CN103616729B - A kind of multiple human body target evaluation methods and system based on UWB bioradar - Google Patents

Abstract

The invention discloses a method and system for estimating multiple human targets based on UWB biological radar, which belongs to the technical field of biological radar life detection. In the processing of collected echo data, the invention firstly adopts breathing detection in each channel Algorithm, adaptive cancellation method and target recognition algorithm are used for processing, and then correlation processing is carried out between channels according to the processing results, and finally thresholds are set to remove residual clutter in the channel, and target classification and number estimation are realized at the same time. The method for estimating the number of human targets of the present invention lays a technical foundation for further realizing the biological radar detection and positioning of multiple human targets, and helps to improve the level and ability of biological radar technology to obtain human body information.

Description

A method and system for estimating multiple human targets based on UWB bioradar

technical field

The invention belongs to the technical field of biological radar life detection, and in particular relates to a method and system for estimating multiple human targets based on UWB biological radar.

Background technique

Bio-radar is a special radar technology for the detection and identification of living objects. This technology mainly uses the human body as the detection target, and uses the electromagnetic wave emitted by the radar as the carrier to obtain the information of the human target. Its biggest advantage is the ability to penetrate obstacles such as walls and ruins, so it has broad application prospects in national security, emergency rescue and other fields.

In practical applications, people first pay attention to whether there is someone, that is, whether there is at least one person. Therefore, the current bio-radar technology is basically based on the detection of a single target. With the realization of this technology, people will further pay attention to the number of human targets, that is, how many people exist. For example, search for survivors buried under the ruins after an earthquake, and monitor terrorists in buildings through partition walls during anti-terrorism operations. But so far, there is no way to solve this problem, especially without any prior knowledge.

Contents of the invention

The purpose of the present invention is to provide a method and system for estimating multiple human targets based on UWB biological radar. The method is based on a multi-antenna UWB biological radar system and can accurately estimate the number of human targets without prior knowledge.

The present invention is achieved through the following technical solutions:

A system for estimating multiple human targets based on UWB biological radar, including a radar host, an antenna array, and a computer, wherein the radar host includes a pulse generator, a transmitter, a central controller, a delayer, and several receiving units; The antenna array described above includes 1 transmitting antenna and 3 receiving antennas, the transmitting antenna is connected with the transmitter of the radar host, and each receiving antenna is connected with a receiving unit to form a radar echo data channel; the receiving unit includes a receiver and an analog-to-digital converter connected to the receiver, the output end of each analog-to-digital converter is connected to the central controller, and the central controller interacts with the computer to realize communication;

The pulse generator generates a periodic narrow pulse signal: one part is sent to the transmitter for shaping, and then radiated out through the transmitting antenna; the other part of the narrow pulse signal is sent to the delayer to generate a control signal, and triggers the receiving unit to adjust the distance of the echo signal for selective reception.

The antenna array adopts a linear array arrangement, with the transmitting antenna as the center, the first receiving antenna is arranged adjacent to the transmitting antenna, the second receiving antenna and the third receiving antenna are symmetrically arranged on both sides of the transmitting antenna, and the second receiving antenna The distances between the antenna, the third receiving antenna and the transmitting antenna are all 0.3-1m.

A reinforced power supply isolation design is adopted between the independent echo data channels.

The UWB biological radar adopts the impulse pulse system, the width of the transmission pulse is controlled at the ns level, and the corresponding spectrum width and center frequency are both 500MHz; both the transmitting antenna and the receiving antenna use a butterfly dipole antenna, and are used between the radar host Shielded coaxial cable connection.

A method for estimating multiple human targets based on UWB bioradar, comprising the following steps:

1) Data collection

After sampling several radar echo data channels of the UWB biological radar system, the obtained echo data is sent to the computer for storage and processing;

2) In-channel processing

Firstly, the breathing detection algorithm is used to obtain the human breathing signal for the obtained echo data; secondly, the adaptive cancellation method is used for the detected human breathing signal to suppress the mutual interference of breathing signals between different human targets; finally, through the target recognition algorithm Set the energy threshold to identify the target individual;

3) Inter-channel processing

For the waveform corresponding to the human body target recognition point in each channel identified in step 2), calculate the correlation coefficient r _ij between channels according to formula (1):

${\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; \&Sigma;S</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}{{<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ]</span>}^{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Among them, S _i and S _j represent the waveforms on two channels, where i, j=1, 2, 3..., and i≠j;

Set the threshold A, compare the calculated correlation coefficient r _ij with the threshold A, and distinguish different targets: For the target points in the recognition results of each channel, there must be another channel target point with a correlation coefficient greater than the set The threshold A, otherwise it is regarded as clutter; the target points with the inter-channel correlation coefficient greater than the set threshold A are regarded as coming from the same target;

4) Finally, remove the clutter in each channel, classify the targets, and obtain the number of human targets.

The breathing detection algorithm described in step 2) is specifically:

First, according to the distance information provided by the UWB bio-radar system, the echo data is subjected to sliding average processing on the distance to improve the signal-to-noise ratio of the echo data;

Secondly, remove the static background clutter in the echo data that does not change with time;

Thirdly, using the narrow-band characteristics of human respiration, an adaptive algorithm is used to separate human respiration from noise;

Finally, low-pass filtering is performed according to the slowly changing characteristics of human respiration to obtain the human respiration signal detected by the radar.

Step 2) The self-adaptive cancellation method is to use the waveform of a certain distance point in the echo data processed by the breathing detection algorithm as the input signal, and the breathing signal of the human target before the distance point as the reference signal, using two channels After the signal is subtracted, the error signal is obtained to correct the coefficient of the adaptive filter; if the input signal corresponds to a certain target distance point, the process will cancel the trailing interference of the previous target and output the breathing signal of the target; if the input signal is the previous target The smearing caused by the target respiratory movement is canceled and the output signal is zero; wherein, the adaptive filter is realized by using the LMS algorithm.

The target recognition algorithm described in step 2) is to accumulate the data processed by the adaptive cancellation method in time to obtain a one-dimensional energy-distance distribution. In this distribution, there will be an energy peak at the corresponding distance of the human target. The target can be identified by setting the threshold value of the energy peak to make a binary judgment on whether there is a target.

First, set the energy threshold Th1 and the related threshold Th2;

The energy threshold Th1 is used to judge whether there is a target. Select the energy E _obj of the target point and the energy E _clu of the clutter point. When E _obj /E _clu is greater than the energy threshold Th1, it is judged that the distance point is a target and the output is 1, otherwise output 0;

The correlation threshold Th2 is used to select clutter points, and the point in the recognition window with the largest energy except the target point and the correlation coefficient between the waveform and the target point waveform is less than the correlation threshold Th2 is regarded as a clutter point, and the energy of this point is E _clu .

Both the energy threshold Th1 and the related threshold Th2 are empirical values, and Th1=2, Th2=0.6.

Compared with the prior art, the present invention has the following beneficial technical effects:

The radar antenna array of the present invention is composed of one transmitting antenna and three receiving antennas, and each receiving antenna and its corresponding receiving unit form three independent radar echo data channels. For the same target, the respiratory signals detected by different channels are highly correlated; while for the respiratory signals and clutter of different targets, the correlation between channels is low. Therefore, in the processing of the collected echo data, the present invention first adopts a breathing detection algorithm, an adaptive cancellation method and a target recognition algorithm in each channel to process, and then performs correlation processing between each channel according to the processing results. Finally, the threshold is set to remove the residual clutter in the channel, and at the same time realize the target classification and number estimation. The method for estimating the number of human targets of the present invention lays a technical foundation for further realizing the biological radar detection and positioning of multiple human targets, and helps to improve the level and ability of biological radar technology to obtain human body information.

Furthermore, the UWB biological radar of the present invention adopts an impulse pulse system, and the width of the transmitted pulse is controlled at the ns level, and the corresponding spectrum width and center frequency are both 500 MHz, which can simultaneously ensure the penetrating ability and the detection ability of human breathing.

Description of drawings

Figure 1 Block diagram of multi-antenna UWB bio-radar system;

Figure 2 The flow chart of echo processing in the channel;

Figure 3 UWB bio-radar echo data processed by breathing detection algorithm;

Fig. 4 Schematic diagram of interference suppression algorithm;

Figure 5 Schematic diagram of the target recognition algorithm in the channel

Figure 6 Target recognition results in the channel

Figure 7 Human target breathing and clutter detected by different channels;

Figure 8 Two human target breaths detected by different channels;

Fig. 9 Clutter detected by different channels;

Figure 10 Inter-channel processing flow;

Figure 11 Schematic diagram of performance evaluation experiment;

Figure 12 results of multi-antenna UWB bio-radar detection and target number estimation when there is no target;

Figure 13 results of multi-antenna UWB bio-radar detection and target number estimation when single target;

Fig. 14 results of multi-antenna UWB bio-radar detection and target number estimation when there are two targets;

Figure 15 Multi-antenna UWB bio-radar detection and target number estimation results for three targets.

detailed description

The present invention will be further described in detail below in conjunction with specific drawings and embodiments, which are explanations rather than limitations of the present invention.

Referring to Fig. 1, the multiple human target estimation system based on UWB biological radar of the present invention is mainly composed of three functional modules: antenna array, radar host and computer. The pulse generator in the radar host generates periodic narrow pulse signals, part of which is sent to the transmitter for shaping, and then radiated out through the transmitting antenna; part of which is sent to the delayer to generate a control signal to trigger the receiver to select the echo signal in terms of distance sexual reception. The delayer is designed with digital multi-channel stepping technology, and the highest precision can reach 10ps. The antenna array consists of four units: 1 transmitting antenna and 3 receiving antennas, all of which use butterfly dipole antennas and are connected to the radar host with shielded coaxial cables. Each receiving antenna is connected with a receiving unit to form a radar echo data channel; the receiving unit includes a receiver and an analog-to-digital converter connected with the receiver. The reinforced power isolation design is adopted between each channel, which can effectively reduce the echo signal coupling between channels. The data sampled by each channel is sent to the computer for storage and processing. The central controller is mainly used for the communication between the computer and the radar host to realize the functions of parameter setting and data transmission. It is connected with the computer by USB.

Since the present invention only involves the estimation of the number of human targets, the antenna array adopts the form of a linear array, and the specific arrangement is as follows: the transmitting antenna is the center; the receiving antenna 1 is located on one side and placed close to it; the receiving antennas 2 and 3 are located The two sides of the transmitting antenna are distributed symmetrically; the distance between them and their centers is 0.3-1m, the lower limit is determined by the size of the antenna, and the upper limit is related to the antenna beam width. The radar transmission pulse repetition frequency is 128KHz, and the three channels adopt synchronous data acquisition. The corresponding parameters are shown in Table 1.

Table 1 Key working parameters of UWB bioradar

time window 80ns starting point 5ns Sampling points 8192 scanning speed 16Hz

The radar echo data processing in the method for estimating multiple human targets based on UWB biological radar of the present invention is described in detail below:

1. In-channel processing

As shown in Figure 2, it is the basic process of processing in the UWB bio-radar channel, and the echo data of each channel is processed by this process. In-channel processing consists of three main steps: a breath detection algorithm, an interference suppression algorithm, and an object recognition algorithm. The breathing detection algorithm mainly realizes the detection of human breathing signals; on this basis, adaptive cancellation is used to suppress the mutual interference of breathing signals between human targets; the energy of the corresponding distance increases due to the breathing of human targets, and finally the energy threshold is set to detect the target. identify.

1. Respiration detection algorithm

Including the following functional modules: signal-to-noise ratio improvement, background removal, adaptive enhancement and low-pass filtering. Among them, the improvement of signal-to-noise ratio is to use the distance information provided by UWB biological radar to averagely improve the signal-to-noise ratio of the echo data in the distance; the background removal is used to remove the static background clutter in the radar echo data that does not change with time. Time-up and subtraction averaging is realized; adaptive enhancement uses the narrow-band characteristics of human breathing to separate it from broadband noise, so as to achieve the purpose of enhancement; low-pass filtering uses digital filters to filter out low-frequency human low-pass signals.

Referring to Fig. 3, it is the UWB bio-radar echo data processed by the breathing detection algorithm, which contains two-dimensional information of time and distance, and the distance is represented by pulse two-way travel time ns. In this data, the human target distance is around 28ns, showing periodic fluctuations over time. By extracting the waveform of the point with the maximum energy distance, the human breathing signal detected by the radar can be obtained. It can be seen from the figure that due to the modulation of human respiratory movement, the UWB bio-radar echo data presents obvious smearing. When the radar is used for multiple target detection, the smearing of the front target will interfere with the rear target, and seriously affect the detection result. Therefore, suppressing it is the key to realize the estimation of the number of multiple human targets.

2. Interference suppression algorithm

The interference suppression processing flow is shown in Figure 4, where the input signal is the waveform of a certain distance point in the UWB bio-radar echo data processed by the breathing detection algorithm, and the reference signal is the breathing signal of the human target before the distance point. After signal subtraction, an error signal is obtained to modify the adaptive filter coefficients. If the input signal corresponds to a target distance point, the process will cancel the tailing of the previous target and output the breathing signal of the target; if the input signal is the tailing caused by the breathing movement of the previous target, the output signal will be zero after cancellation . The adaptive filter is realized by LMS algorithm.

3. Target recognition algorithm

After processing by the above algorithm, the distance point corresponding to the target in the UWB bio-radar echo data will show strong energy due to the effect of human breathing. Accumulate the data over time to obtain a one-dimensional energy-distance distribution. In this distribution, there will be an energy peak at the corresponding distance of the human target, and the target can be identified by setting a threshold for binary judgment of whether there is a target.

The schematic diagram of the target recognition algorithm is shown in Figure 5. Thresholds Th1 and Th2 are set. In the figure, N _win represents the recognition window, which is related to the distance resolution of the system. The number of distance points corresponding to 2 times the UWB radar pulse width is used (N _win =8ns) ; E _obj represents the energy of the target point; E _clu represents the energy of the clutter point.

The energy threshold Th1 is used to judge whether there is a target. When E _obj /E _clu is greater than the energy threshold Th1, it means that the energy of the target point is significantly greater than the energy of the clutter point, and the possibility of being a target is very high (or it can be judged based on this There is a human target), judge that the distance point is a target and output 1, otherwise output 0.

The correlation threshold Th2 is used to select clutter points, and the selection of clutter points follows the following rules: the point in the identification window with the largest energy except the target point and the correlation coefficient between the waveform and the target point waveform is less than the threshold Th2.

Among them, the energy threshold Th1 and the correlation threshold Th2 are empirical values. Th1=2 means that the energy of the target point is significantly higher than that of the clutter point due to human breathing; Th2=0.6. The correlation coefficient is less than this value, which means that the correlation between the two waveforms is low. The probability of correct selection of clutter points is high. The specific processing steps are:

First, calculate the correlation coefficient between the selected clutter point waveform and the target point waveform, and then compare the correlation coefficient with Th2. The correlation coefficient is greater than Th2=0.6, indicating that the two are highly correlated, and the selected clutter point is similar to the target point. If the degree is high, it means that the energy of the clutter point determined at this time is incorrect. Only when the correlation coefficient is less than Th2=0.6, it means that the selected clutter point is the clutter point, and the target point is the target point.

The output result processed by the target recognition algorithm in Figure 5 is shown in Figure 6, which is the data collected by the second channel of the UWB bio-radar system in an experiment to detect two human targets. According to the prior knowledge of the target distance, it can be seen that the two human targets at 14ns and 23ns in the figure are recognized, but residual clutter due to misjudgment appears at 68ns and 84ns. In order to correctly estimate the number of targets, inter-channel processing based on multi-antenna UWB biological radar will continue.

2. Inter-channel processing

Referring to FIG. 7 , it shows the breathing signal of a human target detected by the three radar echo data channels after the multi-antenna UWB biological radar system used in the present invention is used to detect a certain human target. It can be seen from the figure that the three respiratory waveforms are consistent and show a high degree of correlation. This is because for the echo signals generated by the same target, since the three receiving antennas receive and sample independently, the breathing signals detected in different data channels should be completely consistent in theory. The dotted line in the figure indicates the clutter waveform at a certain distance point in a certain channel, which is quite different from the respiratory waveform and has a low degree of correlation.

Fig. 8 shows the breathing signals detected after the multi-antenna UWB bio-radar system used in the present invention detects two human targets, and the waveforms of the two are quite different. This is because there are always differences between different target respiratory rhythms, so the detected respiratory waveforms have a low degree of correlation.

Figure 9 shows two clutter signals detected by two radar echo data channels. Due to the randomness of clutter generation and the fact that it comes from different data channels, there must be no consistency.

The above analysis shows that for multi-antenna UWB bio-radar:

(1) The breathing of the same target detected by different channels is highly correlated;

(2) The respiration correlation of different targets detected by different channels is low;

(3) The correlation between clutter and human target breathing is low;

(4) The correlation between clutters is low.

Therefore, it is considered to carry out correlation processing between channels, which can not only effectively remove clutter, but also distinguish human targets so as to realize the estimation of the number of targets.

Figure 10 shows the basic process of inter-channel processing. Firstly, the correlation coefficient between channels is calculated according to the results of human target recognition in three channels, that is, the waveform corresponding to the target point in each channel recognition result is calculated in pairs between channels. The method as follows:

${\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; ij</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; \&Sigma;S</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}{{<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ]</span>}^{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; /</span>\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Si and Sj represent waveforms on two channels, where i, j=1, 2, 3..., and i≠j.

Set the threshold A, compare the calculated correlation coefficient r _ij with the threshold A, and distinguish different targets. The rules are as follows: (1) For the target points in the recognition results of each channel, there must be another channel target point with which The correlation coefficient is greater than the threshold A, otherwise it is regarded as clutter; (2) The target points with the inter-channel correlation coefficient greater than the threshold A are considered to come from the same target.

The closer the correlation coefficient r _ij is to 1, the higher the degree of correlation between the two waveforms is, that is, the better the consistency. After processing according to the above rules, the residual clutter in each channel is removed, and the targets are classified at the same time, so as to obtain the number of targets. The threshold A here is an empirical value, and a large number of experiments show that the performance is the best when the value is set to 0.6.

The effectiveness of the inventive method is evaluated by specific design experiments below:

As shown in Figure 11, the UWB bio-radar penetrates a 28cm thick brick wall to detect behind the wall. The detection process includes 4 situations: no target, 1 target (P1), 2 targets (P1, P2), 3 targets (P1, P2, P3). Among them, P1 is facing the transmitting antenna and is located about 3m behind the wall; P2 is 20° to the left of the transmitting antenna and is located about 4m behind the wall; P3 is 20° to the right of the transmitting antenna and is located about 5m behind the wall. When there are targets, each target stands facing the wall and remains still.

Figures 12 to 15 show the processing results corresponding to the above four situations. The rows in the figure correspond to channels 1, 2 and 3 from top to bottom; the columns from left to right are the data processed by the interference suppression algorithm in the channel respectively. The one-dimensional energy-distance distribution corresponding to the data, and the human target recognition results. In addition, after inter-channel correlation processing, the clutter points in each channel are marked with crosses, and the distinguished targets are marked with triangles in different directions. It can be seen from these figures that the number of targets identified in the four cases is exactly the same as the actual one. Therefore, without any prior knowledge, the present invention provides an effective method for identifying and estimating the number of multiple human targets in bio-radar, which helps to improve the level and accuracy of bio-radar technology in obtaining human body information. ability.

It should be understood that those skilled in the art can make improvements or changes based on the above description, and all these improvements and changes should belong to the protection scope of the appended claims of the present invention.

Claims (5)

1. the multiple human body target estimating systems based on UWB bioradar, is characterized in that, bagDraw together radar host computer, aerial array and computer, described radar host computer comprise impulse generator, emitter,Central controller, delayer and several receiving elements; Described aerial array comprises 1 transmitting antennaWith 3 reception antennas, transmitting antenna is connected with the emitter of radar host computer, each reception antenna and oneReceiving element is connected to form a radar return data channel; Described receiving element comprise receiver and withThe analog-digital converter that receiver is connected, the output of each analog-digital converter is all connected with central controller,Central controller is realized and being communicated by letter with computer interactive;

Impulse generator produces periodic short pulse signal: a part is delivered to emitter shaping, then passes throughTransmitting antenna radiate; Another part narrow pulse signal is sent to delayer and produces control signal, and touchesTransmission/reception unit selectively receives in distance echo-signal;

Described aerial array adopts linear array arrangement mode, centered by transmitting antenna, the first reception antenna withThe adjacent setting of transmitting antenna, the second reception antenna and the 3rd reception antenna are symmetricly set in respectively transmitting antennaBoth sides, and the distance of the second reception antenna, the 3rd reception antenna and transmitting antenna is 0.3～1m.

2. a kind of UWB bioradar that can estimate multiple human body targets according to claim 1System, is characterized in that, described each independently echo data interchannel employing reinforcement isolated from power design.

3. can estimate multiple human body targets according to a kind of described in any one in claim 1～2UWB bioradar system, is characterized in that, UWB bioradar adopts impulse system, transmittingThe width of pulse is controlled at ns rank, and corresponding spectrum width and centre frequency are 500MHz; TransmittingAntenna and reception antenna all adopt butterfly dipole antenna, and and radar host computer between use shielding coaxial lineCable connects.

4. the multiple human body target evaluation methods based on UWB bioradar, is characterized in that, bagDraw together following steps:

1) data acquisition

After several radar return data channel samplings of UWB bioradar system, by the echo obtainingData are delivered to computer and are carried out Storage and Processing;

2) in passage, process

First, adopt breathing detection algorithm to obtain human body respiration signal to the echo data obtaining; Secondly,Adopt adaptive cancellation method to suppress to breathe and believe between different human body target to the human body respiration signal detectingNumber phase mutual interference; Finally, setting energy threshold by Target Recognition Algorithms identifies target individual;

Described breathing detection algorithm is specially:

First, the range information providing according to UWB bioradar system, to echo data in distanceCarry out moving average processing, improve the signal to noise ratio of echo data;

Secondly, remove time-independent static background clutter in echo data;

Again, utilize the narrow-band characteristic of human body respiration, adopt adaptive algorithm that human body respiration and noise are enteredRow separates;

Finally, carry out LPF according to the gradual feature of human body respiration, obtain the human body that detections of radar arrivesBreath signal;

Described adaptive cancellation method is by a certain distance in the echo data after breathing detection algorithm processFrom the waveform of point as input signal, before this range points the breath signal of human body target as with reference to signal,After subtracting each other with two paths of signals, obtain error signal and go to revise the coefficient of sef-adapting filter; If input is correspondingA certain target range point signal, this flow process can offset the hangover of previous targets and disturb, and exports exhaling of this targetInhale signal; If input signal is the hangover that previous targets respiratory movement causes, output letter after offsetingNumber be zero; Wherein, described sef-adapting filter adopts LMS algorithm to realize;

Described Target Recognition Algorithms is that the data of processing through the method for adaptive cancellation are entered in timeRow is accumulated, and obtains the energy-range distribution of one dimension, in this distribution, in human body target respective distances, there will be energyAmount peak value, has aimless binary judgement to this energy peak setting threshold, can identify target;

The described aimless binary judgement that has refers to:

First, set energy threshold Th1 and dependent thresholds Th2; Energy threshold Th1 has or not for judgementTarget, chooses impact point energy E_obj, clutter point energy E_clu, work as E_obj/E_cluBe greater than energy threshold Th1Time, judge in this range points for target and be output as 1, otherwise output 0; Dependent thresholds Th2 is used forChoose clutter point, in recognition window except impact point energy maximum and waveform and impact point waveform correlation coefficient littleBe regarded as clutter point in the point of dependent thresholds Th2, this energy is as E_clu；

3) interchannel processing

To step 2) waveform corresponding to human body target identification point in each passage of recognizing be according to formula (1)Calculate interchannel correlation coefficient r_ij：

$r_{ij}=\frac{{\mathrm{\&Sigma;S}}_i{S}_j}{{\&lsqb;{\mathrm{\&Sigma;S}}_i^2{\mathrm{\&Sigma;S}}_j^2\&rsqb;}^{1/2}}---\left(1\right)$

Wherein, S_iAnd S_jRepresent the waveform on two passages, wherein i, j=1,2,3 ..., and i ≠ j;

Setting threshold A, by the correlation coefficient r calculating_ijContrast with this threshold value A, carry out differenceThe differentiation of target: for the impact point in each channel recognition result, must have another passage impact pointBe greater than the threshold value A of this setting with its coefficient correlation, otherwise be considered as clutter; Interchannel coefficient correlation is greater than thisThe impact point of the threshold value A of setting is regarded as from same target;

4) last, the clutter in each passage is removed, by target classification, obtain human body target number.

5. a kind of multiple human body target estimations based on UWB bioradar according to claim 4Method, is characterized in that: described energy threshold Th1 and dependent thresholds Th2 are empirical value, andTh1＝2，Th2＝0.6。