received signal and the separated signal calculated by the average method,
similarity between the received signal and the separated signal calculated by the k-average method;
An intruder detection device characterized in that the similarity is calculated using at least one of the following. 請求項１から請求項３のいずれ一項に記載の侵入者検知装置であって、
前記記憶部には、所定の周波数区間単位で、前記各信号受信部が備えるアンテナまたはセンサの周波数特性が記録され、
前記受信信号の時間応答を前記周波数特性を用いて補正するデータ補正部を、さらに有すること
を特徴とする侵入者検知装置。 The intruder detection device according to any one of claims 1 to 3,
In the storage unit, the frequency characteristics of the antenna or sensor included in each signal receiving unit is recorded in a predetermined frequency section unit,
An intruder detection device further comprising: a data correction unit that corrects a time response of the received signal using the frequency characteristic. 侵入者検知装置が行う侵入者検知方法であって、
複数の信号受信部で、任意の周波数の信号を受信する受信ステップと、
前記受信ステップでそれぞれ受信した複数の受信信号から、独立成分分析手法を用いてノイズを除去した複数の分離信号を生成する信号分離ステップと、
各受信信号と各分離信号との間の類似度を算出し、算出した類似度に基づいて受信信号と分離信号とをそれぞれ対応付け、記憶部に記憶する類似度算出ステップと、
分離信号毎に、対象物の移動に伴う信号の変動量を算出する変動量算出ステップと、
算出した各変動量と前記記憶部に記憶された閾値とを比較し、変動量が閾値を超える分離信号については、当該分離信号に対応する受信信号を受信した信号受信部の付近で人が検知されたと判別する判定ステップと、
前記判定ステップの検知結果を通知する通知ステップと、を行うこと
を特徴とする侵入者検知方法。 An intruder detection method performed by an intruder detection device,
A reception step of receiving a signal of an arbitrary frequency with a plurality of signal reception units;
A signal separation step of generating a plurality of separated signals from which noise has been removed using an independent component analysis technique from the plurality of received signals respectively received in the reception step;
A similarity calculation step of calculating a similarity between each received signal and each separated signal, associating each received signal with the separated signal based on the calculated similarity, and storing in the storage unit;
A fluctuation amount calculating step for calculating a fluctuation amount of the signal accompanying the movement of the object for each separated signal;
Each calculated fluctuation amount is compared with the threshold value stored in the storage unit, and a separated signal whose fluctuation amount exceeds the threshold value is detected by a person in the vicinity of the signal receiving unit that has received the reception signal corresponding to the separation signal. A determination step for determining that
An intruder detection method comprising: performing a notification step of notifying a detection result of the determination step. 侵入者検知装置が行う侵入者検知方法であって、
複数の信号受信部で、任意の周波数の信号を受信する受信ステップと、
前記受信ステップでそれぞれ受信した複数の受信信号から、独立成分分析手法を用いてノイズを除去した複数の分離信号を生成する信号分離ステップと、
各受信信号と各分離信号との間の類似度を算出し、算出した類似度に基づいて受信信号と分離信号とをそれぞれ対応付け、記憶部に記憶する類似度算出ステップと、
複数の分離信号の中から、最も信号の変動が少ない分離信号を基準信号として特定し、他の分離信号の各々について前記基準信号との差分である検知信号を生成する検知信号生成ステップと、
生成した検知信号毎に、対象物の移動に伴う信号の変動量を算出する変動量算出ステップと、
算出した各変動量と前記記憶部に記憶された閾値とを比較し、変動量が閾値を超える検知信号については、当該検知信号に対応する受信信号を受信した信号受信部の付近で人が検知されたと判別する判定ステップと、
前記判定ステップの検知結果を通知する通知ステップと、を行うこと
を特徴とする侵入者検知方法。 An intruder detection method performed by an intruder detection device,
A reception step of receiving a signal of an arbitrary frequency with a plurality of signal reception units;
A signal separation step of generating a plurality of separated signals from which noise has been removed using an independent component analysis technique from the plurality of received signals respectively received in the reception step;
A similarity calculation step of calculating a similarity between each received signal and each separated signal, associating each received signal with the separated signal based on the calculated similarity, and storing in the storage unit;
A detection signal generation step of identifying a separation signal with the least signal fluctuation among a plurality of separation signals as a reference signal and generating a detection signal that is a difference from the reference signal for each of the other separation signals;
For each generated detection signal, a fluctuation amount calculating step for calculating a fluctuation amount of the signal accompanying the movement of the object,
Each calculated fluctuation amount is compared with the threshold value stored in the storage unit, and a detection signal whose fluctuation amount exceeds the threshold value is detected by a person in the vicinity of the signal receiving unit that has received the reception signal corresponding to the detection signal. A determination step for determining that
An intruder detection method comprising: performing a notification step of notifying a detection result of the determination step. コンピュータに、
任意の周波数の信号を受信する複数の信号受信手段、
前記複数の信号受信手段がそれぞれ受信した複数の受信信号から、独立成分分析手法を用いて、ノイズを除去した複数の分離信号を生成する信号分離手段、
各受信信号と各分離信号との間の類似度を算出し、算出した類似度に基づいて受信信号と分離信号とをそれぞれ対応付ける類似度算出手段、
前記類似度算出手段が算出した受信信号と分離信号との対応付けを記憶する記憶手段、
分離信号毎に、対象物の移動に伴う信号の変動量を算出する変動量算出手段、
算出した各変動量と前記記憶手段に記憶された閾値とを比較し、変動量が閾値を超える分離信号については、当該分離信号に対応する受信信号を受信した信号受信手段の付近で人が検知されたと判別する判定手段、および、
前記判定手段の検知結果を通知する通知手段、
として機能させるための侵入者検知プログラム。 On the computer,
A plurality of signal receiving means for receiving a signal of an arbitrary frequency;
Signal separating means for generating a plurality of separated signals from which noise has been removed using an independent component analysis technique from the plurality of received signals respectively received by the plurality of signal receiving means,
Similarity calculation means for calculating the similarity between each received signal and each separated signal and associating the received signal with the separated signal based on the calculated similarity,
Storage means for storing the correspondence between the received signal calculated by the similarity calculating means and the separated signal;
Fluctuation amount calculating means for calculating the fluctuation amount of the signal accompanying the movement of the object for each separated signal,
Each calculated fluctuation amount is compared with the threshold value stored in the storage means, and a separation signal whose fluctuation amount exceeds the threshold value is detected by a person in the vicinity of the signal receiving means that has received the reception signal corresponding to the separation signal. Determining means for determining that the
Notification means for notifying the detection result of the determination means;
Intruder detection program to function as コンピュータに、
任意の周波数の信号を受信する複数の信号受信手段、
前記複数の信号受信手段がそれぞれ受信した複数の受信信号から、独立成分分析手法を用いて、ノイズを除去した複数の分離信号を生成する信号分離手段、
各受信信号と各分離信号との間の類似度を算出し、算出した類似度に基づいて受信信号と分離信号とをそれぞれ対応付ける類似度算出手段、
前記類似度算出手段が算出した受信信号と分離信号との対応付けを記憶する記憶手段、
複数の分離信号の中から、最も信号の変動が少ない分離信号を基準信号として特定し、他の分離信号の各々について前記基準信号との差分である検知信号を生成する検出手段、
生成した検知信号毎に、対象物の移動に伴う信号の変動量を算出する変動量算出手段、
算出した各変動量と前記記憶手段に記憶された閾値とを比較し、変動量が閾値を超える検知信号については、当該検知信号に対応する受信信号を受信した信号受信手段の付近で人が検知されたと判別する判定手段、および、
前記判定手段の検知結果を通知する通知手段
として機能させるための侵入者検知プログラム。 On the computer,
A plurality of signal receiving means for receiving a signal of an arbitrary frequency;
Signal separating means for generating a plurality of separated signals from which noise has been removed using an independent component analysis technique from the plurality of received signals respectively received by the plurality of signal receiving means,
Similarity calculation means for calculating the similarity between each received signal and each separated signal and associating the received signal with the separated signal based on the calculated similarity,
Storage means for storing the correspondence between the received signal calculated by the similarity calculating means and the separated signal;
Detecting means for identifying a separated signal having the least signal fluctuation among a plurality of separated signals as a reference signal, and generating a detection signal that is a difference from the reference signal for each of the other separated signals;
For each generated detection signal, a fluctuation amount calculating means for calculating a fluctuation amount of the signal accompanying the movement of the object,
Each calculated fluctuation amount is compared with the threshold value stored in the storage means, and a detection signal whose fluctuation amount exceeds the threshold value is detected by a person in the vicinity of the signal receiving means that has received the reception signal corresponding to the detection signal. Determining means for determining that the
An intruder detection program for functioning as notification means for notifying the detection result of the determination means.

Images