CN114325623A - Method for measuring human body limb movement information based on millimeter wave radar - Google Patents
Method for measuring human body limb movement information based on millimeter wave radar Download PDFInfo
- Publication number
- CN114325623A CN114325623A CN202011220695.6A CN202011220695A CN114325623A CN 114325623 A CN114325623 A CN 114325623A CN 202011220695 A CN202011220695 A CN 202011220695A CN 114325623 A CN114325623 A CN 114325623A
- Authority
- CN
- China
- Prior art keywords
- target
- radar
- distance
- signal
- human body
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
- 230000033001 locomotion Effects 0.000 title claims abstract description 83
- 238000000034 method Methods 0.000 title claims abstract description 72
- 238000005259 measurement Methods 0.000 claims abstract description 53
- 238000001514 detection method Methods 0.000 claims abstract description 34
- 238000000605 extraction Methods 0.000 claims abstract description 30
- 230000001629 suppression Effects 0.000 claims abstract description 18
- 239000011159 matrix material Substances 0.000 claims description 36
- 238000010586 diagram Methods 0.000 claims description 17
- 238000012545 processing Methods 0.000 claims description 17
- 238000000354 decomposition reaction Methods 0.000 claims description 15
- 238000012880 independent component analysis Methods 0.000 claims description 13
- 238000009825 accumulation Methods 0.000 claims description 10
- 230000001427 coherent effect Effects 0.000 claims description 10
- 238000004458 analytical method Methods 0.000 claims description 9
- 230000006835 compression Effects 0.000 claims description 6
- 238000007906 compression Methods 0.000 claims description 6
- 238000002592 echocardiography Methods 0.000 claims description 6
- 238000002360 preparation method Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000002156 mixing Methods 0.000 claims description 4
- 238000004806 packaging method and process Methods 0.000 claims description 4
- 239000013598 vector Substances 0.000 claims description 4
- 238000013459 approach Methods 0.000 claims description 2
- 230000002457 bidirectional effect Effects 0.000 claims description 2
- 230000005540 biological transmission Effects 0.000 claims description 2
- 238000000926 separation method Methods 0.000 claims description 2
- 210000003414 extremity Anatomy 0.000 description 93
- 238000007781 pre-processing Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 3
- 238000002310 reflectometry Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000010355 oscillation Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000003897 fog Substances 0.000 description 1
- 210000003127 knee Anatomy 0.000 description 1
- QVRVXSZKCXFBTE-UHFFFAOYSA-N n-[4-(6,7-dimethoxy-3,4-dihydro-1h-isoquinolin-2-yl)butyl]-2-(2-fluoroethoxy)-5-methylbenzamide Chemical compound C1C=2C=C(OC)C(OC)=CC=2CCN1CCCCNC(=O)C1=CC(C)=CC=C1OCCF QVRVXSZKCXFBTE-UHFFFAOYSA-N 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000149 penetrating effect Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 239000000779 smoke Substances 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Images
Landscapes
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention provides a method for measuring human body limb movement information based on a millimeter wave radar, which comprises the following steps: firstly, the method comprises the following steps: transmitting a millimeter wave electromagnetic signal through radar equipment; II, secondly: receiving a scattering echo signal of a human body target; thirdly, the method comprises the following steps: carrying out target distance-Doppler-angle combined measurement on each frame of data; fourthly, the method comprises the following steps: multi-target detection and information extraction; fifthly: estimating the position of a target trunk; sixthly, the method comprises the following steps: torso scatter suppression; seventhly, the method comprises the following steps: determining the distribution of the limb in a radial distance dimension; eighthly: extracting and identifying characteristics; the invention has higher measurement precision and all-weather working characteristics, can realize the measurement of the limb movement information of a plurality of human bodies, improves the measurement efficiency and improves the measurement precision; the device has the characteristics of simple structure, strong compatibility and wide application range; the method is scientific, has good manufacturability and has wide popularization and application values.
Description
Technical Field
The invention relates to a method for measuring human body limb movement information based on a millimeter wave radar, which adopts related technologies such as radar signal processing and the like to realize the characteristic extraction and identification of the limb movement information and finally realize the targets such as judgment of human body movement intention, judgment of external personnel and the like, has wide application range and belongs to the field of digital signal processing.
Background
The limb movement information measuring method is information brought by relative movement of limbs relative to a trunk when a human body moves, and the movement of the human body can be considered as combination of movement of different parts, namely the trunk and other parts move in a translation mode, and the limbs periodically swing relative to the trunk. Due to the characteristics of large trunk area, high reflectivity and the like, the strength of trunk scattering signals is far greater than that of limb scattering signals, and the accuracy of limb movement information measurement is influenced. In practical application scenarios, a lot of collections in a museum have millions of values, and the museum has a large floor area and a large flow of people and is often threatened by thieves. Once a theft event occurs, a thief can only be searched by monitoring the playback of a video, and the efficiency is extremely low, so that a museum needs a reliable, sensitive and accurate anti-theft system. With the continuous deepening of scientific and technological innovation ideas, the demands of cultural relic protection units on asset anti-theft monitoring systems are increasingly diversified, the high-tech crime rate is continuously increased in recent years, and the defects of high false alarm rate, easiness in damage and the like of conventional technical precaution means cannot meet the ever-increasing safety demands of users. Therefore, the problem of measuring the limb movement information needs to be solved urgently.
In the prior art, an effective solution is not provided at present for the problem that the limb movement information cannot be conveniently and accurately measured in the related art. Since the country was built, seven major medical records of the native palace treasure case recorded by the Beijing police have been published, namely 1959, 1962, 1980, 1987 (two), 1991 and 2011, wherein only five of the records are disclosed. The theft cases always rank the first criminal case for years, generally account for about 80% of the criminal case, and the entrance theft cases generally account for more than six times of the whole theft cases. It can be said that the frequent occurrence of theft cases, especially burglary cases, is the most important problem which disturbs the current security management order. At present, the anti-theft technology adopted by most museums is limited to capturing images by a camera, and a background is used for manually judging whether a theft action occurs, so that the method consumes manpower and financial resources and does not meet the increasing safety requirement.
The equipment mainly used by the invention is the millimeter wave radar which has the characteristics of higher measurement precision and all-weather and all-day working and can still exert better detection performance in the severe environments of weak light, rain, snow and the like. Compared with centimeter wave radar, the millimeter wave radar has the characteristics of small volume, light weight and high spatial resolution, and has stronger capability of penetrating fog, smoke and dust compared with optical radars such as infrared and laser. The system can work in a plurality of modulation modes, wherein a linear Frequency Modulated Continuous Wave (FMCW) mode can realize high-precision measurement of distance and speed at the same time.
Aiming at the above mentioned situations and practical application requirements, the invention provides a method for measuring human body limb movement information based on a millimeter wave radar, which can measure information brought by relative movement of limbs relative to a trunk when a human body moves, can realize the targets of judgment of human body movement intention, judgment of external personnel and the like according to the limb movement information, and has strong compatibility and wide application range.
Disclosure of Invention
The invention aims to provide a method for measuring human body limb movement information based on a millimeter wave radar. The system flow diagram is shown in fig. 1.
The invention relates to a method for measuring human body limb movement information based on a millimeter wave radar, which specifically comprises the following steps as shown in figure 1:
the method comprises the following steps: transmitting millimeter wave electromagnetic signals by radar devices
The radar equipment is arranged indoors, and a transmitting end of the radar equipment transmits linear frequency modulation continuous waves which are consistent in initial phase and repeated at a certain frequency to a space through a plurality of transmitting antennas; the case of transmitting a millimeter wave electromagnetic signal by a radar device is shown in fig. 2;
assuming that the slope of emission sweep frequency of the radar at the starting moment is k, the period is T, and the initial phase is
Of the radar emitting signal s for an arbitrary time tT(t) can be expressed as:
where A is the signal amplitude, exp is a complex representation, j is a complex symbol, f0M is the initial frequency of the signal, 1,2, and M is the serial number of each pulse;
the method for transmitting the millimeter wave electromagnetic signal by the radar equipment comprises the following steps: firstly, obtaining FMCW linear frequency modulation signals, converting the FMCW linear frequency modulation signals into Analog signals through a Digital-to-Analog Converter (DAC) after forming multi-channel signals, and sending the Analog signals to a detection target through a transmitting antenna to form an effective radar action area;
step two: receiving scattering echo signals of human body targets
The target scattering echo signal is a backward scattering echo received by a transmitting signal after being reflected by a human target, a radar echo and a local oscillator signal received by an antenna are subjected to frequency mixing to output a difference frequency signal, and the difference frequency signal is converted into a mathematical signal convenient to process through an anti-aliasing filter (AAF) and an Analog-to-Digital Converter (ADC); when receiving signals, a mode of receiving echo signals by multiple antennas is adopted, angles can be measured through phase differences among the multiple antennas, and the motion states of multiple targets can be observed at the same time; the situation of receiving the scattering echo signal of the human target is shown in fig. 3;
because there is more than one scattering point on the human body target, the actually obtained echo signal should be the superposition of the echo signals of each scattering point, and it is assumed that there are N scattering points on the human body targetIf the speed direction is positive when the target approaches the radar, the echo signal s of the human target is a single receiving channelR(t) can be expressed as:
wherein sigmanIs the electromagnetic scattering intensity, t, of the n-th scattering centern0Delaying the transmission of the bidirectional signal from the nth scattering center to the radar;
the method for receiving the scattering echo signal of the human body target comprises the following steps: the scattered echo signals are received by a plurality of receiving antennas of the radar and then converted into analog electric signals, radar echoes received by the antennas and local oscillator signals are mixed to output difference frequency signals, the difference frequency signals are filtered by an anti-aliasing filter, and then the difference frequency signals are converted into digital signals convenient for fast processing by an ADC (analog to digital converter); the signals received by different receiving antennas jointly form a plurality of paths of parallel signals so as to realize the discrimination of targets in different directions;
step three: performing target distance-Doppler-angle joint measurement on each frame of data
After receiving a backscattering echo signal of a target human body, packaging data by taking a frame as a unit, performing pulse compression on a distance dimension of a target echo of each frame of data to obtain a one-dimensional Range profile matrix, and performing coherent accumulation processing on the one-dimensional Range profile matrix to obtain energy distribution of the signal in a Range-Doppler two-dimensional plane, namely a Range-Doppler (RD) matrix, wherein an RD image can be drawn through the RD matrix; because the Doppler value and the velocity value are in a linear relation, the RD matrix can also be converted into a distance-velocity matrix; angle information can be obtained through the phase difference relation among the multiple receiving antennas, namely the distance-Doppler-angle combined information is finally obtained;
the method for carrying out target distance-Doppler-angle joint measurement on each frame of data comprises the following steps: pulse compression and coherent accumulation are carried out on each frame of data, the delay time of a target echo is measured, the relative distance between a radar and a target is obtained through calculation, then the Doppler frequency of the echo is extracted, the relative speed between the target and the radar is obtained through calculation, and finally distance-Doppler-angle joint information is obtained through the phase difference relation among multiple receiving antennas;
step four: multi-target detection and information extraction
The data received by the multiple receiving antennas can obtain the running information of the targets in the radar sight line through signal processing, namely the position and the speed of one or more targets in the RD diagram can be determined, and multi-target detection and information extraction are completed;
the method for multi-target detection and information extraction comprises the following steps: performing Constant False Alarm Rate (CFAR) detection on the RD matrix after the target echo signal processing, wherein each peak value after the CFAR detection corresponds to a target, and the distance and the speed of the target corresponding to the peak value point position can be obtained corresponding to the position of the peak value after the CFAR detection in the RD matrix; the multi-target detection and information extraction are completed through the speed, distance, position and other information of the targets in a combined manner, and preparation is made for the subsequent measurement of the body and limb movement information;
step five: target torso position estimation
In order to determine the limb movement information, the position of a target trunk must be determined firstly, the azimuth angle and the pitch angle of the target trunk relative to the radar are different due to different multi-target positions, and the position of the target trunk can be obtained through distance-angle combined information; secondly, a basis is provided for the subsequent measurement of limb movement information through the restriction relationship of the limb relative to the body position;
the method for estimating the position of the target trunk is as follows: after the peak positions of the targets are obtained, the radar scattering cross-sectional area of the trunk in the human body target is the largest, namely the area is the largest on the radar sight line, so that the peak positions can be generally regarded as distance units where the trunk is located; the position of the trunk can be specifically determined in the distance-angle diagram, namely the range of the distance-angle unit of the limb is defined, and preparation is made for the subsequent measurement of the limb movement information;
step six: trunk scatter suppression
In an actual measurement scene, due to the characteristics of large trunk area, high reflectivity and the like, after the target distance-Doppler-angle combined measurement is carried out, the trunk scattering signal intensity is far greater than the limb scattering signal intensity, so that the accuracy of limb movement information measurement is influenced; therefore, the method for suppressing the trunk scattering is adopted to preprocess the echo signals, the frequency components in a specific frequency band can be removed by using the method, and the method has the function of reducing the influence of trunk signal components on the measurement accuracy of the limb movement information to the maximum extent; torso scatter suppression is shown in fig. 4;
the trunk scattering suppression method comprises the following steps:
1) recording the amplitude and the phase of the peak position of the target trunk in the obtained RD image and the position of the maximum point;
2) calculating Doppler frequency according to the peak position, and reconstructing a time domain signal of the subharmonic;
3) subtracting the reconstructed time domain signal from the original signal to obtain a new time domain signal;
4) performing coherent accumulation on the new time domain signal to obtain a new RD matrix and drawing an RD diagram;
5) repeating the steps 1 to 4 until the maximum iteration number is reached;
the advantage of preprocessing the signals by applying the torso scattering suppression method is that after the preprocessing, the limb movement signals are easier to distinguish in the RD diagram than before the preprocessing, which is more beneficial to the measurement of the later limb movement information;
step seven: determining the distribution of a limb in a radial distance dimension
The radial distance is the projection of the distance between the radar and the target on the radar sight line, and the radar sight line direction is the orientation of the radar; because the limb moves relative to the trunk, the distance unit where the limb is located in the RD image has a certain Doppler dimension broadening, the broadening corresponds to a peak value on the Doppler dimension, and the distance unit where the limb is located can be determined through the broadening on the Doppler dimension; the situation of determining the distribution of the limb in the radial distance dimension is shown in fig. 5;
the method for determining the distribution of the limb in the radial distance dimension comprises the following steps:
1) determining a widened Doppler range in the trunk distance range on the RD image according to the trunk distance dimensional range determined in the step five;
2) weighting and summing each distance unit where the trunk is located;
3) the summation result of the distance units where the limbs are located is larger than the summation result of the distance units on the two sides of the limbs, so that the distance unit where the trunk exists is determined;
for the distance units with limbs, whether the distance units are larger than the summation mean value or not can be divided into two types of distance units for determining the existence of the limbs and distance units for possibly having the limbs, and because the distance units are farther away from the radar in an actual measurement scene, the signal intensity is weaker, so that the two types of distance units provide a basis for accurately judging the motion information of the limbs;
step eight: feature extraction and recognition
Performing time-frequency analysis on the one-dimensional range profile sequence of the range units corresponding to the radial range dimension distribution of the limb determined in the step seven, and then applying an Independent Component Analysis (ICA) method to the spectrogram to perform decomposition and reconstruction so as to complete feature extraction and identification; the ICA method is mostly applied to the blind source separation problem, a human body target in motion can be regarded as superposition of motion of each part, echo waves of the human body target can be expressed as superposition of echo waves of each part, and fluctuation and rotation motion of various forms contained in the superposition can be described as combination of various micro-motion information sources; the feature extraction and recognition situation is shown in fig. 6;
the method for extracting and identifying the features comprises the following steps: .
1) Performing ICA analysis on the whole data set to obtain a plurality of independent Information Sources (ICs);
2) the original signal is reconstructed by these ICs to obtain the decomposition coefficients (a) of the original signal on each IC1,a2…an) Taking the decomposition coefficient as a feature vector of the original data to train a classifier;
3) when new data to be classified appear, decomposing the data by using the trained IC, and sending the obtained decomposition coefficient into a classifier to complete classification;
after the limb movement information is classified, feature extraction and recognition are completed, namely limb movement information measurement is completed, and then subsequent targets such as human movement intention judgment, external person judgment and the like can be performed.
The advantages and the effects are as follows:
1) the device has the characteristics of high measurement precision and all-weather working, and can still exert good detection performance in severe environments such as weak light and the like;
2) the multi-channel signal simultaneous transmitting and receiving function is provided, and the angle estimation of the target can be completed, so that the limb movement information of a plurality of human bodies can be measured, and the measurement efficiency is improved;
3) the trunk scattering suppression method is used for preprocessing the target scattering echo signals, unnecessary clutter signals are filtered, interference of the trunk signals on measurement of the limb movement information is effectively avoided, and the measurement precision is improved;
4) compared with the traditional limb movement information measuring method, the method has the characteristics of simple structure, strong compatibility and wide application range; the method is scientific, has good manufacturability and has wide popularization and application values.
Drawings
FIG. 1 is a flow chart of the system of the present invention.
Fig. 2 is a diagram of the invention for transmitting millimeter wave electromagnetic signals by a radar device.
Fig. 3 is a diagram of the situation of receiving the scattering echo signal of the human target according to the invention.
Fig. 4 is a diagram of torso scatter suppression in accordance with the present invention.
FIG. 5 is a graph of the present invention determining the distribution of the limb in the radial distance dimension.
FIG. 6 is a diagram of feature extraction and recognition in accordance with the present invention.
The symbols in the figures are as follows:
a DAC digital-to-analog converter; an AAF anti-aliasing filter; an ADC analog-to-digital converter; ICA independent component analysis; an IC independent information source; (a)1,a2…an) The decomposition coefficient.
Detailed Description
In specific engineering practice, millimeter wave radars are installed in indoor places such as museums and the like, millimeter wave electromagnetic signals are transmitted and received, target distance-Doppler-angle joint measurement is carried out on each frame of data, then multi-target detection and information extraction are carried out, the position of a target trunk can be estimated, trunk scattering suppression is carried out to determine distribution of limbs in a radial distance dimension, feature extraction and identification are carried out on the distribution positions of the limbs, measurement of limb movement information is completed, and finally, judgment of human movement intentions, judgment of external personnel and other targets are achieved.
The invention provides a method for measuring human body limb movement information based on a millimeter wave radar, a system flow chart is shown in figure 1, and the specific implementation mode comprises the following steps:
the method comprises the following steps: transmitting millimeter wave electromagnetic signals by radar devices
The radar equipment is arranged indoors, and a transmitting end of the radar equipment transmits linear frequency modulation continuous waves which are consistent in initial phase and repeated at a certain frequency to the space through a plurality of transmitting antennas. Firstly, FMCW linear frequency modulation signals are obtained to form multi-channel signals, then the multi-channel signals are converted into analog signals through a digital-to-analog converter, and the analog signals are transmitted to a detection target through a transmitting antenna to form an effective radar action area. The millimeter wave radar has the characteristics of small volume, light weight and high spatial resolution, and improves the accuracy of the measurement of the limb movement information. The case of transmitting a millimeter wave electromagnetic signal by a radar device is shown in fig. 2;
step two: receiving scattering echo signals of human body targets
After millimeter wave electromagnetic signals are transmitted, scattered echo signals are received by a plurality of receiving antennas of a radar and then converted into analog electric signals, radar echoes received by the antennas and local oscillation signals are subjected to frequency mixing to output difference frequency signals, filtering is performed through an anti-aliasing filter, the difference frequency signals are converted into digital signals convenient for rapid processing through an analog/digital converter, and the signals received by different receiving antennas jointly form a plurality of paths of parallel signals so as to realize the discrimination of targets in different directions; the millimeter wave radar has the characteristics of high measurement precision and all-weather and all-day work, and can play a good role in detection in various severe environments; the situation of receiving the scattering echo signal of the human target is shown in fig. 3;
step three: performing target distance-Doppler-angle joint measurement on each frame of data
After receiving a backscattering echo signal of a target human body, packaging data by taking a frame as a unit, performing pulse compression on a distance dimension of a target echo of each frame of data to obtain a one-dimensional range profile matrix, and performing coherent accumulation processing on the one-dimensional range profile matrix to obtain energy distribution of the signal on a range-Doppler two-dimensional plane, namely a range-Doppler matrix, wherein an RD (distance-Doppler) diagram can be drawn through the RD matrix; because the Doppler value and the velocity value are in a linear relation, the RD matrix can also be converted into a distance-velocity matrix; angle information can be obtained through the phase difference relation among the multiple receiving antennas, namely the distance-Doppler-angle combined information is finally obtained; the linear frequency modulation continuous wave mode can simultaneously realize high-precision measurement of distance, speed and angle, and improve the measurement precision of human limb movement;
step four: multi-target detection and information extraction
After the RD matrix of each frame of data is obtained through measurement, CFAR detection is carried out on the RD matrix after target echo signal processing, each peak value after CFAR detection corresponds to a target, the distance and the speed of the target corresponding to the position of the peak value point can be obtained corresponding to the position of the peak value after CFAR detection in the RD matrix, and multi-target detection and information extraction are completed through combined processing of the speed, the distance, the position and other information of a plurality of targets;
step five: target torso position estimation
After information such as the speed, the distance and the position of multiple targets is obtained, the radar scattering cross section area of the trunk in the human body target is the largest, namely the area on the radar sight line is the largest, so the peak position can be generally regarded as a distance unit where the trunk is located; the position of the trunk can be specifically determined in the distance-angle diagram, namely the range of the distance-angle unit of the limb is defined, and preparation is made for the subsequent measurement of the limb movement information;
step six: trunk scatter suppression
In an actual measurement scene, due to the characteristics of large trunk area, high reflectivity and the like, after the target distance-Doppler-angle combined measurement is carried out, the trunk scattering signal intensity is far greater than the limb scattering signal intensity, so that the accuracy of limb movement information measurement is influenced; preprocessing an echo signal by adopting a trunk scattering suppression method, after estimating the position of a target trunk, recording the amplitude, the phase and the maximum position of the peak position, calculating Doppler frequency and reconstructing a time domain signal, subtracting the reconstructed time domain signal from an original signal, performing coherent accumulation on a new time domain signal, and performing loop iteration until the maximum iteration number to complete suppression of the trunk signal; by using the method, the frequency components in a specific frequency band can be removed, and the influence of the body signal component on the measurement accuracy of the limb movement information is reduced to the maximum extent; torso scatter suppression is shown in fig. 4;
step seven: determining the distribution of a limb in a radial distance dimension
The radial distance is the projection of the distance between the radar and the target on the radar sight line, and the radar sight line direction is the orientation of the radar. Because the limb moves relative to the trunk, the distance unit where the limb is located in the RD image has a certain Doppler dimension broadening, the broadening corresponds to a peak value on the Doppler dimension, and the distance unit where the limb is located can be determined through the broadening on the Doppler dimension; after the body scattering is inhibited, firstly determining the broadened Doppler dimension range, then carrying out weighted summation on each distance unit where the body is located, determining the distance units where the limbs are located, and dividing the distance units into two types of determining that the limbs exist and possibly the limbs exist; the situation of determining the distribution of the limb in the radial distance dimension is shown in fig. 5;
step eight: feature extraction and recognition
Performing time-frequency analysis on the one-dimensional range profile sequence of the range units corresponding to the radial range dimension distribution of the limb determined in the step seven, and performing ICA (independent component analysis) analysis on the whole result after the time-frequency analysis to obtain a plurality of independent information sources (IC); the original signal is reconstructed by these ICs to obtain the decomposition coefficients (a) of the original signal on each IC1,a2…an) Taking the decomposition coefficient as a feature vector of the original data to train a classifier; when new data to be classified appear, decomposing the data by using the trained IC, and sending the obtained decomposition coefficient into a classifier to complete classification; for body movement messageAfter the information is classified, feature extraction and recognition are completed, namely limb movement information measurement is completed, and then subsequent targets such as human body movement intention judgment, external person judgment and the like can be performed; the method is scientific, has good manufacturability and has wide popularization and application value; the feature extraction and recognition is shown in fig. 6.
Specific embodiments are as follows:
in the prior art, an effective solution is not provided at present for the problem that the limb movement information cannot be conveniently and accurately measured in the related art. Since the country was built, seven major medical records of the native palace treasure case recorded by the Beijing police have been published, namely 1959, 1962, 1980, 1987 (two), 1991 and 2011, wherein only five of the records are disclosed. In specific engineering practice, millimeter wave radars can be installed in museums or other indoor places, millimeter wave electromagnetic signals are transmitted and received, target distance-Doppler joint measurement is carried out on each frame of data, then multi-target detection and information extraction are carried out, the position of a target trunk can be estimated, trunk scattering suppression is carried out to determine distribution of limbs in a radial distance dimension, then feature extraction and identification are carried out on the distribution positions of the limbs, and measurement of the motion information of the limbs is completed.
The invention relates to a method for measuring human body limb movement information based on a millimeter wave radar, which comprises the following steps:
the method comprises the following steps: transmitting millimeter wave electromagnetic signals by radar devices
The radar equipment can be installed in a museum or other indoor places, and the transmitting end of the radar equipment transmits linear frequency modulation continuous waves which are initially consistent and are repeated at a certain frequency to the space by a plurality of transmitting antennas; firstly, obtaining FMCW linear frequency modulation signals, converting the FMCW linear frequency modulation signals into analog signals through a digital-to-analog converter after forming multi-channel signals, and sending the analog signals to a detection target through a transmitting antenna to form an effective radar action area;
step two: receiving scattering echo signals of human body targets
After millimeter wave electromagnetic signals are transmitted, scattered echo signals are received by a plurality of receiving antennas of a radar and then converted into analog electric signals, radar echoes received by the antennas and local oscillation signals are subjected to frequency mixing to output difference frequency signals, filtering is performed through an anti-aliasing filter, the difference frequency signals are converted into digital signals convenient for rapid processing through an analog/digital converter, and the signals received by different receiving antennas jointly form a plurality of paths of parallel signals so as to realize the discrimination of targets in different directions; because the number of human body targets is usually not unique in indoor scenes such as museums, the identification of the targets at different angles through a plurality of receiving antennas is beneficial to the measurement of the later limb movement information.
Step three: performing target distance-Doppler-angle joint measurement on each frame of data
After receiving a backscattering echo signal of a target human body, packaging data by taking a frame as a unit, performing pulse compression on a distance dimension of a target echo of each frame of data to obtain a one-dimensional range profile matrix, and performing coherent accumulation processing on the one-dimensional range profile matrix to obtain energy distribution of the signal on a range-Doppler two-dimensional plane, namely a range-Doppler matrix, wherein an RD (distance-Doppler) diagram can be drawn through the RD matrix; because the Doppler value and the velocity value are in a linear relation, the RD matrix can also be converted into a distance-velocity matrix; angle information can be obtained through the phase difference relation among the multiple receiving antennas, namely the distance-Doppler-angle combined information is finally obtained; the linear frequency modulation continuous wave mode can simultaneously realize high-precision measurement of distance, speed and angle, and improve the measurement precision of human limb movement;
step four: multi-target detection and information extraction
After the RD matrix of each frame of data is obtained through measurement, CFAR detection is carried out on the RD matrix after target echo signal processing, each peak value after CFAR detection corresponds to a target, the distance and the speed of the target corresponding to the position of the peak value point can be obtained corresponding to the position of the peak value after CFAR detection in the RD matrix, and multi-target detection and information extraction are completed through combined processing of the speed, the distance, the position and other information of a plurality of targets;
step five: target torso position estimation
After information such as the speed, the distance and the position of multiple targets is obtained, the radar scattering cross section area of the trunk in the human body target is the largest, namely the area on the radar sight line is the largest, so the peak position can be generally regarded as a distance unit where the trunk is located; the position of the trunk can be specifically determined in the distance-angle diagram, namely the range of the distance-angle unit of the limb is defined, and preparation is made for the subsequent measurement of the limb movement information; because the structures of different human bodies are similar, after the position of the target body is estimated, the position of the target limb can be roughly estimated;
step six: torso signal scatter suppression
Preprocessing an echo signal by adopting a trunk scattering suppression method, after estimating the position of a target trunk, recording the amplitude, the phase and the maximum position of the peak position, calculating Doppler frequency and reconstructing a time domain signal, subtracting the reconstructed time domain signal from an original signal, performing coherent accumulation on a new time domain signal, and performing loop iteration until the maximum iteration number to complete suppression of the trunk signal; because the strength of the body scattering signal is far greater than that of the limb scattering signal, after the body signal scattering is inhibited, the limb signal is more obvious than before, and the position and the motion state of the limb are easier to estimate;
step seven: determining the distribution of a limb in a radial distance dimension
Because the limb moves relative to the trunk, the distance unit where the limb is located in the RD image has a certain Doppler dimension broadening, the broadening corresponds to a peak value on the Doppler dimension, and the distance unit where the limb is located can be determined through the broadening on the Doppler dimension; after the body scattering is inhibited, firstly determining the broadened Doppler dimension range, then carrying out weighted summation on each distance unit where the body is located, determining the distance units where the limbs are located, and dividing the distance units into two types of determining that the limbs exist and possibly the limbs exist; the human body limbs basically comprise arms, knees and feet, the movement of the three limbs is periodic in an RD diagram, in addition, the position distribution of the limbs and the body of the human body is fixed, and the distribution correctness of the limbs in a radial distance dimension can be verified according to the point;
step eight: feature extraction and recognition
Performing time-frequency analysis on the one-dimensional range profile sequence of the distance units corresponding to the radial distance dimensional distribution of the limb determined in the step seven, performing independent component analysis on the whole result after the time-frequency analysis to obtain a plurality of independent information sources, reconstructing the original signal by using the independent information sources to obtain the decomposition coefficients of the original signal on each independent information source, and training a classifier by using the decomposition coefficients as the feature vectors of the original data; when new data to be classified appear, decomposing the data by using the trained independent source, and sending the obtained decomposition coefficient into a classifier to complete classification; after classification, the limb movement information of the external personnel and the limb movement information of the internal personnel can be compared, and the goals of judging the human body movement intention, judging the external personnel and the like are completed; currently, the anti-theft technology adopted by most museums is only limited to capturing images by a camera, and a background is used for manually judging whether theft happens or not, so that the method consumes manpower and financial resources and does not meet the increasing safety requirements.
Claims (9)
1. A method for measuring human body limb movement information based on a millimeter wave radar is characterized in that: the method comprises the following steps:
the method comprises the following steps: transmitting millimeter wave electromagnetic signals by radar devices
The radar equipment is arranged indoors, and a transmitting end of the radar equipment transmits linear frequency modulation continuous waves which are initially consistent and are repeated at a preset frequency to the space through a plurality of transmitting antennas;
the radar emits sweep frequency with slope k at the initial time, period T and initial phase
Of the radar emitting signal s for an arbitrary time tT(t) can be expressed as:
where A is the signal amplitude, exp is a complex representation, j is a complex symbol, f0M is the initial frequency of the signal, 1,2, and M is the serial number of each pulse;
step two: receiving scattering echo signals of human body targets
The target scattering echo signal is a backward scattering echo received by a transmitting signal after being reflected by a human body target, a radar echo and a local oscillator signal received by an antenna are subjected to frequency mixing to output a difference frequency signal, the difference frequency signal passes through an anti-aliasing filter (AAF), and is converted into a mathematical signal convenient to process through an analog/digital converter (ADC); when receiving signals, a mode of receiving echo signals by multiple antennas is adopted, angles can be measured through phase differences among the multiple antennas, and the motion states of multiple targets can be observed at the same time;
the actual echo signal is the superposition of a plurality of scattering point echo signals due to more than one scattering point on the human body target, if the human body target is provided with N scattering points and the speed direction is positive when the target approaches a radar, the echo signal s of the human body target is a single receiving channelR(t) can be expressed as:
wherein sigmanIs the electromagnetic scattering intensity, t, of the n-th scattering centern0Delaying the transmission of the bidirectional signal from the nth scattering center to the radar;
step three: performing target distance-Doppler-angle joint measurement on each frame of data
After receiving a backscattering echo signal of a target human body, packaging data by taking a frame as a unit, performing pulse compression on a distance dimension of a target echo of each frame of data to obtain a one-dimensional range profile matrix, performing coherent accumulation processing on the one-dimensional range profile matrix to obtain energy distribution of the signal on a range-Doppler two-dimensional plane, namely a range-Doppler RD matrix, and drawing an RD image through the RD matrix; because the Doppler value and the velocity value are in a linear relation, the RD matrix can also be converted into a distance-velocity matrix; angle information can be obtained through the phase difference relation among the multiple receiving antennas, namely the distance-Doppler-angle combined information is finally obtained;
step four: multi-target detection and information extraction
The data received by the multiple receiving antennas can obtain the running information of the targets in the radar sight line through signal processing, namely the positions and the speeds of one or more targets in the RD diagram can be determined, and multi-target detection and information extraction are completed;
step five: target torso position estimation
In order to determine the limb movement information, the position of a target trunk must be determined firstly, the azimuth angle and the pitch angle of the target trunk relative to the radar are different due to different multi-target positions, and the position of the target trunk can be obtained through distance-angle combined information; secondly, a basis is provided for the subsequent measurement of limb movement information through the restriction relationship of the limb relative to the body position;
step six: trunk scatter suppression
The method for suppressing the trunk scattering is adopted to preprocess the echo signals, and the method can remove frequency components in a specific frequency band, thereby reducing the influence of trunk signal components on the accuracy of measuring the limb movement information to the maximum extent;
the radial distance is the projection of the distance between the radar and the target on the radar sight line, and the radar sight line direction is the orientation of the radar; because the limb moves relative to the trunk, the distance unit where the limb is located in the RD image has a certain Doppler dimension broadening, the broadening corresponds to a peak value on the Doppler dimension, and the distance unit where the limb is located is determined through the broadening on the Doppler dimension;
step eight: feature extraction and recognition
Performing time-frequency analysis on the one-dimensional range profile sequence of the range units corresponding to the radial range dimension distribution of the limb determined in the step seven, and then applying an Independent Component Analysis (ICA) method to the spectrogram to perform decomposition and reconstruction so as to complete feature extraction and identification; the ICA method is mostly applied to the blind source separation problem, a human body target in motion is regarded as superposition of motion of each part, echoes are represented as superposition of echoes of each part, and fluctuation and rotation motion of various forms are contained in the superposition and are described as combination of various micro-motion information sources.
2. The method for measuring the human body limb movement information based on the millimeter wave radar as claimed in claim 1, wherein the method comprises the following steps: in the step one, the method of "transmitting millimeter wave electromagnetic signals by radar equipment" includes: firstly, FMCW linear frequency modulation signals are obtained to form multi-channel signals, then the multi-channel signals are converted into analog signals through a digital-to-analog converter (DAC), and the analog signals are transmitted to a detection target through a transmitting antenna to form an effective radar action area.
3. The method for measuring the human body limb movement information based on the millimeter wave radar as claimed in claim 1, wherein the method comprises the following steps: the method for receiving the scattering echo signal of the human body target in the step two comprises the following steps: the scattered echo signals are received by a plurality of receiving antennas of the radar and then converted into analog electric signals, radar echoes received by the antennas and local oscillator signals are mixed to output difference frequency signals, the difference frequency signals are filtered by an anti-aliasing filter, and then the difference frequency signals are converted into digital signals convenient for fast processing by an ADC (analog to digital converter); the signals received by different receiving antennas jointly form a plurality of paths of parallel signals so as to realize the discrimination of different direction targets.
4. The method for measuring the human body limb movement information based on the millimeter wave radar as claimed in claim 1, wherein the method comprises the following steps: the method of "performing target range-doppler-angle joint measurement on each frame of data" described in step three is as follows: pulse compression and coherent accumulation are carried out on each frame of data, delay time of target echo is measured, relative distance between a radar and a target is obtained through calculation, Doppler frequency of the echo is extracted, relative speed between the target and the radar is obtained through calculation, and finally distance-Doppler-angle joint information is obtained through a phase difference relation among multiple receiving antennas.
5. The method for measuring the human body limb movement information based on the millimeter wave radar as claimed in claim 1, wherein the method comprises the following steps: the multi-target detection and information extraction method in the fourth step comprises the following steps: performing Constant False Alarm Rate (CFAR) detection on the RD matrix processed by the target echo signal, wherein each peak value after the CFAR detection corresponds to a target, and the distance and the speed of the target corresponding to the peak value point position can be obtained corresponding to the position of the peak value after the CFAR detection in the RD matrix; and completing multi-target detection and information extraction by combining the information of the speed, the distance and the position of a plurality of targets, and preparing for the subsequent measurement of the body and limb movement information.
6. The method for measuring the human body limb movement information based on the millimeter wave radar as claimed in claim 1, wherein the method comprises the following steps: the "target torso position estimation" in step five is performed by: after the peak positions of the targets are obtained, the radar cross-sectional area of the trunk in the human body target is the largest, namely the area is the largest on the radar sight line, so that the peak positions can be generally regarded as distance units where the trunk is located; the position of the trunk can be specifically determined in the distance-angle diagram, namely the range of the distance-angle unit of the limb is defined, and preparation is made for the subsequent measurement of the limb movement information.
7. The method for measuring the human body limb movement information based on the millimeter wave radar as claimed in claim 1, wherein the method comprises the following steps: "torso scatter suppression" as described in step six, by:
1) recording the amplitude and the phase of the peak position of the target trunk in the obtained RD image and the position of the maximum point;
2) calculating Doppler frequency according to the peak position, and reconstructing a time domain signal of the subharmonic;
3) subtracting the reconstructed time domain signal from the original signal to obtain a new time domain signal;
4) performing coherent accumulation on the new time domain signal to obtain a new RD matrix and drawing an RD diagram;
5) repeating steps 1 to 4 until the maximum number of iterations.
8. The method for measuring the human body limb movement information based on the millimeter wave radar as claimed in claim 1, wherein the method comprises the following steps: in step seven, the "determining the distribution of the limbs in the radial distance dimension" is performed by:
1) determining a widened Doppler range in the trunk distance range on the RD image according to the trunk distance dimensional range determined in the step five;
2) weighting and summing each distance unit where the trunk is located;
3) the summation result of the distance units where the limbs are located is larger than the summation result of the distance units on the two sides of the limbs, so that the distance unit where the trunk exists is determined;
for the distance units with limbs, the distance units determining the existence of the limbs and the distance units possibly having the limbs can be divided into two types according to whether the distance units are larger than the summation mean value or not, and because the farther the distance units are away from the radar in an actual measurement scene, the weaker the signal intensity is, the two types of distance units provide bases for accurate judgment of the limb movement information.
9. The method for measuring the human body limb movement information based on the millimeter wave radar as claimed in claim 1, wherein the method comprises the following steps: the method for extracting and identifying the features in the step eight comprises the following steps:
1) performing ICA analysis on the whole data set to obtain an independent information source, namely IC;
2) the original signal is reconstructed by these ICs to obtain the decomposition coefficients (a) of the original signal on each IC1,a2…an) Taking the decomposition coefficient as a feature vector of the original data to train a classifier;
3) when new data to be classified appear, decomposing the data by using the trained IC, and sending the obtained decomposition coefficient into a classifier to complete classification;
after the limb movement information is classified, feature extraction and recognition are completed, namely limb movement information measurement is completed, and then subsequent human movement intention judgment and external person judgment target can be carried out.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011220695.6A CN114325623B (en) | 2020-11-05 | 2020-11-05 | Human body limb movement information measurement method based on millimeter wave radar |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011220695.6A CN114325623B (en) | 2020-11-05 | 2020-11-05 | Human body limb movement information measurement method based on millimeter wave radar |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114325623A true CN114325623A (en) | 2022-04-12 |
| CN114325623B CN114325623B (en) | 2024-05-28 |
Family
ID=81031864
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011220695.6A Active CN114325623B (en) | 2020-11-05 | 2020-11-05 | Human body limb movement information measurement method based on millimeter wave radar |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114325623B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114563784A (en) * | 2022-04-27 | 2022-05-31 | 武汉理工大学 | Shipborne environment intrusion detection method and system based on double millimeter wave radar |
| CN115857595A (en) * | 2023-03-02 | 2023-03-28 | 安徽星辰智跃科技有限责任公司 | Functional environment adjusting method, system and device based on user mood |
| CN117930213A (en) * | 2024-03-25 | 2024-04-26 | 湘江实验室 | Method for detecting human existence based on millimeter wave radar |
| CN117930213B (en) * | 2024-03-25 | 2024-06-04 | 湘江实验室 | Method for detecting human existence based on millimeter wave radar |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110118966A (en) * | 2019-05-28 | 2019-08-13 | 长沙莫之比智能科技有限公司 | Personnel's detection and number system based on millimetre-wave radar |
| US20200116850A1 (en) * | 2018-10-16 | 2020-04-16 | Infineon Technologies Ag | Estimating Angle of Human Target Using mmWave Radar |
| CN111830508A (en) * | 2020-06-23 | 2020-10-27 | 北京航空航天大学 | Barrier gate anti-smashing system and method adopting millimeter wave radar |
-
2020
- 2020-11-05 CN CN202011220695.6A patent/CN114325623B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200116850A1 (en) * | 2018-10-16 | 2020-04-16 | Infineon Technologies Ag | Estimating Angle of Human Target Using mmWave Radar |
| CN110118966A (en) * | 2019-05-28 | 2019-08-13 | 长沙莫之比智能科技有限公司 | Personnel's detection and number system based on millimetre-wave radar |
| CN111830508A (en) * | 2020-06-23 | 2020-10-27 | 北京航空航天大学 | Barrier gate anti-smashing system and method adopting millimeter wave radar |
Non-Patent Citations (1)
| Title |
|---|
| 贺峰;黄晓涛;刘承兰;李欣;周智敏;: "行进人体目标雷达瞬时多普勒特征分析", 信号处理, no. 09, 30 September 2010 (2010-09-30) * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114563784A (en) * | 2022-04-27 | 2022-05-31 | 武汉理工大学 | Shipborne environment intrusion detection method and system based on double millimeter wave radar |
| CN114563784B (en) * | 2022-04-27 | 2022-08-09 | 武汉理工大学 | Shipborne environment intrusion detection method and system based on double millimeter wave radar |
| CN115857595A (en) * | 2023-03-02 | 2023-03-28 | 安徽星辰智跃科技有限责任公司 | Functional environment adjusting method, system and device based on user mood |
| CN117930213A (en) * | 2024-03-25 | 2024-04-26 | 湘江实验室 | Method for detecting human existence based on millimeter wave radar |
| CN117930213B (en) * | 2024-03-25 | 2024-06-04 | 湘江实验室 | Method for detecting human existence based on millimeter wave radar |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114325623B (en) | 2024-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Yang et al. | Dense people counting using IR-UWB radar with a hybrid feature extraction method | |
| Chang et al. | An algorithm for UWB radar-based human detection | |
| Nag et al. | Preliminary interferometric images of moving targets obtained using a time-modulated ultra-wide band through-wall penetration radar | |
| Zheng et al. | A target detection scheme with decreased complexity and enhanced performance for range-Doppler FMCW radar | |
| CN114325623B (en) | Human body limb movement information measurement method based on millimeter wave radar | |
| Zhang et al. | Enhanced detection of Doppler-spread targets for FMCW radar | |
| CN110187342B (en) | Vital sign detection and imaging method based on FMCW mobile platform | |
| CN112731380A (en) | Intelligent human body monitoring method and monitoring equipment based on millimeter waves | |
| Tan et al. | A new method for multipath interference suppression in through-the-wall UWB radar imaging | |
| Cai et al. | Foreign objects intrusion detection using millimeter wave radar on railway crossings | |
| Gao et al. | Learning to detect open carry and concealed object with 77 GHz radar | |
| Aristov et al. | Ultra-wideband pulse radar with discrete stroboscopic receiver for detection of small targets behind dielectric obstacles | |
| Dubey et al. | Generative adversial network based extended target detection for automotive mimo radar | |
| Vlasak et al. | The analysis of small RCS target detection in primary radar system | |
| CN113866756A (en) | Small unmanned aerial vehicle target tracking method based on MIMO radar | |
| Zhang et al. | Comparison of impulse radar and spread-spectrum radar in through-wall imaging | |
| Park et al. | Length prediction of moving vehicles using a commercial FMCW radar | |
| Chen et al. | Radar signal processing for low-observable marine target-challenges and solutions | |
| Kocur et al. | Imaging method: A strong tool for moving target tracking by a multistatic UWB radar system | |
| Lee et al. | Reduction of false alarm rate in SAR-MTI based on weighted kurtosis | |
| CN115079118A (en) | Millimeter wave radar gait recognition method based on retrieval task | |
| CN112835006B (en) | Method and system for detecting small targets on sea by tracking radar based on interframe accumulation | |
| Huang et al. | Regional people counting approach based on multi-motion states models using MIMO radar | |
| Feng et al. | Detection of human targets using ultra-wide band through-the-wall radar | |
| Mimouna et al. | LSTM-based system for multiple obstacle detection using ultra-wide band radar |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |