Resting human body identification method based on civil millimeter wave radar and related device
The multi-frame chirped signal was transmitted through civilian millimeter wave radar, the sum of the maximum Doppler speeds of each frame was calculated, and the distance fast Fourier transform was performed, which solved the problem of difficulty in recognizing the resting human body and achieved efficient and accurate recognition effect.
Patent Information
- Application Number
- CN202311574063.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-11-23
- Publication Date
- 2025-05-23
AI Technical Summary
The prior art has difficulties in identifying resting human bodies, especially when Doppler motion is low, traditional methods require a large amount of data and complex calculations for model training, and the calculation is large.
Multi-frame chirped signals are transmitted through civilian millimeter wave radar, the sum of the maximum Doppler velocities of each frame is calculated, and the distance fast Fourier transform is performed to accumulate inter-frame sampling data to identify the resting human body.
It realizes the identification of resting human bodies under less data volume and simple calculation conditions, which improves the accuracy and efficiency of recognition and reduces the calculation amount.
Smart Images
Figure CN120028764A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of civil radar detection technology, and in particular to a method and related device for identifying a stationary human body based on a civil millimeter wave radar. Background Art
[0002] Millimeter-wave radar is applied to civilian indoor human detection, which can monitor elderly people living alone and people in public places. Since the human body reflects radar signals weakly, the radar signals reflected by the human body in complex indoor scenes are easily obliterated by environmental signals. Therefore, the traditional signal processing method is to first remove the DC in the Doppler domain of the reflected radar signal, and then analyze the signal after DC removal, so that it is easier to extract the radar signal reflected by the active human body.
[0003] When the human body is in a resting state such as sleeping or sitting still, the Doppler motion of the human body is very low. When performing Doppler domain DC removal, the radar signal reflected by the human body will also be removed, making it difficult to identify a resting human body. Because the human body has low Doppler frequency movements such as breathing and heartbeat in a resting state, some solutions will extract the radar's micro-Doppler signal, extract features from the extracted micro-Doppler signal, and then classify based on the extracted features through neural networks or support vector machines, and finally identify whether it is a resting human body.
[0004] However, the recognition of static human bodies through support vector machines and neural networks requires model training in the early stage, which requires a large amount of data and high computational complexity. Moreover, when the radar signal reflected by the human body disappears, it is necessary to search all the detectable range of the radar and then extract the features, which requires even greater computation. Summary of the invention
[0005] In view of this, the present disclosure provides a method and related device for identifying a stationary human body based on a civilian millimeter-wave radar, aiming to achieve the identification of a stationary human body with less data and simpler calculations.
[0006] According to a first aspect of the present disclosure, a method for identifying a static human body based on a civilian millimeter wave radar is provided, wherein the civilian millimeter wave radar transmits a plurality of frames and each frame includes a plurality of chirp signals, and the method for identifying a static human body includes:
[0007] Determine the maximum Doppler velocity of the effective target at each sampling distance in the frame through the echo data of multiple chirp signals in the same frame, and calculate the sum of the maximum Doppler velocity of each frame at the same sampling distance;
[0008] Performing range fast Fourier transform on echo data of chirp signals of multiple frames and accumulating them to obtain inter-frame sampling data;
[0009] Based on the inter-frame sampling data, a stationary human body is identified from each sampling distance where the sum of the maximum Doppler velocities of each frame is not greater than a threshold.
[0010] Optionally, a sampling distance at which the sum of the maximum Doppler velocities of each frame is not greater than a threshold is a sampling distance to be processed, and based on the inter-frame sampling data, identifying a stationary human body from the sampling distance to be processed includes:
[0011] Find the frame whose Doppler velocity is greater than zero at the sampling distance to be processed, and obtain the frame to be processed;
[0012] Resetting the data corresponding to the sampling distance to be processed and the frame to be processed in the inter-frame sampling data to zero to obtain processed inter-frame sampling data;
[0013] Based on the processed inter-frame sampling data, a stationary human body is identified from the sampling distance to be processed.
[0014] Optionally, based on the processed inter-frame sampling data, identifying a stationary human body from the to-be-processed sampling distance includes:
[0015] Performing Doppler fast Fourier transform on the processed inter-frame sampling data to obtain velocity information and then generate a first range Doppler map including the velocity information;
[0016] Calculate the average value of each data amplitude at the same sampling distance in the inter-frame sampling data to obtain the average energy at each sampling distance;
[0017] In the constant false alarm rate calculation process, a peak in the first range Doppler map whose intensity value is greater than a threshold of the constant false alarm rate and higher than the average energy at the sampling distance is determined as the presence of a resting human body, wherein the position of the peak represents the distance information and speed information of the resting human body.
[0018] Optionally, the method for identifying a stationary human body further includes: calculating an arrival angle of the identified stationary human body.
[0019] Optionally, determining the maximum Doppler velocity of a valid target at each sampling distance in the frame through echo data of a plurality of chirp signals in the same frame includes:
[0020] Performing range fast Fourier transform on echo data of multiple chirp signals in the same frame and accumulating them to obtain intra-frame sampling data;
[0021] Performing Doppler fast Fourier transform on the intra-frame sampling data to obtain speed information indicating the maximum Doppler speed and then generating a second range Doppler map including the speed information;
[0022] In the constant false alarm rate calculation process, a valid target of the frame is identified from the second range Doppler map and the speed information corresponding to the valid target in the second range Doppler map is determined as the maximum Doppler speed at the sampling distance of the valid target in the frame.
[0023] Optionally, determining the maximum Doppler velocity of a valid target at each sampling distance in the frame through echo data of multiple chirp signals in the same frame further includes: recording the maximum Doppler velocity through a velocity matrix, wherein the dimension of the velocity matrix is determined by the number of sampling distances and the number of frames;
[0024] Calculating the sum of the maximum Doppler velocities of each frame at the same sampling distance includes: calculating the sum of each element corresponding to the same sampling distance in the velocity matrix.
[0025] Optionally, each frame transmitted by the civil millimeter-wave radar includes two subframes, the first subframe of the two subframes includes multiple chirp signals, the period of the first subframe is less than 1 / 2 of the frame period, and the second subframe of the two subframes is transmitted at 1 / 2 of the frame period, and the second subframe transmitted by each antenna includes only one chirp signal.
[0026] Optionally, determining the maximum Doppler velocity of the valid target at each sampling distance in the frame through echo data of multiple chirp signals in the same frame includes: determining the maximum Doppler velocity of the valid target at each sampling distance in the frame through echo data of multiple chirp signals in the first subframe in a frame;
[0027] Performing a range fast Fourier transform on the echo data of the chirp signals of multiple frames and accumulating them to obtain inter-frame sampling data, including: performing a range fast Fourier transform on the echo data of the first chirp signal of the first subframe in the multiple frames and a chirp signal in the second subframe and accumulating them to obtain the inter-frame sampling data.
[0028] According to a second aspect of the present disclosure, an electronic device is provided, comprising: a processor, a memory, and a program stored in the memory and executable on the processor, wherein when the program is executed by the processor, any one of the resting human body recognition methods described in the first aspect is implemented.
[0029] According to a third aspect of the present disclosure, a computer-readable storage medium is provided, on which a computer program or instruction is stored. When the computer program or instruction is executed by a processor, any one of the resting human body recognition methods described in the first aspect is implemented.
[0030] Beneficial effects of the present disclosure:
[0031] The present disclosure provides a method for identifying a static human body based on a civil millimeter wave radar, comprising: determining the maximum Doppler velocity of a valid target at each sampling distance in the frame through echo data of multiple chirp signals in the same frame, and calculating the sum of the maximum Doppler velocities of each frame at the same sampling distance; performing a distance fast Fourier transform on the echo data of the chirp signals of multiple frames and accumulating them to obtain inter-frame sampling data; based on the inter-frame sampling data, identifying a static human body from each sampling distance where the sum of the maximum Doppler velocities of each frame is not greater than a threshold. This static human body recognition method does not involve support vector machines and neural networks, and does not require a large amount of data and complex calculations for preliminary model training. Among them, the time span of the chirp signals of multiple frames for obtaining the inter-frame sampling data is large, so the speed resolution can meet the requirements of identifying a static human body based on movements such as breathing and heartbeat; and the sampling distances where the sum of the maximum Doppler velocities of each frame is greater than the threshold are targets with a large movement speed, so abandoning these sampling distances and directly identifying a static human body from the sampling distances where the sum of the Doppler velocities of each frame is not greater than the threshold can effectively reduce the amount of calculation.
[0032] It should be noted that the above general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] Figure 1 A flow chart showing a method for identifying a static human body according to an embodiment of the present disclosure is shown;
[0034] Figure 2 An exemplary inter-frame sampling data according to an embodiment of the present disclosure is shown;
[0035] Figure 3 A flow chart showing a method for identifying a static human body according to an embodiment of the present disclosure for generating a detection point cloud of a civil millimeter-wave radar is shown;
[0036] Figure 4 A schematic structural diagram of an electronic device according to another embodiment of the present disclosure is shown. DETAILED DESCRIPTION
[0037] In order to facilitate the understanding of the present disclosure, the present disclosure will be described more comprehensively below with reference to the relevant drawings. The preferred embodiments of the present disclosure are given in the drawings. However, the present disclosure can be implemented in different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present disclosure more thorough and comprehensive.
[0038] The signal transmission of millimeter wave radar is managed by frames. A frame can be composed of multiple subframes and each subframe includes one or more chirp signals. The signal transmitted by the millimeter wave radar will be reflected back after encountering the target and collected by the radar sensor as echo data. The target can be identified based on the echo data collected by the radar sensor.
[0039] At present, millimeter-wave radar is applied to civilian indoor human body detection. Since the radar reflection signal of the human body is very weak and the indoor environment is relatively complex, the human body is mainly detected through Doppler signals. For a static human body with a low movement speed, the Doppler signal is close to the indoor environment, and it is difficult to detect the human body through Doppler signals. In this case, when using millimeter-wave radar detection, it is necessary to combine neural networks or support vector machines, etc. Since the use of neural networks and support vector machines requires model training in the early stage, a large amount of data is required and the calculation complexity is high. Moreover, when the radar signal reflected by the human body disappears, it is necessary to search all the detectable ranges of the radar and then extract features, which requires a greater amount of calculation.
[0040] In view of this, an embodiment of the present disclosure provides a method for identifying a static human body based on a civilian millimeter-wave radar according to the radar Doppler principle, aiming to achieve static human body identification with less data and simpler calculation. The radar Doppler measurement principle is: in a short time, the movement of the target within a distance unit causes the echo phase difference of multiple chirp signals, and the resolution V of the radar Doppler velocity is doppler As shown in formula (1), λ is the wavelength of the chirp signal during propagation, N is the number of accumulated chirp signals, and T D is the time interval between adjacent chirp signals. Therefore, when the wavelength λ of the chirp signal during propagation remains unchanged, the number N of the accumulated multiple chirp signals and the time interval T between adjacent chirp signals are greatly increased by using chirp signals of different frames. D The Doppler velocity resolution can be improved by multiplying the time interval T between adjacent chirp signals. D Under the condition of being constant, the number N of the accumulated multiple chirp signals and the time interval T between adjacent chirp signals are D The product of is the time span of the accumulated multiple chirp signals. Therefore, when the time span of the accumulated multiple chirp signals is increased, the Doppler velocity resolution is improved. When the Doppler velocity resolution is improved, the breathing and heartbeat of the resting human body, which are slower movements, can be identified, that is, the resting human body can be identified.
[0041]
[0042] Figure 1The figure shows a flow chart of a static human body recognition method based on a civil millimeter wave radar provided by an embodiment of the present disclosure, wherein the civil millimeter wave radar transmits multiple frames and each frame includes multiple chirp signals. Figure 1 , the static human body recognition method based on civilian millimeter wave radar includes:
[0043] Step S110, determining the maximum Doppler velocity of the effective target at each sampling distance in the frame through the echo data of multiple chirp signals in the same frame, and calculating the sum of the maximum Doppler velocity of each frame at the same sampling distance.
[0044] Specifically, multiple sampling distances can be preset, for example, the sampling distance is set based on the range unit of the radar, and finally it is identified whether there is a target at an integer multiple of the range unit and the Doppler velocity of the target. The range unit of the radar refers to the minimum measurement used in the radar system to measure the distance between the target and the radar.
[0045] It should be noted that, through the echo data of multiple chirp signals in the same frame, multiple different Doppler speed targets at a sampling distance in the frame can be determined, and the above-mentioned maximum Doppler speed at a sampling distance in the frame is the maximum value of the Doppler speed among the multiple Doppler speed targets at the sampling distance in the frame.
[0046] In step S120, the echo data of the chirp signal of multiple frames are subjected to a range fast Fourier transform (FFT) and accumulated to obtain inter-frame sampling data.
[0047] Specifically, echo data of chirp signals of multiple frames are accumulated by performing range fast Fourier transform, that is, one or more echo data of chirp signals are taken from each frame and then stored in time series after performing range fast Fourier transform to form inter-frame sampling data.
[0048] Exemplarily, each frame includes two subframes, and the echo data of the first chirp signal of each subframe is taken and fast Fourier transformed and then stored in time sequence to form inter-frame sampling data. Wherein, if the number of sampling distances is RN, and the number of accumulated chirp signals is CN (if there is no special explanation in the following description, RN and CN are explained here), then when the inter-frame sampling data is stored through the matrix S, the data at each sampling distance obtained after the echo data of a chirp signal is fast Fourier transformed can be stored through a column of the matrix S, so that the matrix S is CN columns and RN rows, and the data obtained after the echo data of multiple chirp signals are fast Fourier transformed from the distance are stored in time sequence in each sampling distance where the sum of the maximum Doppler velocity of each frame is not greater than the threshold to identify a stationary human body.
[0049] Step S130 , based on the inter-frame sampling data, identifying a stationary human body from each sampling distance where the sum of the maximum Doppler velocities of each frame is not greater than a threshold.
[0050] The method for identifying a static human body based on a civil millimeter-wave radar provided by the embodiment of the present disclosure obtains chirp signals of multiple frames of inter-frame sampling data with a larger time span than multiple chirp signals in a single frame or a single subframe, so that the velocity resolution can meet the requirements of identifying a static human body based on movements such as breathing and heartbeat. Moreover, at the sampling distance (exemplarily using an integer multiple of a distance unit as a sampling distance) where the maximum Doppler velocity statistic of each frame (i.e., the sum of the above-mentioned maximum Doppler velocities) is greater than the threshold, there is a target with a relatively large moving speed, so abandoning these sampling distances can reduce false alarms caused by moving targets; in addition, abandoning these sampling distances and directly identifying a static human body from the sampling distance where the sum of the maximum Doppler velocities of each frame is not greater than the threshold can effectively reduce the amount of calculation and simplify the calculation.
[0051] The following is a detailed description of the static human body recognition method based on civilian millimeter-wave radar provided in an embodiment of the present disclosure.
[0052] Step S110, determining the maximum Doppler velocity of the effective target at each sampling distance in the frame through the echo data of multiple chirp signals in the same frame, and calculating the sum of the maximum Doppler velocity of each frame at the same sampling distance.
[0053] In some examples, each frame transmitted by the civil millimeter wave radar includes two subframes, and the first subframe of the two subframes includes multiple chirp signals. The execution of step S110 can be to determine the maximum Doppler velocity of the valid target at each sampling distance in the frame through the echo data of the multiple chirp signals of the first subframe in a frame, where the time interval between two adjacent chirp signals in the multiple chirp signals included in the first subframe is uniform. In this case, the first subframe in each frame is used to determine the maximum Doppler velocity of the valid target at each sampling distance in the frame, and the second subframe can be as follows Figure 2 The example shown includes only one chirp signal to be combined with the first chirp signal in the first subframe to form inter-frame sampling data ( Figure 2 The first subframe in each frame includes W chirp signals).
[0054] Specifically, the first subframe of each frame in this example may be a general subframe, using a multiple input multiple output (MIMO) signal modulation method such as time division multiple access (TDMA) or Doppler division multiple access (DDMA).
[0055] The echo signals of multiple chirp signals in the first subframe of each frame can be subjected to 2D Fast Fourier Transform (2D FFT), incoherent accumulation, constant false alarm rate (CFAR), and direction of arrival (DoA) calculation to obtain the distance, Doppler information, and angle of the normal moving target. Then, the maximum Doppler velocity of the normal moving target at each sampling distance in a frame is recorded.
[0056] Based on multiple chirp signals of the first subframe in each frame, the following steps can be specifically performed to determine the maximum Doppler speed of the effective target at each sampling distance in the frame: first, the echo data of the multiple chirp signals of the first subframe are accumulated by performing a range fast Fourier transform to obtain intra-frame sampling data; then, the intra-frame sampling data is subjected to a Doppler fast Fourier transform to obtain speed information indicating the maximum Doppler speed and then generate a second range Doppler map including the speed information; then, in a constant false alarm rate calculation process, the effective target of the frame is identified from the second range Doppler map and the speed information corresponding to the effective target in the second range Doppler map is determined as the maximum Doppler speed at the sampling distance where the effective target in the frame is located.
[0057] It should be noted that the above distance fast Fourier transform can determine each possible target within the detection range of the millimeter wave radar, the Doppler fast Fourier transform can determine the Doppler velocity of the possible target, and the constant false alarm rate calculation is to screen out the effective target from the possible targets. Furthermore, the angle of arrival can also be calculated to determine the direction of the effective target in the millimeter wave radar. Based on these processes, the point cloud of the millimeter wave radar detecting the moving human body can be finally obtained. Since the above-mentioned distance fast Fourier transform, Doppler fast Fourier transform, constant false alarm rate calculation and arrival angle calculation can be implemented by using the existing relevant technologies based on the detection of indoor moving human bodies by civil millimeter wave radar, they will not be described in detail here.
[0058] The maximum Doppler velocity of each sampling distance in a frame obtained by the above calculation can be recorded by the velocity matrix D. The dimension of the velocity matrix D is determined by the number of sampling distances and the number of frames, that is, the size of one dimension of the velocity matrix D is the number of sampling distances, and the size of the other dimension is the number of frames. In this case, calculating the sum of the maximum Doppler velocities of each frame at the same sampling distance in step S120 includes: calculating the sum of each element corresponding to the same sampling distance in the velocity matrix D.
[0059] For example, the velocity matrix D may be a matrix of RN rows and CN / 2 columns, and the element of the i-th row and j-th column in the velocity matrix D represents the maximum Doppler velocity of the j-th frame at the i-th sampling distance, and calculating the sum of the maximum Doppler velocities of the i-th sampling distance in each frame is to calculate the sum of the elements in the i-th row of the velocity matrix D. In other examples, the velocity matrix D may also be a matrix of CN / 2 rows and RN columns, in which case the element of the i-th row and j-th column in the velocity matrix D represents the maximum Doppler velocity of the i-th frame at the j-th sampling distance, and calculating the sum of the maximum Doppler velocities of the j-th sampling distance in each frame is to calculate the sum of the elements in the j-th column of the velocity matrix D.
[0060] It should be understood that if there is no valid target at a sampling distance in a frame or there is a valid target but the determined maximum Doppler velocity is 0, the maximum Doppler velocity of the sampling distance at the frame is recorded as 0, and the corresponding element in the velocity matrix D is set to 0.
[0061] exist Figure 2 In the example shown, when the maximum Doppler velocity of the effective target at each sampling distance in the frame is determined by the echo data of multiple chirp signals in the first subframe of a frame, since the period of the first subframe is less than half of the frame period T, the above formula (1) is only applicable to the recognition of moving human bodies, and mainly detects people with a moving speed of 0.3m / s-3m / s. People with a speed lower than 0.3m / s will be close to stationary targets and be annihilated.
[0062] It should be noted that the above process of determining the maximum Doppler velocity of a valid target at each sampling distance in a frame through the echo data of multiple chirp signals in the first subframe in a frame is also applicable to determining the maximum Doppler velocity of a valid target at each sampling distance in the frame through the echo data of other multiple chirp signals in the same frame.
[0063] Step S120, performing range fast Fourier transform on echo data of chirp signals of multiple frames and accumulating them to obtain inter-frame sampling data.
[0064] Specifically, the time span of the chirp signals of multiple frames only needs to meet the requirement so that the speed resolution obtained based on the above formula (1) can detect low-speed movements such as human breathing and heartbeat. In practice, one or more chirp signals in one frame can be used. It should be noted that when using multiple chirp signals in one frame, the time intervals between the accumulated multiple chirp signals of multiple frames must be the same.
[0065] In some examples, each frame transmitted by a civil millimeter-wave radar includes two subframes, and the second subframe of the two subframes is transmitted at 1 / 2 of the frame period T. In step S120, echo data of chirp signals of multiple frames are accumulated by performing a range fast Fourier transform to obtain inter-frame sampling data, including: performing a range fast Fourier transform on the echo data of the first chirp signal of each subframe in the multiple frames and accumulating them to obtain inter-frame sampling data, so that the time interval between two adjacent chirp signals in the multiple accumulated chirp signals is constant, which is half of the frame period T.
[0066] Figure 2 An exemplary inter-frame sampling data according to an embodiment of the present disclosure is shown. Figure 2 Civilian millimeter-wave radar transmits multiple frames, each frame has the same period and a time interval, each frame includes two subframes, the lengths of the two subframes are different, among which the period of the first subframe is less than T / 2 and includes multiple chirp signals; the second subframe is transmitted at time T / 2, there is a time interval between it and the first subframe and includes a chirp signal. Figure 2 As shown, the generation of inter-frame sampling data involves 32 frames such as frame U-31 to frame U, and the echo data of 64 chirp signals are subjected to distance fast Fourier transform and accumulated to obtain inter-frame sampling data. In other examples, the number of frames can be increased. For example, the generation of inter-frame sampling data involves 64 frames such as frame U-63 to frame U, and the echo data of 128 chirp signals are subjected to distance fast Fourier transform and accumulated to obtain inter-frame sampling data, so that Doppler velocity with smaller granularity can be detected. For commonly used millimeter waves, the inter-frame sampling data generated by the above-mentioned 64 frames can detect Doppler velocities as low as 0.001m / s, which fully meets the requirements for detecting low-speed movements such as breathing and heartbeats.
[0067] Step S130, based on the inter-frame sampling data, identify the static human body from each sampling distance where the sum of the maximum Doppler velocities of each frame is not greater than the threshold. For the purpose of ease of description, the sampling distance where the sum of the maximum Doppler velocities of each frame is not greater than the threshold is recorded as the sampling distance to be processed.
[0068] Since the time span of the inter-frame sampling data is very large, a target with a high moving speed may cross the distance unit within this time span. Figure 2In the example shown, taking a frame period of 50ms as an example, the time span of accumulating 32 or 64 frames of data is 1.6s / 3.2s. The moving human body will move 1-4m in this time span, while the radar's distance unit is generally 0.1m-0.3m. It can be seen that the moving distance of the moving human body in the time span of the inter-frame sampling data has exceeded the radar's distance unit. For situations where the time span is large and the target may cross the distance unit in the time span, the Doppler fast Fourier transform, constant false alarm rate calculation, and arrival angle calculation directly based on the inter-frame sampling data may result in more noise, which ultimately leads to the identification of incorrect targets and speeds.
[0069] In order to eliminate the noise caused by the large time span of the inter-frame sampling data, the inter-frame sampling data needs to be processed first even when identifying a stationary human body from the sampling distance to be processed based on the inter-frame sampling data. Step S130, identifying a stationary human body from the sampling distance to be processed based on the inter-frame sampling data, may include: searching for a frame whose Doppler velocity at the sampling distance to be processed is greater than zero to obtain a frame to be processed; setting the data corresponding to the sampling distance to be processed and the frame to be processed in the inter-frame sampling data to zero to obtain the processed inter-frame sampling data; and identifying a stationary human body from the sampling distance to be processed based on the processed inter-frame sampling data.
[0070] Specifically, as mentioned above, the inter-frame sampling data includes a total of RN×CN data, which can be recorded by a matrix S, where one dimension of the matrix S is RN and the other dimension is CN. Take the matrix S as an example with RN rows and CN columns, where the data in the Ri-th row and the n-th column corresponds to the Ri-th sampling distance and the n-th chirp signal. If the sum of the maximum Doppler velocities of each frame at the Ri-th sampling distance is not greater than the threshold, that is, the Ri-th sampling distance is the sampling distance to be processed described above. Next, if it is determined based on the calculation result of step S110 that the Doppler velocity of the frame where the n-th chirp signal is located at the Ri-th sampling distance is greater than zero, then the data in the Ri-th row and the n-th column in the matrix S is set to zero.
[0071] In the disclosed embodiment, the sampling distance to be processed is a sampling distance at which the sum of the maximum Doppler velocities of each frame is not greater than a threshold value. There is no target with a large Doppler velocity at this sampling distance, or there is occasionally a target with a large Doppler velocity. The above-mentioned zeroing step is to eliminate the interference of such a short-lived target with a large Doppler velocity, so that there is no noise in the result of resting human body recognition based on the processed inter-frame sampling data, thereby improving the recognition accuracy.
[0072] Furthermore, the above-mentioned identification of a stationary human body from a sampling distance to be processed based on the processed inter-frame sampling data may include: performing a Doppler fast Fourier transform on the processed inter-frame sampling data to obtain velocity information and then generate a first distance Doppler map including the velocity information; calculating the average value of each data amplitude at the same sampling distance in the inter-frame sampling data to obtain the average energy at each sampling distance; and in the constant false alarm rate calculation process, identifying a stationary human body from the first distance Doppler map based on the average energy.
[0073] Taking the matrix S as an example with RN rows and CN columns, the average energy M(Ri) at the Ri-th sampling distance is shown in formula (2), where abs is the modulo operator, and abs(S(Ri,i)) represents the amplitude of the data corresponding to the Ri-th sampling distance and the i-th frame in the inter-frame sampling data.
[0074]
[0075] Specifically, the intensity value of the peak in the range Doppler image represents the intensity of the echo data of the target. The above-mentioned identification of a stationary human body from the first range Doppler image based on the average energy is that a peak in the first range Doppler image whose intensity value is greater than the CFAR threshold and higher than the average energy of the sampling distance is determined as the existence of a stationary human body, and the position of the peak represents the distance information and speed information of the stationary human body.
[0076] Furthermore, the above-mentioned static human body recognition method also includes: calculating the arrival angle of the recognized static human body, so as to finally obtain the point cloud of the static person detected by the millimeter wave radar.
[0077] In some embodiments, based on the inter-frame sampling data, for the sampling distance where the sum of the maximum Doppler velocities is not greater than a threshold, interference from normal moving targets is eliminated, and then Doppler FFT, incoherent accumulation, CFAR and energy comparison, and DoA are performed to obtain stationary human target information.
[0078] Figure 3 FIG. 1 is a flow chart of generating a detection point cloud of a millimeter wave radar based on a static human body recognition method according to an embodiment of the present disclosure, wherein the generation process of the inter-frame sampling data is as follows: Figure 2 See Figure 3 , the process includes:
[0079] Steps S311 to S315, i.e., based on the first subframe, perform range fast Fourier transform, Doppler fast Fourier transform, non-coherent accumulation, constant virtual warning calculation and arrival angle calculation in sequence, and finally obtain a motion point cloud (i.e., a point cloud obtained by detecting a moving human body). For details, reference may be made to the detailed description of step 110 above, where the non-coherent accumulation added between the Doppler fast Fourier transform and the constant virtual warning calculation refers to the signal accumulation method of different transmitting antennas or receiving antennas using the value of amplitude superposition to obtain a second range Doppler map.
[0080] Steps S321 to S323, that is, obtaining the velocity matrix D and determining whether the sum of the maximum Doppler velocities of each frame at a sampling distance is greater than a threshold based on the velocity matrix D, and the sampling distance that is not greater than the threshold is discarded. Here, the sampling distance can be set as the distance unit of the radar.
[0081] Step 331 to step S337, wherein step 331 to step S332 are to obtain inter-frame sampling data based on the first chirp signal of each subframe, and specific reference may be made to the detailed description of step 120 above; step S333 is to calculate the average value of each data amplitude at the same sampling distance in the inter-frame sampling data as described above to obtain the average energy at each sampling distance, and the subsequent steps are for the sampling distances that have not been discarded. First, in step S334, the inter-frame sampling data is processed, i.e., the data corresponding to the sampling distance to be processed and the frame to be processed in the inter-frame sampling data is set to zero as described above, so as to obtain the processed inter-frame sampling data; then, steps S335 to S337, i.e., based on the processed inter-frame sampling data, perform Doppler fast Fourier transform, incoherent accumulation, constant virtual warning calculation and arrival angle calculation in sequence, so as to finally obtain a static point cloud (i.e., a point cloud obtained by detecting a stationary human body), for details, please refer to the detailed description of step 130 above, wherein the incoherent accumulation added between the Doppler fast Fourier transform and the constant virtual warning calculation refers to the signal accumulation method of different transmitting antennas or receiving antennas adopting the value of amplitude superposition to obtain the first distance Doppler diagram, and the constant virtual warning calculation process uses the average energy calculated in step S333.
[0082] Step S340: merge the moving point cloud and the static point cloud and output them.
[0083] In the whole process of generating the detection point cloud of the millimeter wave radar, step S332 is to obtain the inter-frame sampling data based on the first chirp signal of each subframe in multiple frames, so the time span is large, and the speed resolution can meet the needs of identifying the static human body based on the movement of breathing and heartbeat. In other words, improving the speed resolution through inter-frame sampling data is equivalent to amplifying the tiny Doppler changes such as breathing and heartbeat during the detection process, thereby realizing the acquisition of static targets, wherein the time for accumulating the inter-frame sampling data is generally in the second level, and the environmental change has little effect on the system for realizing the acquisition of static targets. In addition, the implementation of step S322 ensures that the detection of static human bodies will be performed only at the sampling distance where the maximum Doppler speed of each frame is not greater than the threshold, thereby effectively reducing the amount of calculation. The execution of step S344 is to eliminate the interference of the target with a large Doppler speed that appears briefly, so that the result of the static human body recognition in each sampling distance where the sum of the maximum Doppler speed of each frame is not greater than the threshold will not have noise, thereby improving the accuracy of recognition. Figure 3 The static point cloud shown in includes the stationary human body, and the moving human body signal is filtered.
[0084] Corresponding to the resting human body recognition method provided in the above embodiment, another embodiment of the present disclosure further provides an electronic device, Figure 4 The figure shows the structure of the electronic device. Figure 4 , the electronic device 800 includes a processor 810, at least one storage unit 820, and a bus 840 connecting different system components (including the processor 810 and the storage unit 820). The storage unit 820 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 8201 and / or a cache storage unit 8202, and may further include a read-only storage unit (ROM) 8203. The storage unit 820 may also include a dynamic random access memory / tightly coupled memory 8204 storing a set of program modules 8205, and the program modules 8205 include but are not limited to: one or more application programs and program data, which enable the electronic device 800 to implement system functions. It should be noted that when the partial program included in the program module 8205 is executed by the processor 810, each process of each embodiment of the above-mentioned static human body recognition method can be implemented and the same technical effect can be achieved.
[0085] Bus 840 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.
[0086] The electronic device 800 may also communicate with one or more terminal devices that enable a user to interact with the electronic device 800, and / or any transmission device that enables the electronic device 800 to communicate with one or more other computing devices. Such communication may be performed through an input / output (I / O) interface 850, which is connected to the bus 840. In addition, the electronic device 800 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 860 connected to the bus 840, and the communication is performed based on a communication protocol.
[0087] It can be understood by a person of ordinary skill in the art that all or part of the steps in the various methods of the above-mentioned embodiments can be completed by instructions, or by controlling related hardware through instructions, and the instructions can be stored in a computer-readable storage medium and loaded and executed by a processor. To this end, the embodiments of the present disclosure also provide a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by the processor, each process of each embodiment of the above-mentioned method for identifying a resting human body can be implemented. Among them, computer-readable storage media, such as U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., can store program codes.
[0088] Since the program stored in the readable storage medium can execute the steps in any of the resting human body recognition methods provided in the embodiments of the present disclosure, the beneficial effects that can be achieved by any of the resting human body recognition methods provided in the embodiments of the present disclosure can be achieved. For details, please refer to the previous embodiments, which will not be repeated here. The specific implementation of each of the above operations can be referred to the previous embodiments, which will not be repeated here.
[0089] It should be noted that when describing each embodiment in this specification, the focus is on the differences from other embodiments, and the same or similar parts between the embodiments can be understood by reference to each other. For each device embodiment, since it is basically similar to the method embodiment, the relevant parts can refer to the description of the method embodiment. Since each device embodiment has the beneficial effects that can be achieved by the above-mentioned method embodiment, please refer to the previous embodiment for details, which will not be repeated here.
[0090] In addition, it should be noted that in the apparatus and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. Moreover, the steps of performing the above-mentioned series of processing can be naturally performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order, and some steps can be performed in parallel or independently of each other. For those of ordinary skill in the art, it is understood that all or any steps or components of the method and apparatus of the present invention can be implemented in any computing device (including processors, storage media, etc.) or a network of computing devices in hardware, firmware, software or a combination thereof, which can be achieved by those of ordinary skill in the art using their basic programming skills after reading the description of the present invention.
[0091] Finally, it should be noted that the terms "first", "second" and "third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, the above-described embodiments are only specific implementation methods of the present disclosure, which are used to illustrate the technical solutions of the present disclosure, rather than to limit them. The protection scope of the present disclosure is not limited thereto. Although the present disclosure is described in detail with reference to the aforementioned embodiments, a person of ordinary skill in the art should understand that any person of ordinary skill in the art can still modify the technical solutions recorded in the aforementioned embodiments within the technical scope disclosed in the present disclosure, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present disclosure, and should all be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure shall be subject to the protection scope of the claims.
Claims
1. A static human body recognition method based on a civilian millimeter wave radar, wherein the civilian millimeter wave radar transmits multiple frames and each frame includes multiple chirp signals. include: Determine the maximum Doppler velocity of the effective target at each sampling distance in the frame through the echo data of multiple chirp signals in the same frame, and calculate the sum of the maximum Doppler velocity of each frame at the same sampling distance; Performing range fast Fourier transform on echo data of chirp signals of multiple frames and accumulating them to obtain inter-frame sampling data; Based on the inter-frame sampling data, a stationary human body is identified from each sampling distance where the sum of the maximum Doppler velocities of each frame is not greater than a threshold.
2. The method for identifying a resting human body according to claim 1, in, The sampling distance at which the sum of the maximum Doppler velocities of each frame is not greater than a threshold is a sampling distance to be processed, and based on the inter-frame sampling data, a stationary human body is identified from the sampling distance to be processed, including: Find the frame whose Doppler velocity is greater than zero at the sampling distance to be processed, and obtain the frame to be processed; Resetting the data corresponding to the sampling distance to be processed and the frame to be processed in the inter-frame sampling data to zero to obtain processed inter-frame sampling data; Based on the processed inter-frame sampling data, a stationary human body is identified from the sampling distance to be processed.
3. The method for identifying a resting human body according to claim 2, in, Based on the processed inter-frame sampling data, identifying a stationary human body from the to-be-processed sampling distance includes: Performing Doppler fast Fourier transform on the processed inter-frame sampling data to obtain velocity information and then generate a first range Doppler map including the velocity information; Calculate the average value of each data amplitude at the same sampling distance in the inter-frame sampling data to obtain the average energy at each sampling distance; In the constant false alarm rate calculation process, a peak in the first range Doppler map whose intensity value is greater than a threshold of the constant false alarm rate and higher than the average energy at the sampling distance is determined as the presence of a resting human body, wherein the position of the peak represents the distance information and speed information of the resting human body.
4. The method for identifying a human body at rest according to claim 3, further comprising: include: Calculate the arrival angle of the identified static human body.
5. The method for identifying a resting human body according to claim 1, in, Determining the maximum Doppler velocity of a valid target at each sampling distance in the frame through echo data of multiple chirp signals in the same frame includes: Performing range fast Fourier transform on echo data of multiple chirp signals in the same frame and accumulating them to obtain intra-frame sampling data; Performing Doppler fast Fourier transform on the intra-frame sampling data to obtain speed information indicating the maximum Doppler speed and then generating a second range Doppler map including the speed information; In the constant false alarm rate calculation process, a valid target of the frame is identified from the second range Doppler map and the speed information corresponding to the valid target in the second range Doppler map is determined as the maximum Doppler speed at the sampling distance of the valid target in the frame.
6. The method for identifying a resting human body according to claim 5, in, Determining the maximum Doppler velocity of a valid target at each sampling distance in the frame through echo data of multiple chirp signals in the same frame also includes: recording the maximum Doppler velocity through a velocity matrix, wherein the dimension of the velocity matrix is determined by the number of sampling distances and the number of frames; The calculating the sum of the maximum Doppler velocities of each frame at the same sampling distance includes: calculating the sum of each element corresponding to the same sampling distance in the velocity matrix.
7. The method for identifying a resting human body according to any one of claims 1 to 6, in, Each frame transmitted by the civil millimeter-wave radar includes two subframes, the first subframe of the two subframes includes multiple chirp signals, the period of the first subframe is less than 1 / 2 of the frame period, and the second subframe of the two subframes is transmitted at 1 / 2 of the frame period, and the second subframe transmitted by each antenna includes only one chirp signal.
8. The method for identifying a resting human body according to claim 7, in, Determining the maximum Doppler velocity of the effective target at each sampling distance in the frame through the echo data of multiple chirp signals in the same frame includes: determining the maximum Doppler velocity of the effective target at each sampling distance in the frame through the echo data of multiple chirp signals in the first subframe in a frame; The step of performing a range fast Fourier transform on the echo data of the chirp signals of multiple frames and accumulating them to obtain the inter-frame sampling data includes: performing a range fast Fourier transform on the echo data of the first chirp signal of the first subframe in the multiple frames and accumulating them to obtain the inter-frame sampling data.
9. An electronic device, include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein when the program is executed by the processor, the method for identifying a resting human body as described in any one of claims 1 to 8 is implemented.
10. A computer-readable storage medium having a computer program or instruction stored thereon, wherein the computer program or instruction, when executed by a processor, implements the method for identifying a resting human body according to any one of claims 1 to 8.