CN109975797A - A kind of arm motion details cognitive method based on doppler radar signal - Google Patents
A kind of arm motion details cognitive method based on doppler radar signal Download PDFInfo
- Publication number
- CN109975797A CN109975797A CN201910285758.7A CN201910285758A CN109975797A CN 109975797 A CN109975797 A CN 109975797A CN 201910285758 A CN201910285758 A CN 201910285758A CN 109975797 A CN109975797 A CN 109975797A
- Authority
- CN
- China
- Prior art keywords
- signal
- arm
- radar
- radar signal
- arm motion
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/02—Systems using reflection of radio waves, e.g. primary radar systems; Analogous systems
- G01S13/50—Systems of measurement based on relative movement of target
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/28—Details of pulse systems
- G01S7/285—Receivers
- G01S7/292—Extracting wanted echo-signals
- G01S7/2923—Extracting wanted echo-signals based on data belonging to a number of consecutive radar periods
- G01S7/2927—Extracting wanted echo-signals based on data belonging to a number of consecutive radar periods by deriving and controlling a threshold value
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
- G01S7/411—Identification of targets based on measurements of radar reflectivity
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
- G01S7/415—Identification of targets based on measurements of movement associated with the target
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The present invention proposes that a kind of arm motion details cognitive method based on doppler radar signal refers to that the movements such as example lifting hand to common gesture behavior, wave, push and pull carries out fine granularity identification, specifically includes the big contents of information analyses two such as the differentiation of gesture type, operating angle amplitude or direction.We, which, by the combinatory analysis from the original signal of two radars, obtain the gesture classification that user does, is differentiated for gesture type.Its working contents includes that signal detection is extracted, identification is classified, conclusion generates.Arm motion details cognitive method proposed in this paper based on doppler radar signal without carrying out the training of mass data for the different movements in each angle, amplitude, direction, therefore has saved cost.Using multi-level cognitive method, i.e., first differentiate that type analyzes details again, can improve treatment effeciency and accuracy.Finally it can recognize that arm lifts the details action message such as direction of the angle put down, the amplitude that arm is brandished in front of body, push pull maneuver.
Description
Technical field
The present invention relates to the human body behaviors based on radio magnetic wave signal to perceive field more particularly to miniature Doppler radar
Fine granularity perception is carried out to the arm motion amplitude of user, angle etc. and knows method for distinguishing.
Background technique
With the increase of human-computer interaction demand and the development of technology, people are acted using wireless device more and more
Identify work.There is additional requirement for illumination different from the action identification method of computer vision, and is passed using acceleration
To people the case where having any problem in whole action recognition, the action recognition of wireless device has the action identification method of sensor perception
There is pervasive, easy-operating advantage.Simultaneously with the development of the smart machines such as smart phone, wireless device is also deep into life
Every aspect, popularity is significantly increased.At present there are many based on wireless cognition technology, such as in 2016
" the WiFinger:talk to your smart devices with finger-grained that UbiComp is delivered
Gesture " article using Wi-Fi signal pass through human body when its changed characteristic of CSI signal strength, come what is made to user
Gesture is identified, more natural human-computer interaction is realized;Patent US20120139708A1 then illustrates a kind of based on RFID's
Gesture identification method wears RFID radar physically by user and receives the RFID label tag transmission that user is worn on hand
Signal, obtain the spatial position of user's hand and then identify the gesture that user makes.However, in the existing method, Wi-
The distribution of Fi wireless signal is unstable, makes it difficult to be applied to actual conditions vulnerable to the characteristics of interference, and RFID identification technology then needs
Additional equipment, affect experience are worn to user.In addition, they are substantially for the purpose of identification maneuver type, it is few to examine
Consider the analysis of gesture detailed information.Light is the movement lifting hand or waving, and also has different lift hand angles, amplitude of waving etc.
Difference, and user is likely to wish to carry out some special human-computer interactions with this, such as fine tuning indoor light brightness.Radar signal
Have the advantages that low noise, bandwidth, loss are small, more accurate, stable knowledge can be made to the movement of user in identical environment
Not.The perception of human arm movement being carried out using miniature Doppler radar, not only identification maneuver type, also analysis acts details,
It can provide more preferable more convenient and fast man-machine interaction experience, can bring certain practical significance in fields such as smart home, work entertainments.
Summary of the invention
In view of the above problems, the present invention provides and a kind of can recognize that such as arm lifts the angle put down, arm in body
Human arm motion's details based on Doppler radar of the details action message such as direction of amplitude, push pull maneuver that front is brandished
Cognitive method.
The technical solution of the present invention is as follows: a kind of arm motion details cognitive method based on doppler radar signal, including
Following steps:
Step 1: collected radar signal first being backed up, is then filtered;
Step 2: filtered radar signal is based on, using rule-based double-threshold comparison algorithm, to two radars
Have in signal and fluctuate, it may be possible to which movement causes the signal segment generated to be detected and extracted;Meanwhile in step 1
The original radar signal backed up also extracts signal segment on same time point;
Step 3: the action signal that filtered two radars respectively extract being subjected to discrete wavelet variation drop and is adopted
Then sample uses dual-stage classification, obtain movement type preliminary conclusion;Preliminary conclusion include: class lift hand, class put down, analogize,
Class is drawn, class rotation, can not be judged;
Step 4: the movement type preliminary conclusion respectively obtained according to two radars designs conclusion rule of combination, completes tool
The arm action type of body judges;Specific conclusion include: forwards lift hand and put down, push away and draw forwards, to the right lift hand with
Put down, push away and draw to the right, move to right side in front of the body, from right side of body translate back in front of, in front of body it is clockwise
Rotation, rotates counterclockwise in front of body;
Step 5: movement category identification after the completion of, if the movement belong to lift hand down, push and pull, translating in any one,
Then continue to carry out step backward;If the movement belongs to rotation classification, further identification will be stopped;
Step 6: the action signal that two radars in step 3 without filtering are respectively extracted generates spectrogram;
Step 7: usable floor area comparison method analyzes spectrogram, obtains Doppler's frequency that two radars respectively detect
Ratio is moved, to reflect speed ratio of the arm motion on two components;
Step 8: arctan function being used to Ratio of Doppler Shift, obtains angle, amplitude or the direction tool of arm motion
Body detailed information.
Further, a kind of arm motion details cognitive method based on doppler radar signal, it is two-door in the step 2
Limit detection algorithm, which uses, crosses threshold rate and short-time energy in short-term as threshold value progress two-stage judgement.
Further, a kind of arm motion details cognitive method based on doppler radar signal first carries out signal
Framing calculates separately each frame and crosses threshold rate and short-time energy in short-term;Crossing threshold rate formula in short-term is
Short-time energy isWherein i represents frame number, and T represents threshold value, and threshold value TZ, TE is set separately;
Only the threshold rate excessively in short-term of continuous multiple frames is more than TZ, and the short-time energy summation of these frames is more than TE, is just determined as
Effective action signal.
Further, a kind of arm motion details cognitive method based on doppler radar signal, as long as there is any one
Radar produces qualified signal according to double-threshold comparison method, and no matter at this time whether another radar detects in the time zone
To qualified signal, then equal acts of determination detection comes into force.
Further, a kind of arm motion details cognitive method based on doppler radar signal, by repeatedly discrete small
Wave variation is by an action signal sequence length control between 160-320 point.
Further, a kind of arm motion details cognitive method based on doppler radar signal, in dual-stage classification side
In the first stage of method, need that hand signal is first divided into close, separate two major classes;In the second-order of dual-stage classification method
The classification of motion is further given category, including class lift hand, class are put down, analogize, class according to the conclusion of previous stage by Duan Zhong
It draws, class rotation, can not identify.
Further, a kind of arm motion details cognitive method based on doppler radar signal, calculates separately two thunders
Darker regions area (area) up to frequency spectrum subtracts base area (offset);Then by the calculated value of front radar divided by side
The area value of radar is as area ratio, formulaWherein A and B respectively represents the thunder of a front surface and a side surface
It reaches, this area ratio is equivalent to the Ratio of Doppler Shift that two detections of radar arrive.
Further, a kind of arm motion details cognitive method based on doppler radar signal, by the ratio acquired into
Row arctangent computation goes out angle, and formula is θ=arctan (r);For lift hand and movement is put down, θ represents arm and lifts direction
With the angle in direction immediately ahead of body;Push and pull is acted, θ represents direction immediately ahead of the direction that arm pushs out and body
Angle;Movement is moved horizontally for arm, θ represents the amplitude of horizontal movement before arm stop motion.
Arm motion details perception based on doppler radar signal is referred to for example lifting hand to common gesture behavior, be waved
The movements such as hand, push-and-pull carry out fine granularity identification, specifically include the information analyses such as the differentiation of gesture type, operating angle amplitude or direction
Two big contents.We, which, by the combinatory analysis from the original signal of two radars, obtain the hand that user does, is differentiated for gesture type
Gesture classification.Its working contents includes that signal detection is extracted, identification is classified, conclusion generates.For operating angle amplitude direction
We have proposed a kind of methods based on spectrogram for analysis, obtain the detail information for this gesture that user does, such as lift
Hand angle, amplitude of waving etc..Its working contents includes the conversion of signal spectrum figure, frequency than calculating, conclusion generation.Its intermediate frequency
Rate proposes area comparison method calculating thinking than calculating link, it can be deduced that the movement velocity ratio of arm action in the two-dimensional direction
Value, so that indirect analysis goes out the angle moved or amplitude information.Arm fortune proposed in this paper based on doppler radar signal
Dynamic details cognitive method without carrying out the training of mass data for the different movements in each angle, amplitude, direction, therefore is saved
Cost.In addition, using multi-level cognitive method, i.e., first differentiate that type analyzes details again, treatment effeciency and accurate can be improved
Property.Finally it can recognize that such as arm lifts the direction of the angle put down, the amplitude that arm is brandished in front of body, push pull maneuver
Etc. details action message.
Detailed description of the invention
Fig. 1 is a kind of flow chart of the arm motion details cognitive method based on doppler radar signal of the present invention;
Fig. 2 is that a kind of experimental provision of the arm motion details cognitive method based on doppler radar signal of the present invention is overlooked
Schematic diagram;
Fig. 3 is a kind of having handled through small echo for arm motion details cognitive method based on doppler radar signal of the invention
Complete signal;
Fig. 4 is that a kind of parabola of the arm motion details cognitive method based on doppler radar signal of the present invention constrains effect
Fruit figure;
Fig. 5 is that rule are combined in a kind of perception of the arm motion details cognitive method based on doppler radar signal of the present invention
Then;
Fig. 6 is a kind of perceived spectral figure of the arm motion details cognitive method based on doppler radar signal of the present invention.
Specific embodiment
Further describe technical solution of the present invention with reference to the accompanying drawing:
Two identical small-sized 24GHz doppler radar sensors are placed on about 1.4 meters of high positions from the ground, are passed through
USB data line is connected to computer;One of radar is placed on about 1.5 meters immediately ahead of user of position, another radar is placed on
The position that about 1.5 meters of user's front-right, two radars are towards user.Entire experiment schematic top plan view is as shown in Figure 2.Radar passes
Sensor is by the arm behavior of perception user, to the time-domain signal of computer transmission binary channels I, Q, sample rate 44100Hz, and two
Radar one, which is met together, generates two groups of I, Q signals.Computer then these data of real-time collecting, carry out latter step processing.
Step 1: collected radar signal first being backed up, is then filtered, reduces noise to movement kind
The interference of alanysis conclusion.Since it is considered that filter high frequency noise to greatest extent is needed, using I type Chebyshev's low pass filtered
Wave device.Parameters setting value is cut-off frequecy of passband 40Hz, stopband cutoff frequency 75Hz, sideband region decaying in filter
0.1dB, cut-off region decaying 30dB, sample rate, that is, radar sensor output frequency is 44100Hz.
Step 2: filtered radar signal is based on, using rule-based double-threshold comparison algorithm, to two radars
Have in signal and fluctuate, it may be possible to which movement causes the signal segment generated to be detected and extracted.At the same time, in step
The original radar signal backed up in 2 also extracts signal segment on same time point, i.e., obtains altogether so far following
Signal: the respective I of filtered two radars, Q signal segment and the respective I of two radars, Q signal piece without filtering
Section.Double-threshold comparison algorithm, which has used, crosses threshold rate and short-time energy in short-term as threshold value progress two-stage judgement.First to signal into
Row framing calculates separately each frame and crosses threshold rate and short-time energy in short-term.Crossing threshold rate formula in short-term is
Short-time energy isWherein i represents frame number, and T represents threshold value and is set as 15, and is set separately
Threshold value TZ=10, TE=10.Only the threshold rate excessively in short-term of continuous multiple frames is more than TZ, and the short-time energy summation of these frames is super
TE is crossed, is just determined as effective action signal.Further, since there are horizontal direction check frequencies for Doppler radar, therefore
Some gesture motions can only may be arrived by a detections of radar, rule-based detection method presented further herein.That is: as long as
There is any one radar to produce qualified signal according to double-threshold comparison method, no matter at this time whether another radar is in the time
Qualified signal is detected in region, then equal acts of determination detection comes into force.
Step 3: the action signal that filtered two radars respectively extract being subjected to discrete wavelet variation drop and is adopted
One action signal sequence length is controlled between 160-320 point by the variation of multiple discrete wavelet, is subsequent meter by sample
Calculating, which reduces complexity, improves efficiency.The signal that is disposed as shown in Figure 3 (note: it is more intuitive in order to show, front and back is remained in figure
Extra signal is to observe, the signal segment of the only central marker actually get).
Then dual-stage classification is used, obtains movement type preliminary conclusion.Preliminary conclusion includes: that class lift hand (refers to similar
In the movement of lift hand, afterwards together), class is put down, analogizes, class is drawn, class rotation, can not judge.Due to the rule design of step 2, if
The signal segment is inherently underproof, then being directly classified as can not judge, is otherwise identified using dual-stage classification.
In the first stage of dual-stage classification method, need that hand signal is first divided into close, separate two major classes.?
In the first stage of dual-stage classification method, need that hand signal is first divided into close, separate two major classes.According to Doppler's thunder
Up to working principle, it can be simply considered that the expression formula of orthogonal demodulation signal I, Q areWherein A representation signal intensity, fdIt represents more
General Le frequency displacement,Representing leads to the initial phase generated by distance between user and radar.Thus two signal phases can be calculated
Poor θ=tan-1The π vt/ λ of Q (t)/I (t)=4, wherein v represent arm motion speed (taken just when close to radar motion, it is on the contrary
It takes negative).Next the positive negativity of v can be obtained indirectly according to the value variation tendency of θ and concludes therefrom that the direction of motion is proximate to
Or it is separate.If growth trend is presented in θ at any time, illustrate this gesture motion be close to radar, it is on the contrary then separate
Radar.In addition, it is dull continuous in open interval (- pi/2, pi/2) due to arctan function in this step, it will appear at interval endpoint
SPA sudden phase anomalies are needed to obtaining phase difference and do phase unwrapping to restore phase.Expansion formula is θu,i=θu,i-1+mod(θw,i-
θw,i-1-π,2π)+π。
It is further to refer to by the classification of motion according to the conclusion of previous stage in the second stage of dual-stage classification method
Determine type.It is further specified by the classification of motion according to the conclusion of previous stage in the second stage of dual-stage classification method
Type, including class lift hand, class are put down, analogize, class is drawn, class rotation, can not identify.Use dynamic time warping algorithm (DTW)+neighbour
Nearly algorithm (kNN) mode carries out proximity matching, so that action signal be classified.Wherein DTW is carried out on the basis of primal algorithm
The improvement of constrained path.Parabola the way of restraint is specifically used, i.e., in the distance matrix of two bars, with two parabolas
Matrix area is divided as boundary, limitation coupling path is in intermediate one piece of region and can not cross the border, so that apart from meter
It is more reasonable accurate to calculate, while also improving efficiency.Parabola constrained path formula is recommended to use 2nx2/3m2+nx/3m2-10≤y
≤-2nx2/3m2+5nx/3m2+ 10 wherein m, n respectively represent the length of two signals compared.Binding effect figure is shown in Fig. 4.
Step 4: the movement type preliminary conclusion respectively obtained according to two radars designs conclusion rule of combination, completes tool
The arm action type of body judges.Rule of combination is as shown in Figure 5.Specific conclusion includes: to lift hand forwards and put down, push away forwards
With draw, lift and is put down hand to the right, push away and draw to the right, move to right side in front of body, from right side of body translate back in front of,
It rotates clockwise in front of body, is rotated counterclockwise in front of body.For example, when radar conclusion in front is " class lift hand ", side
Radar conclusion is " class is put down ", then practical arm action belongs to " moving to right in front of from body ".
Step 5: movement category identification after the completion of, if the movement belong to lift hand down, push and pull, translating in any one,
Then continue to carry out step backward.If the movement belongs to rotation classification, it is recognized herein that action recognition granularity has little significance at this time, it will
Stop further identification.
Step 6: the action signal that two radars in step 3 without filtering are respectively extracted generates spectrogram.
In view of during category identification filtering and discrete wavelet variation etc. operations lost a large amount of detailed information, it is therefore necessary to
Original signal segment is reused, i.e., without the data of filtering, directly carries out the generation of FFT spectrum figure, as shown in Figure 6 (note: for
Expression is more intuitive, the spectrogram of continuous three action signals is illustrated in figure, it can be observed that the pole of the movement of different directions
It is worth different).Each radar can generate a similar amplitude-frequency spectrogram.
Step 7: usable floor area comparison method analyzes spectrogram, obtains Doppler's frequency that two radars respectively detect
Ratio is moved, to reflect speed ratio of the arm motion on two components.It is marked according to Fig. 6, calculates separately two radar frequencies
The darker regions area (area) of spectrum subtracts base area (offset).Then by the calculated value of front radar divided by side radars
Area value as area ratio, formula isWherein A and B respectively represents the radar of a front surface and a side surface.
Step 8: arctan function being used to Ratio of Doppler Shift, obtains angle, amplitude or the direction tool of arm motion
Body detailed information.Formula is θ=arctan (r).For lift hand and movement is put down, θ represents arm and lifting direction and body just
The angle of forward direction;Push and pull is acted, θ represents the angle in direction immediately ahead of the direction that arm pushs out and body;
Movement is moved horizontally for arm, θ represents the amplitude of horizontal movement before arm stop motion.Realize arm herein as a result,
The detailed information of movement is analyzed.
Claims (8)
1. a kind of arm motion details cognitive method based on doppler radar signal, it is characterised in that: the following steps are included:
Step 1: collected radar signal first being backed up, is then filtered;
Step 2: filtered radar signal is based on, using rule-based double-threshold comparison algorithm, to the signal of two radars
In have and fluctuate, it may be possible to movement causes the signal segment generated to be detected and extracted;Meanwhile it backing up in step 1
The original radar signal crossed also extracts signal segment on same time point;
Step 3: the action signal progress discrete wavelet variation that filtered two radars are respectively extracted is down-sampled, so
Dual-stage classification is used afterwards, obtains movement type preliminary conclusion;Preliminary conclusion include: class lift hand, class are put down, analogize, class is drawn,
Class rotation can not judge;
Step 4: the movement type preliminary conclusion respectively obtained according to two radars designs conclusion rule of combination, completes specific
The judgement of arm action type;Specific conclusion includes: to lift hand forwards and put down, push away forwards and draw, lift hand to the right and put down,
It pushes away and draws to the right, move to right side in front of body, translate go back to front from right side of body, rotated clockwise in front of body,
It is rotated counterclockwise in front of body;
Step 5: movement category identification after the completion of, if the movement belong to lift hand down, push and pull, translating in any one, after
It is continuous to carry out step backward;If the movement belongs to rotation classification, further identification will be stopped;
Step 6: the action signal that two radars in step 3 without filtering are respectively extracted generates spectrogram;
Step 7: usable floor area comparison method analyzes spectrogram, obtains the Doppler frequency shift that two radars respectively detect
Than to reflect speed ratio of the arm motion on two components;
Step 8: arctan function being used to Ratio of Doppler Shift, show that the angle, amplitude or direction of arm motion are specifically thin
Save information.
2. a kind of arm motion details cognitive method based on doppler radar signal according to claim 1, feature
Be: double-threshold comparison algorithm, which uses, in the step 2 crosses threshold rate and short-time energy in short-term as threshold value progress two-stage judgement.
3. a kind of arm motion details cognitive method based on doppler radar signal according to claim 2, feature
It is: framing is carried out to signal first, each frame is calculated separately and crosses threshold rate and short-time energy in short-term;Threshold rate is crossed in short-term
Formula is
Short-time energy isWherein i represents frame number, and T represents threshold value, and threshold value TZ, TE is set separately;Only
Continuous multiple frames cross threshold rate more than TZ in short-term, and the short-time energy summation of these frames is more than TE, is just determined as effectively
Action signal.
4. a kind of arm motion details cognitive method based on doppler radar signal according to claim 3, feature
Be: as long as there is any one radar to produce qualified signal according to double-threshold comparison method, no matter at this time another radar is
No that qualified signal is detected in the time zone, then equal acts of determination detection comes into force.
5. a kind of arm motion details cognitive method based on doppler radar signal according to claim 1, feature
It is: is controlled an action signal sequence length between 160-320 point by the variation of multiple discrete wavelet.
6. a kind of arm motion details cognitive method based on doppler radar signal according to claim 1, feature
It is: in the first stage of dual-stage classification method, needs that hand signal is first divided into close, separate two major classes;Double
It is further given category, including class by the classification of motion according to the conclusion of previous stage in the second stage of Stage Classification method
Lift hand, class are put down, analogize, class is drawn, class rotation, can not identify.
7. a kind of arm motion details cognitive method based on doppler radar signal according to claim 1, feature
Be: the darker regions area (area) for calculating separately two radar frequency spectrums subtracts base area (offset);It then will be positive
Divided by the area value of side radars as area ratio, formula is the calculated value of radarWherein A and B points
The radar of a front surface and a side surface is not represented, this area ratio is equivalent to the Ratio of Doppler Shift that two detections of radar arrive.
8. a kind of arm motion details cognitive method based on doppler radar signal according to claim 1, feature
It is: the ratio acquired progress arctangent computation is gone out into angle, formula is θ=arctan (r);For lift hand and put down movement, θ
Represent the angle that arm lifts direction immediately ahead of direction and body;Push and pull is acted, θ represents the side that arm pushs out
To the angle with direction immediately ahead of body;Movement moved horizontally for arm, θ represents horizontal movement before arm stop motion
Amplitude.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910285758.7A CN109975797A (en) | 2019-04-10 | 2019-04-10 | A kind of arm motion details cognitive method based on doppler radar signal |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910285758.7A CN109975797A (en) | 2019-04-10 | 2019-04-10 | A kind of arm motion details cognitive method based on doppler radar signal |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109975797A true CN109975797A (en) | 2019-07-05 |
Family
ID=67083976
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910285758.7A Pending CN109975797A (en) | 2019-04-10 | 2019-04-10 | A kind of arm motion details cognitive method based on doppler radar signal |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109975797A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110347260A (en) * | 2019-07-11 | 2019-10-18 | 歌尔科技有限公司 | A kind of augmented reality device and its control method, computer readable storage medium |
CN110412566A (en) * | 2019-07-22 | 2019-11-05 | 西北工业大学 | A kind of fine granularity human arm motion's recognition methods based on Doppler radar time and frequency domain characteristics |
CN111753678A (en) * | 2020-06-10 | 2020-10-09 | 西北工业大学 | Multi-device cooperative gait perception and identity recognition method based on ultrasonic waves |
CN113591684A (en) * | 2021-07-29 | 2021-11-02 | 北京富奥星电子技术有限公司 | Gesture recognition method based on Doppler radar of CW system |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2011099703A (en) * | 2009-11-04 | 2011-05-19 | Oki Electric Industry Co Ltd | Data processor, motion recognition system, method of motion determination, and program |
CN103793059A (en) * | 2014-02-14 | 2014-05-14 | 浙江大学 | Gesture recovery and recognition method based on time domain Doppler effect |
CN104094194A (en) * | 2011-12-09 | 2014-10-08 | 诺基亚公司 | Method and apparatus for identifying a gesture based upon fusion of multiple sensor signals |
CN105786185A (en) * | 2016-03-12 | 2016-07-20 | 浙江大学 | Non-contact type gesture recognition system and method based on continuous-wave micro-Doppler radar |
CN107132512A (en) * | 2017-03-22 | 2017-09-05 | 中国人民解放军第四军医大学 | UWB radar human motion micro-Doppler feature extracting method based on multichannel HHT |
CN107526437A (en) * | 2017-07-31 | 2017-12-29 | 武汉大学 | A kind of gesture identification method based on Audio Doppler characteristic quantification |
CN108371545A (en) * | 2018-02-02 | 2018-08-07 | 西北工业大学 | A kind of human arm action cognitive method based on Doppler radar |
CN109583436A (en) * | 2019-01-29 | 2019-04-05 | 杭州朗阳科技有限公司 | A kind of gesture recognition system based on millimetre-wave radar |
-
2019
- 2019-04-10 CN CN201910285758.7A patent/CN109975797A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2011099703A (en) * | 2009-11-04 | 2011-05-19 | Oki Electric Industry Co Ltd | Data processor, motion recognition system, method of motion determination, and program |
CN104094194A (en) * | 2011-12-09 | 2014-10-08 | 诺基亚公司 | Method and apparatus for identifying a gesture based upon fusion of multiple sensor signals |
CN103793059A (en) * | 2014-02-14 | 2014-05-14 | 浙江大学 | Gesture recovery and recognition method based on time domain Doppler effect |
CN105786185A (en) * | 2016-03-12 | 2016-07-20 | 浙江大学 | Non-contact type gesture recognition system and method based on continuous-wave micro-Doppler radar |
CN107132512A (en) * | 2017-03-22 | 2017-09-05 | 中国人民解放军第四军医大学 | UWB radar human motion micro-Doppler feature extracting method based on multichannel HHT |
CN107526437A (en) * | 2017-07-31 | 2017-12-29 | 武汉大学 | A kind of gesture identification method based on Audio Doppler characteristic quantification |
CN108371545A (en) * | 2018-02-02 | 2018-08-07 | 西北工业大学 | A kind of human arm action cognitive method based on Doppler radar |
CN109583436A (en) * | 2019-01-29 | 2019-04-05 | 杭州朗阳科技有限公司 | A kind of gesture recognition system based on millimetre-wave radar |
Non-Patent Citations (3)
Title |
---|
XINYE LOU ET AL.: "Gesture-Radar: Enabling Natural Human-Computer Interactions with Radar-Based Adaptive and Robust Arm Gesture Recognition ", 《2018 IEEE INTERNATIONAL CONFERENCE ON SYSTEMS, MAN, AND CYBERNETICS》 * |
刘熠辰等: "基于雷达技术的手势识别", 《中国电子科学研究院学报》 * |
鲁勇等: "基于WiFi信号的人体行为感知技术研究综述", 《计算机学报》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110347260A (en) * | 2019-07-11 | 2019-10-18 | 歌尔科技有限公司 | A kind of augmented reality device and its control method, computer readable storage medium |
CN110347260B (en) * | 2019-07-11 | 2023-04-14 | 歌尔科技有限公司 | Augmented reality device, control method thereof and computer-readable storage medium |
CN110412566A (en) * | 2019-07-22 | 2019-11-05 | 西北工业大学 | A kind of fine granularity human arm motion's recognition methods based on Doppler radar time and frequency domain characteristics |
CN111753678A (en) * | 2020-06-10 | 2020-10-09 | 西北工业大学 | Multi-device cooperative gait perception and identity recognition method based on ultrasonic waves |
CN113591684A (en) * | 2021-07-29 | 2021-11-02 | 北京富奥星电子技术有限公司 | Gesture recognition method based on Doppler radar of CW system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109975797A (en) | A kind of arm motion details cognitive method based on doppler radar signal | |
CN108371545B (en) | Human body arm action sensing method based on Doppler radar | |
US11256335B2 (en) | Fine-motion virtual-reality or augmented-reality control using radar | |
JP2023021967A (en) | Gesture recognition with sensor | |
Liu et al. | M-gesture: Person-independent real-time in-air gesture recognition using commodity millimeter wave radar | |
Li et al. | Towards domain-independent and real-time gesture recognition using mmwave signal | |
CN105807935B (en) | A kind of gesture control man-machine interactive system based on WiFi | |
CN106899968A (en) | A kind of active noncontact identity identifying method based on WiFi channel condition informations | |
CN110412566A (en) | A kind of fine granularity human arm motion's recognition methods based on Doppler radar time and frequency domain characteristics | |
CN104699241B (en) | It is determined that the method and apparatus at action and/or action position | |
Liu et al. | Spectrum-based hand gesture recognition using millimeter-wave radar parameter measurements | |
CN110059612A (en) | A kind of gesture identification method and system that the position based on channel state information is unrelated | |
Zinnen et al. | An analysis of sensor-oriented vs. model-based activity recognition | |
CN110866468A (en) | Gesture recognition system and method based on passive RFID | |
CN113051972A (en) | Gesture recognition system based on WiFi | |
CN113064483A (en) | Gesture recognition method and related device | |
Pan et al. | Dynamic hand gesture detection and recognition with WiFi signal based on 1d-CNN | |
Guendel et al. | Phase-based classification for arm gesture and gross-motor activities using histogram of oriented gradients | |
CN116343261A (en) | Gesture recognition method and system based on multi-modal feature fusion and small sample learning | |
CN109189219A (en) | The implementation method of contactless virtual mouse based on gesture identification | |
Dong et al. | Review of research on gesture recognition based on radar technology | |
Kabir et al. | CSI-DeepNet: A lightweight deep convolutional neural network based hand gesture recognition system using Wi-Fi CSI signal | |
Qin et al. | WiASL: American sign language writing recognition system using commercial WiFi devices | |
CN112380903B (en) | Human body activity recognition method based on WiFi-CSI signal enhancement | |
CN113449711A (en) | Micro Doppler image sign language perception identification method based on direction density characteristics |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190705 |