CN105738862A - Behind-the-wall human movement orientation detection method based on dynamic time warping - Google Patents

Behind-the-wall human movement orientation detection method based on dynamic time warping Download PDF

Info

Publication number
CN105738862A
CN105738862A CN201610040965.2A CN201610040965A CN105738862A CN 105738862 A CN105738862 A CN 105738862A CN 201610040965 A CN201610040965 A CN 201610040965A CN 105738862 A CN105738862 A CN 105738862A
Authority
CN
China
Prior art keywords
signal
step
wall
waveform
similarity
Prior art date
Application number
CN201610040965.2A
Other languages
Chinese (zh)
Other versions
CN105738862B (en
Inventor
史治国
张志浩
陈积明
程鹏
孙优贤
Original Assignee
浙江大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 浙江大学 filed Critical 浙江大学
Priority to CN201610040965.2A priority Critical patent/CN105738862B/en
Publication of CN105738862A publication Critical patent/CN105738862A/en
Application granted granted Critical
Publication of CN105738862B publication Critical patent/CN105738862B/en

Links

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01SRADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S3/00Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received
    • G01S3/02Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received using radio waves
    • G01S3/14Systems for determining direction or deviation from predetermined direction
    • G01S3/16Systems for determining direction or deviation from predetermined direction using amplitude comparison of signals derived sequentially from receiving antennas or antenna systems having differently-oriented directivity characteristics or from an antenna system having periodically-varied orientation of directivity characteristic
    • G01S3/22Systems for determining direction or deviation from predetermined direction using amplitude comparison of signals derived sequentially from receiving antennas or antenna systems having differently-oriented directivity characteristics or from an antenna system having periodically-varied orientation of directivity characteristic derived from different combinations of signals from separate antennas, e.g. comparing sum with difference

Abstract

The invention discloses a behind-the-wall human movement orientation detection method based on dynamic time warping. The method comprises the following steps: first, receiving a pre-coded signal waveform from a transmitter through a receiver; second, dividing received signals into signal segments, performing short-time Fourier transform on each signal segment to get a transformation matrix, and calculating the variance vector of the transformation matrix; third, segmenting the variance vector to isolate useful signals; and finally, using a dynamic time warping algorithm to calculate the similarity between the isolated signals and different movement orientation standard signal waveforms, and judging the current behind-the-wall human movement orientation. According to the invention, through the dynamic time warping algorithm, identification and detection of behind-the-wall human movement orientation is realized, and a variety of different movement orientations can be identified.

Description

基于动态时间规整的隔墙人体运动朝向检测方法 Partition body motion based on dynamic time warping toward the detection method

技术领域 FIELD

[0001] 本发明涉及一种隔墙人体运动朝向检测方法,更具体地说是一种基于动态时间规整的隔墙人体运动朝向检测方法。 [0001] The present invention relates to a partition wall toward the body-movement detecting method, and more particularly to a dynamic time warping toward the partition wall body-movement detecting method.

背景技术 Background technique

[0002] -般视距内的人体检测,可以使用诸如红外、摄像机等光电设备来进行检测。 [0002] - like the human body detection in the line of sight can be used, such as infrared cameras and other optical equipment to detect. 这些技术常见于艺术馆和银行的入侵检测中。 These techniques common in intrusion detection Museum of Art and the bank's. 但是这些技术有很大的局限性,无法胜任对于石木质、混凝土等非透明介质墙体(或遮蔽物)后方物体的检测,所以采用的检测技术需具有透视效果。 However, these techniques have significant limitations, can not be qualified for the detection of objects behind the non-transparent medium stone wall wood, concrete, etc. (or screen), the detection technique employed needs to have the perspective effect. 目前具有透视效果的检测技术常见有基于X射线和超声波回波等方式,可是这几种透视技术都不能很好地适应目前对于穿墙人体检测的需求。 Currently it has a perspective effect common with X-ray detection technology and ultrasonic echo, etc. based, but these types of perspective technologies are not well adapted to the current demand for human detection through walls. X射线属于高能量射线,虽然能够穿透墙体,但是对人体有很大的伤害;而超声波回波对分层的介质有比较大的衰减。 X-rays are high-energy rays, even though it can penetrate walls, but there is great harm to the human body; and a relatively large ultrasonic echo attenuation of layered media. 综上所述,采用对墙体有良好穿透性、对人体伤害可以忽略不计的特定频率电磁波作为隔墙人体运动检测的发射信号具有很好的可行性。 In summary, using a good penetration of the wall, the human body is negligible damage to the particular frequency of electromagnetic waves as a transmission signal of the motion detection of the body wall has a good feasibility. 电磁波作为发射信号,可穿透木门、混凝土墙等非金属介质,实现对墙后运动目标的探测。 An electromagnetic wave as a transmission signal, penetrate wooden, concrete walls and other non-metallic medium, to achieve detection of the target wall motion.

[0003] 在防暴和紧急救援等特殊行动中,能否有效探测出房间内或墙壁后的人体运动信息将对作战和救援产生重大的影响,可以大幅度地减少伤亡人数。 [0003] In a special anti-riot operations and emergency rescue, the ability to effectively detect human motion in the room or the wall after the rescue operations and the information will have a significant impact, can significantly reduce the number of casualties. 因此,能够对墙壁、木门等非金属、透明介质后方物体的检测技术受到了越来越多的关注。 Accordingly, it is possible for walls, doors and other non-metallic, detection of objects behind the transparent medium has been more and more attention.

[0004] 传统的穿墙超宽带雷达虽然能够实现隔墙人体运动的检测,但是其占用大量的带宽,发射功率大,且有非常大的天线阵列。 [0004] The conventional entrapment of UWB wall while detecting human motion can be achieved, but it takes a lot of bandwidth, transmit power, and there is a very large antenna array. 而占用带宽小,发射功率低、体积较小的无线通信设备来实现隔墙人体运动检测具有非常大的挑战性,要在强噪声下实现弱目标的检测。 A small bandwidth occupied, low transmit power, a smaller volume of the wireless communication device to implement partition body motion detector having a very big challenge to achieve detection of weak targets in strong noise. 目前关于这种便携式设备实现的隔墙人体运动检测方法和隔墙人体运动朝向检测的技术有待深入研究与探讨。 In-depth study and discussion on the current partition wall body motion detection method and human movement towards the realization of such a portable device detection technology to be.

发明内容 SUMMARY

[0005] 本发明的目的在于提出一种基于动态时间规整的隔墙人体运动朝向检测方法,能够有效地检测出隔墙人体运动的朝向。 [0005] The object of the present invention is to provide a dynamic time warping toward the partition wall body-movement detecting method can effectively detect movement of the body toward the wall.

[0006] 本发明的目的是通过以下技术方案来实现的:一种基于动态时间规整的隔墙人体运动朝向检测方法,该方法包括以下步骤: [0006] The object of the present invention is achieved by the following technical solutions: A dynamic time warping toward the partition wall body-movement detecting method, the method comprising the steps of:

[0007] 步骤1,在墙的一侧布置第一发射机、第二发射机和接收机;首先第一发射机发送原始信号,接收机接收信号后,第二发射机发送同样的原始信号,接收机接收信号;然后通过两次接收的信号计算第二发射机的预编码信号;最后两台发射机同时发射信号,第一发射机发送原始信号,第二发射机发送预编码信号; [0007] Step 1, in a side wall of the first transmitter arrangement, a second transmitter and receiver; first transmitter transmits a first original signal, the receiver receives the signal, the second transmitter transmits the same original signal, a receiver receiving a signal; signal is then received by two precoded signal calculating a second transmitter; last two transmitters simultaneously transmit signals, the first transmitter transmits the original signal, the second transmitter transmits the pre-coded signal;

[0008] 步骤2,接收机接收到两台发射机同时发送的叠加后的信号,并对接收到的信号按时间进行均匀分割; [0008] Step 2, the receiver receives the two transmitters superposed signal simultaneously transmitted and received signals by time division uniform;

[0009] 步骤3,对步骤2分割的每段信号进行短时傅里叶变换,得到一个短时傅里叶变换矩阵AmXn,m代表傅里叶变换(FFT)的频率点个数,n是根据窗函数大小以及重叠数计算得到的每段信号的时间点个数,矩阵中的元素表示在i频率,j时间点的短时傅里叶变换值; [0009] Step 3, Step 2 for each segment divided signal short-time Fourier transform, short time Fourier transform to obtain a matrix AmXn, the frequency point number m represents the Fourier transform (FFT), n is the the number of each segment of the signal point of time and the size of the window functions overlap number calculated, the elements of the matrix represents values ​​STFT frequency i, j time points;

[0010] 步骤4,对步骤3得到的短时傅里叶变换矩阵AmXn进行方差统计,即计算每个时间点上所有频率点对应的短时傅里叶变换值的方差Vj,最终得到这段信号所有时间点上的方差向量VIXn; [0010] Step 4, to short-time Fourier transform matrix AmXn Step 3 was subjected to statistical variance, i.e. all frequencies is calculated at each time point corresponding to the variance of the values ​​Vj short time Fourier transform, this finally obtained VIXn variance vector signal on all time points;

[0011] 步骤5,对方差向量vixn进行分割,具体采用阈值滤波器进行滤波,阈值选择为方差向量vixn的平均值level; [0011] Step 5, the other divided vixn difference vector, particularly using a threshold filter to filter, the threshold value is selected to the average value of the variance of the vector vixn Level;

[level,v;<level [Level, v; <level

[0012]v. (1) [0012] v. (1)

[vp otherwise [Vp otherwise

[0013]步骤6,对滤波后的方差向量波形信号减去阈值滤波器的阈值,分离出有用的信号波形; [0013] Step 6, the variance of a vector of the filtered waveform signal by subtracting the thresholds filter, isolation to a useful signal waveform;

[0014] 步骤7,分别根据步骤1-6得到隔墙人体不同运动朝向的分离后的信号波形,每个运动朝向训练多次,采用动态时间规整算法计算同一运动朝向的任意两个分离后的信号波形的相似度similiarityi,比较计算得到的多个相似度,选取最大相似度max(similarity) 所对应的分离波形作为该运动朝向的标准信号波形; [0014] Step 7, Step 1-6 were obtained after the separation of the signal waveform of various body motion towards the partition wall according to the movement toward each training times, in arbitrary two separate movement toward the same dynamic time warping algorithm of similiarityi signal waveform similarity, compares the calculated degrees of similarity obtained by selecting the maximum similarity separation waveform max (similarity) corresponding to the signal waveform of the movement as the standard orientation;

[0015]步骤8,在进行隔墙人体运动朝向检测时,根据步骤1-6得到分离后的有用信号波形,并采用动态时间规整算法分别与步骤7得到的隔墙不同运动朝向的标准信号波形计算相似度similiarityi,比较计算得到的多个相似度,选取最大相似度所对应的标准信号波形的运动朝向作为此时刻的隔墙人体运动朝向,对步骤2分割的每段信号重复该步骤,从而给出每个时间段的隔墙人体运动朝向。 [0015] Step 8, the body motion towards the partition wall during detection, steps 1-6 to obtain the useful signal waveform according separated, using standard dynamic time warping algorithm is a signal waveform obtained in step 7 respectively different motion toward the walls of calculating the similarity similiarityi, compares the calculated degrees of similarity obtained, the standard motion selection signal waveform corresponding to the maximum similarity as a partition wall toward the body motion towards this point, this step is repeated for each segment signal dividing step 2, whereby for each given period of body movement toward the wall. 本发明所述的基于动态时间规整的隔墙人体运动朝向检测方法,可以检测出多种不同的隔墙人体运动朝向。 Based on dynamic time warping toward the partition wall body-movement detecting method, it can be detected in many different human motion towards the partition wall of the present invention. 与现有技术相比,本发明具有如下优势: Compared with the prior art, the present invention has the following advantages:

[0016] 1.采用动态时间规整计算信号波形的相似度,相比其他一些计算相似度的方法(比如欧几里得距离、余弦相似度、Jaccard系数等)相比,具有不受波形长度、幅度的限制, 只关心波形的相似性; [0016] 1. Using the similarity calculating dynamic time warping signal waveform, as compared compared to other methods (such as Euclidean distance, cosine similarity, Jaccard coefficient) calculating the similarity, from a waveform having a length, limiting the amplitude of the waveform concerned only similarity;

[0017] 2.可以实现实时检测,根据接收到的信号进行相应的信号处理,并实时给出检测出的运动朝向的结果; [0017] 2. The real-time detection can be achieved, signal processing corresponding to the received signal, and gives results in real time orientation of detected motion;

[0018] 3.可以适应不同的环境以及不同的人体运动,而不用事先针对环境进行相应的改变; [0018] 3. You can adapt to different environments and different human movement, without a corresponding change in advance for the environment;

[0019] 4.检测盲区小,在有效的检测区域都可以实现检测。 [0019] 4. Small blind spot detection, the detection of the effective detection region can be achieved.

附图说明 BRIEF DESCRIPTION

[0020] 图1是发射机和接收机的流程图; [0020] FIG. 1 is a flow diagram of a transmitter and a receiver;

[0021] 图2是基于动态时间规整的隔墙人体运动朝向检测的信号处理流程图; [0021] FIG 2 is a flowchart illustrating the signal processing of dynamic time warping toward the partition body motion detection;

[0022] 图3是平行于墙体运动的标准信号波形; [0022] FIG. 3 is a signal waveform of a standard motion parallel to the wall;

[0023]图4是垂直于墙体运动的标准信号波形。 [0023] FIG 4 is a signal waveform standard movement perpendicular to the wall.

具体实施方式 Detailed ways

[0024]以下结合附图对本发明作进一步详细说明。 [0024] conjunction with the drawings of the present invention will be further described in detail.

[0025] 本发明给出了一种基于动态时间规整的隔墙人体运动朝向检测方法,信号的发送和接收过程如图1所示,所用到的是两台发射机和一台接收机。 [0025] The present invention presents a dynamic time warping toward the partition wall body-movement detecting method, the transmission and reception signals as shown, is used as a transmitter and two receivers 1. 首先,第一发射机发送信号, 接收机接收到信号;其次第二发射机发送与第一发射机同样的信号,接收机接收到信号;然后根据两次接收到的信号,计算出预编码后的信号;最后让两台发射机同时发送信号,接收机接收信号。 Then according to the received two signals, calculate the pre-coding; First, a first transmitter transmits the signal receiver receives the signal; second transmitter transmits a second signal similar to the first transmitter, the receiver receives signals signal; and let the two transmitters simultaneously transmit signals, the signal received by the receiver. 这里第一发射机还是发送原来的信号,而第二发射机则是发送刚刚计算出来的预编码后的信号。 Here the first transmitter or to transmit the original signal, the second signal transmitter is transmitted immediately after the calculated precoding.

[0026] 在上述信号发送与接收的基础上,本发明所述的检测方法,如图2所示,包括以下步骤: [0026] On the basis of the signal transmission and reception, the detection method according to the present invention, shown in Figure 2, comprising the steps of:

[0027] 步骤1,首先让接收机和两台发射机放在墙的一侧运行一段时间,接收机将接收到来自墙后以及墙这边的多种反射信号叠加的信号; [0027] Step 1, let a receiver and a transmitter on the two side walls running for some time, the receiver will receive signals from a plurality of side walls and the walls of the superposed reflected signals;

[0028] 步骤2,对接收到的信号按时间进行均匀分割,将其分割成一段段的小信号,这里具体分割成Is的信号数据 [0028] Step 2, the received signal is split equally in time, which is divided into a section of the small signal, particularly where divided into signal data of Is

[0029]步骤3,对分割后的每段小信号进行短时傅里叶变换(STFT)STFT(t,《 ) =JS(t') « ,得到一个短时傅里叶变换矩阵AmXn,该矩阵的行数m代表了使用多少点的傅里叶变换(FFT),即有多少个频率点;而矩阵的列数n则是根据窗函数大小以及重叠数计算得到的每段小信号的时间点个数。 [0029] Step 3, for each segment of the divided small signal short-time Fourier transform (STFT) STFT (t, ") = JS (t ')«, to obtain a short-time Fourier transform matrix AmXn, the the number m of rows of the matrix represent the number of points used in the Fourier transform (FFT), i.e., the number of frequency points; and n number of columns of the matrix is ​​a small-signal time of each segment is calculated according to the obtained size and the number of overlaps window function The number of points. 所以该变换矩阵不仅与频率有关,而且与时间也有关,矩阵中的元素表示在i频率,j时间点的短时傅里叶变换值; Therefore, not only the frequency of the transform matrix, but also with time-related, matrix elements represent STFT frequency values ​​i, j time points;

[0030] 步骤4,由于隔墙静止与隔墙人体运动的短时傅里叶变换存在显著的区别,采用方差统计的方法来分析短时傅里叶变换的变化趋势。 [0030] Step 4, there are significant differences due to the short time Fourier transform and the stationary partition wall body movement, using statistical methods to analyze the variance tendency of short time Fourier transform. 具体是对每列进行方差统计,即计算每个时间点上所有频率点对应的短时傅里叶变换值的方差vj,它反映了在当前时刻在所有频率点上的波动情况。 Is specific statistical variance for each column, i.e. all frequencies is calculated at each time point corresponding to the variance value vj short time Fourier transform, it reflects fluctuations in the current time at all the frequencies. 计算完每列的方差后可以得到这段信号所有时间点上的方差向量V1xn; After calculating the variance of each column can be obtained on this signal variance vector V1xn all time points;

[0031] 步骤5,对方差向量v1Xn进行分割,具体采用公式(1)所示的阈值滤波器进行滤波, 一般阈值选择为方差向量的平均值level; .一. level,v:< level [0031] Step 5, the other divided v1Xn difference vector, particularly using equation (1) filter to filter a threshold value, the threshold value is generally chosen to be the average of the variance of vector level; a level, v:.. <Level

[0032] v.=<^ (1) 1 v ,otherwise [0032] v. = <^ (1) 1 v, otherwise

[0033] 步骤6,对滤波后的方差向量波形信号统一减去阈值滤波器的阈值,分离出有用的信号波形; [0033] Step 6, the variance of a vector of the filtered waveform signal by subtracting the uniform thresholds filter, isolation to a useful signal waveform;

[0034] 步骤7,分别根据步骤1-6得到隔墙人体不同运动朝向的分离后的信号波形,每个运动朝向训练多次,采用动态时间规整算法计算同一运动朝向的任意两个分离后的信号波形的相似度similiarityi,比较计算得到的多个相似度,选取最大相似度max(similarity) 所对应的分离波形作为该运动朝向的标准信号波形; [0034] Step 7, Step 1-6 were obtained after the separation of the signal waveform of various body motion towards the partition wall according to the movement toward each training times, in arbitrary two separate movement toward the same dynamic time warping algorithm of similiarityi signal waveform similarity, compares the calculated degrees of similarity obtained by selecting the maximum similarity separation waveform max (similarity) corresponding to the signal waveform of the movement as the standard orientation;

[0035] 步骤8,在进行隔墙人体运动朝向检测时,根据步骤1-6得到分离后的有用信号波BSseg,采用动态时间规整算法分别计算与不同的运动朝向标准信号波形Stcy^相似度similiarity:,具体操作如下,为了匹配两个波形序列,定义公式(2)所示的一个匹配路径W,其中m、n分别为Sseg和Stdi的长度,Wj包含Sseg和Stdi的索引,通过公式(3)计算分割信号与标准信号的最小路径距离DTWUse^Stck)。 [0035] Step 8, the body motion towards the partition wall during detection, steps 1-6 to obtain the useful signal wave according BSseg separated using dynamic time warping algorithm to calculate the motion direction different from the standard waveform similarity Stcy ^ similiarity :, as follows, to match the sequence of two waveforms, shown in the definition of formula (2) is a matching path W, where m, n are the length and Stdi of Sseg, comprising Sseg of Wj and Stdi index by the equation (3 minimum path) calculated dividing the signal from the standard signal DTWUse ^ Stck). 比较计算得到的多个相似度,选取最大相似度max(simi1arity)所对应的运动朝向作为此时刻隔墙人体运动的朝向; Comparing the calculated degrees of similarity obtained, selecting a maximum similarity max (simi1arity) corresponding to this point as the movement toward the partition wall towards the body movement;

[0036] ff=wi,W2,. . .,wkmax(m,n) <K<m+n-1 (2) (3) [0036] ff = wi, W2 ,..., Wkmax (m, n) <K <m + n-1 (2) (3)

[0038] 本发明所述检测方法采取的技术方案是:预先提取不同隔墙人体运动朝向的标准信号波形,其次将接收到的信号分割成一段段信号,对每段信号进行短时傅里叶变换得到变换矩阵,并计算变换矩阵的方差向量;然后对方差向量进行分割,分离出有用信号,最后采用动态时间规整算法计算分离波形与提取的标准信号波形的相似度,选取最大相似度所对应的朝向作为此时刻的运动朝向。 [0038] The detection method of the present invention is adopted technical solution is: extracting predetermined standard signal waveforms at different body movement toward the wall, followed by the received signal into a section of the signal, the signal short-time Fourier each segment transformation matrix transform, and calculating the variance of a vector of the transformation matrix; and variance vector to separate the useful signal, and finally isolation waveform similarity is calculated with the standard signal waveform extracted by dynamic time warping algorithm selected corresponding to the maximum similarity as movement towards this point orientation.

[0039] 本发明中的标准信号波形是事先多次进行不同隔墙人体运动朝向实验得出的,该标准信号波形具有一般性、可靠性,可以适应不同的环境以及不同的人体运动。 [0039] The reference signal waveform a plurality of times prior to the present invention is different body motion towards the partition wall experimentally derived, the signal waveform having the general standard, reliability, can adapt to different environments and different body movement.

[0040] 本发明采用带宽小、发射功率低的发射机即可实现隔墙人体运动朝向检测,并可保证检测精度。 [0040] The present invention uses a small bandwidth, low transmit power of the transmitter can be realized wall toward human motion detection, the detection accuracy can be guaranteed. 相比于传统穿墙超宽带雷达那样占用大量的带宽、高发射功率及非常大的天线阵列,本发明具有显著优势。 Compared to the conventional ultra-wideband radar as a wall through a lot of bandwidth, high power and a very large transmit antenna array, the present invention has significant advantages.

[0041 ] 实施例 [0041] Example

[0042] 将两台发射机和一台接收机布置在墙的一侧,运动人体在墙的另一侧随意地行走。 [0042] The two transmitters and one receiver are arranged on one side of the wall, the body motion freely walk on the other side of the wall. 两台发射机和接收机在同一水平面上等距排列,且与墙面距离相等。 Both transmitter and receiver are arranged equidistantly on the same horizontal plane, and is equal to the distance from the wall. 实验的墙体为25cm 厚的混凝土墙,其衰减为20dB。 Experiments wall of 25cm thick concrete walls, its attenuation is 20dB. 发射机的带宽为1MHz,发射功率为100mW,发射频率为2.4GHz,包含3个定向天线。 The bandwidth of the transmitter is 1MHz, the transmit power of 100mW, the emission frequency of 2.4GHz, comprising three directional antennas. 运动人体在墙的另一侧行走,主要是两种典型的朝向运动,1)平行于墙面行走和2)垂直墙面行走。 Movement of human walking on the other side of the wall, mainly two typical movement towards, 1) and running parallel to the wall 2) running vertical wall.

[0043] 图3和图4展示了人体平行于墙体运动和垂直于墙体运动的标准信号波形,从图中可以看出两种运动朝向的标准信号波形有非常明显的区别。 [0043] FIG. 3 and FIG. 4 shows the standard waveform and the body motion parallel to the walls perpendicular to the movement of the wall, it can be seen from the figure the two movements toward the standard waveform has a very clear distinction. 实际实验接收到的信号波形经一系列的信号处理后,与这两种标准信号波形相匹配,可以准确可靠地确定当前时刻的隔墙人体运动朝向。 Actual experiment received signal waveform after a series of signal processing, to match the two standard waveforms, the partition wall can be accurately and reliably determining the current time toward the body movement.

[0044] 根据本发明方法,对隔墙人体运动朝向的检测率可达90%,相对于传统穿墙超宽带雷达占用大量的带宽、高发射功率,本发明方法在窄带宽和低发射功率的条件下也具有较高的检测精度。 [0044] The method of the present invention, the detection rate of body movement toward the wall up to 90%, relative to the conventional through-wall of UWB lot of bandwidth, high transmission power, the method of the present invention in a narrow bandwidth and low transmission power under conditions having high detection accuracy.

Claims (1)

1. 一种基于动态时间规整的隔墙人体运动朝向检测方法,其特征在于,该方法包括以下步骤: 步骤1,在墙的一侧布置第一发射机、第二发射机和接收机;首先第一发射机发送原始信号,接收机接收信号后,第二发射机发送同样的原始信号,接收机接收信号;然后通过两次接收的信号计算第二发射机的预编码信号;最后两台发射机同时发射信号,第一发射机发送原始信号,第二发射机发送预编码信号; 步骤2,接收机接收到两台发射机同时发送的叠加后的信号,并对接收到的信号按时间进行均匀分割; 步骤3,对步骤2分割的每段信号进行短时傅里叶变换,得到一个短时傅里叶变换矩阵AmXn,m代表傅里叶变换(FFT)的频率点个数,n是根据窗函数大小以及重叠数计算得到的每段信号的时间点个数,矩阵中的元素^表示在i频率,j时间点的短时傅里叶变换值; 步骤4, CLAIMS 1. A method for detecting orientation, characterized in that the dynamic time warping partition wall body motion, the method comprising the following steps: Step 1, a first transmitter arranged at one side of the wall, the second transmitter and receiver; First the first transmitter transmits the original signal, the receiver receives the signal, the second transmitter transmits the same original signal, the signal received by the receiver; precoded signal and a second transmitter signal received through two calculated; the last two transmitter machine simultaneously transmitted signals, a first transmitter transmits the original signal, the second signal transmitter transmits the pre-coded; step 2, the receiver receives the two transmitters superposed signal is simultaneously transmitted and received signals by time dividing evenly; step 3, step 2 for each segment divided signal short-time Fourier transform, short time Fourier transform to obtain a matrix AmXn, the frequency point number m represents the Fourier transform (FFT), n is the the number and size of the time window functions overlapping each segment of a signal obtained by calculating the number of matrix elements in STFT ^ represents frequency values ​​of i, j time points; step 4, 对步骤3得到的短时傅里叶变换矩阵AmXn进行方差统计,即计算每个时间点上所有频率点对应的短时傅里叶变换值的方差Vj,最终得到这段信号所有时间点上的方差向量VlXn; 步骤5,对方差向量V1Xn进行分割,具体采用阈值滤波器进行滤波,阈值选择为方差向量vixn的平均值level; On the short time Fourier transform matrix AmXn Step 3 was subjected to statistical variance, i.e., short time Fourier transform calculation values ​​of all the frequency points corresponding to each point in time Vj of the variance, to give the final period of the signals of all time points variance vector VlXn; step 5, the other divided V1Xn difference vector, particularly using a threshold filter to filter, the threshold value is selected to the average value of the variance of the vector vixn Level;
Figure CN105738862AC00021
(1) 步骤6,对滤波后的方差向量波形信号减去阈值滤波器的阈值,分离出有用的信号波形; 步骤7,分别根据步骤1-6得到隔墙人体不同运动朝向的分离后的信号波形,每个运动朝向训练多次,采用动态时间规整算法计算同一运动朝向的任意两个分离后的信号波形的相似度similiarityi,比较计算得到的多个相似度,选取最大相似度max(similarity)所对应的分离波形作为该运动朝向的标准信号波形; 步骤8,在进行隔墙人体运动朝向检测时,根据步骤1-6得到分离后的有用信号波形,并采用动态时间规整算法分别与步骤7得到的隔墙不同运动朝向的标准信号波形计算相似度similiarityi,比较计算得到的多个相似度,选取最大相似度所对应的标准信号波形的运动朝向作为此时刻的隔墙人体运动朝向,对步骤2分割的每段信号重复该步骤,从而给出每个时间段的隔墙人体运动 (1) Step 6, the variance of a vector of the filtered waveform signal by subtracting the thresholds filter, isolation to a useful signal waveform; step 7, steps 1-6 were obtained signal separation wall toward the body motion according to different waveform, each moving toward the training times, dynamic time warping algorithm to calculate the similarity similiarityi same movement toward the signal waveform after the separation of any two, compares the calculated degrees of similarity obtained, selecting a maximum similarity max (similarity) isolated waveform corresponding to the moving direction as a standard waveform; step 8, the body wall during movement towards the detector, useful according to the signal waveform obtained after the separation step 1-6, and the use of dynamic time warping algorithm steps 7 respectively the resulting movement towards the wall of different standard signal waveform similarity is calculated similiarityi, compares the calculated degrees of similarity obtained, the standard motion selection signal waveform corresponding to the maximum similarity as a partition wall toward the body motion toward this point, step 2 each segment of the signal division step is repeated to give the walls of each body movement period 向。 To.
CN201610040965.2A 2016-01-21 2016-01-21 Partition body motion based on dynamic time warping toward the detection method CN105738862B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610040965.2A CN105738862B (en) 2016-01-21 2016-01-21 Partition body motion based on dynamic time warping toward the detection method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610040965.2A CN105738862B (en) 2016-01-21 2016-01-21 Partition body motion based on dynamic time warping toward the detection method

Publications (2)

Publication Number Publication Date
CN105738862A true CN105738862A (en) 2016-07-06
CN105738862B CN105738862B (en) 2018-04-20

Family

ID=56246543

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610040965.2A CN105738862B (en) 2016-01-21 2016-01-21 Partition body motion based on dynamic time warping toward the detection method

Country Status (1)

Country Link
CN (1) CN105738862B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5457394A (en) * 1993-04-12 1995-10-10 The Regents Of The University Of California Impulse radar studfinder
CN202583456U (en) * 2012-04-28 2012-12-05 电子科技大学 Building perspective detection device based on hybrid waveforms
CN104820246A (en) * 2015-04-24 2015-08-05 芜湖航飞科技股份有限公司 Through-the-wall radar human body detecting device
CN105137423A (en) * 2015-09-30 2015-12-09 武汉大学 Real-time detection and separation method of multiple moving objects by through-the-wall radar

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5457394A (en) * 1993-04-12 1995-10-10 The Regents Of The University Of California Impulse radar studfinder
US5512834A (en) * 1993-05-07 1996-04-30 The Regents Of The University Of California Homodyne impulse radar hidden object locator
CN202583456U (en) * 2012-04-28 2012-12-05 电子科技大学 Building perspective detection device based on hybrid waveforms
CN104820246A (en) * 2015-04-24 2015-08-05 芜湖航飞科技股份有限公司 Through-the-wall radar human body detecting device
CN105137423A (en) * 2015-09-30 2015-12-09 武汉大学 Real-time detection and separation method of multiple moving objects by through-the-wall radar

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
CH SENG等: ""Image segmentation for through-the-wall radar target detection"", 《IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS》 *
JING LI等: ""Through-wall detection of human being’s movement by UWB radar"", 《IEEE GEOSCIENCE AND REMOTE SENSING LETTERS》 *
S S RAM等: ""Through-wall tracking of human movers using joint doppler and array processing"", 《IEEE GEOSCIENCE AND REMOTE SENSING LETTERS》 *

Also Published As

Publication number Publication date
CN105738862B (en) 2018-04-20

Similar Documents

Publication Publication Date Title
Cuomo et al. Ultrawide-band coherent processing
Adib et al. See through walls with WiFi!
EP2681586B1 (en) Systems, methods, apparatus, and computer-readable media for source localization using audible sound and ultrasound
US8625846B2 (en) Object and movement detection
Zhang et al. Trilinear decomposition-based transmit angle and receive angle estimation for multiple-input multiple-output radar
US8077072B2 (en) Static RF imaging for inside walls of a premises
US9329728B2 (en) User interfaces
Liang et al. Opportunistic sensing in wireless sensor networks: theory and application
WO2007010460A3 (en) Apparatus for the detection of heart activity
US3716823A (en) Polychromatic sonar object identification system
Moore et al. Beamwidth control and angular target detection in an echolocating bottlenose dolphin (Tursiops truncatus)
US7123546B2 (en) Processing technique for forward looking sonar
US9264806B2 (en) Apparatus and method for tracking locations of plurality of sound sources
Li et al. A novel method for respiration-like clutter cancellation in life detection by dual-frequency IR-UWB radar
US20130308790A1 (en) Methods and systems for doppler recognition aided method (dream) for source localization and separation
Lapierre et al. New methods for handling the range dependence of the clutter spectrum in non-sidelooking monostatic STAP radars
CN101644774A (en) Method based on linear programming for locating near-field targets and system thereof
CN104569948B (en) Sea clutter lower sub-band adaptive detection method glrt-ltd
US9264799B2 (en) Method and apparatus for acoustic area monitoring by exploiting ultra large scale arrays of microphones
CN101907709B (en) Method for searching and positioning moving human object by through-wall detecting radar (TWDR)
Kotus et al. Detection and localization of selected acoustic events in acoustic field for smart surveillance applications
Han et al. Wideband Gaussian source processing using a linear nested array
CN102866391A (en) Multi-target detection method based on short-time Fourier transform and fractional Fourier transform
US20090003134A1 (en) Method and Apparatus for Improved Active Sonar Using Singular Value Decomposition Filtering
Robinson et al. Topological localization via signals of opportunity

Legal Events

Date Code Title Description
C06 Publication
C10 Entry into substantive examination
GR01