CN116087943A - Indoor falling detection method and system based on millimeter wave radar - Google Patents

Abstract

The disclosure provides an indoor falling detection method and system based on millimeter wave radar, which relate to the technical field of millimeter wave radar signals, and comprise the steps of transmitting continuous frequency modulation millimeter wave radar signals and receiving, and performing FFT operation on the obtained millimeter wave radar signals; performing self-adaptive constant false alarm rate detection on the millimeter wave radar signals after FFT operation, and selecting effective target motion information; filtering the effective target motion information, detecting whether a person exists in the room, if so, acquiring the motion information of the human target by using a deep learning LSTM network, and judging whether a falling event occurs according to the motion information of the human target. By means of the combined calculation of the obtained information of the human body movement and judging whether the human body falls down, the problem that the false alarm rate is high due to single information calculation is avoided.

Description

Indoor fall detection method and system based on millimeter wave radar

technical field

The present disclosure relates to the technical field of millimeter-wave radar signals, and in particular to an indoor fall detection method and system based on millimeter-wave radar.

Background technique

The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art.

Population aging has become a common phenomenon in many parts of the world. In particular, the current situation of population aging in my country is becoming more and more severe, and presents the following characteristics: a large elderly population, a rapid aging rate, and a sharp increase in the old-age dependency ratio. At that time, it is imperative to strengthen the technological and intelligent upgrading of elderly products.

According to statistics, falls are the number one cause of death in Shanghai and even death for the elderly over 65 years old. Even for elderly people who are usually in good physical condition, 17.7% of them fall and cause serious injuries. The current fall detection systems are mainly divided into wearable and non-wearable. However, wearable devices need to be charged frequently, and it is impossible to ensure that the elderly wear the devices all the time. For non-wearable detection systems, cameras, thermal infrared or millimeter wave radars are mainly used. However, the use of cameras or infrared is easily affected by optical fibers and obstructions, which cannot guarantee round-the-clock detection, and easily exposes user privacy. But for now, the use of millimeter-wave radar for indoor fall detection still has the problems of low accuracy and high false alarm rate.

Contents of the invention

In order to solve the above problems, the present disclosure proposes an indoor fall detection method and system based on millimeter wave radar, provides a non-contact indoor fall detection method that can accurately identify the fall state of indoor personnel, and uses a deep learning network to train a classifier, Establish a human fall judgment mechanism, which is suitable for indoor environments and ensures the accuracy and reliability of detection.

According to some embodiments, the present disclosure adopts the following technical solutions:

An indoor fall detection method based on millimeter wave radar, including:

Transmit and receive continuous frequency modulation millimeter-wave radar signals, and perform FFT operations on the acquired millimeter-wave radar signals;

Carry out adaptive constant false alarm rate detection on the millimeter-wave radar signal after FFT operation, and select effective target motion information;

Filter the effective target motion information to detect whether there are people in the room. If there is someone in the room, use the deep learning LSTM network to obtain the motion information of the human target, and judge whether there is a fall event based on the motion information of the human target.

According to some embodiments, the present disclosure adopts the following technical solutions:

Indoor fall detection system based on millimeter wave radar, including:

The signal transmission and recovery module is used to transmit and receive continuous frequency modulation millimeter-wave radar signals; perform FFT calculation on the acquired millimeter-wave radar signals;

The signal calculation and processing module is used to perform FFT calculation on the acquired millimeter-wave radar signal; perform adaptive constant false alarm rate detection on the millimeter-wave radar signal after the FFT calculation, and select effective target motion information;

The model judgment and detection module is used to filter the effective target motion information to detect whether there are people in the room. If it detects the presence of people in the room, use the deep learning LSTM network to obtain the motion information of the human target, and judge according to the motion information of the human target Whether there is a fall incident.

According to some embodiments, the present disclosure adopts the following technical solutions:

A non-transitory computer-readable storage medium, the non-transitory computer-readable storage medium is used to store computer instructions, and when the computer instructions are executed by a processor, the described indoor fall detection method based on millimeter-wave radar is realized .

According to some embodiments, the present disclosure adopts the following technical solutions:

An electronic device, comprising: a processor, a memory, and a computer program; wherein, the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processor executes the computer program stored in the memory, To make the electronic equipment execute and realize the indoor fall detection method based on the millimeter wave radar.

Compared with the prior art, the beneficial effects of the present disclosure are:

1. The detection method of the present disclosure adopts a non-contact method, which has no impact on the daily life of the human body, does not need to collect videos, can protect the privacy and security of users, and can accurately and quickly identify the fall state of indoor personnel. It is easy to use and safe. With the advantages of comfortable use, it is suitable for indoor environments such as homes, nursing homes, and wards to ensure the health of the elderly in their daily lives.

The components are simple and easy to install. It can quickly identify the status of widows and elderly people living alone and generate early warning information in time. It has the advantages of ease of use, safety, and comfort. It is suitable for elderly care models such as families, institutions, and communities, and improves the safety of the elderly. sex and health.

2. The working method of the detection system of the present disclosure adopts the sound sensor for indoor human body target recognition, which reduces false alarms in fall detection, avoids alarms due to animal movement during the detection process, and ensures detection accuracy and reliability.

3. This disclosure uses a deep learning network to train a classifier, uses the established data set to train the classifier, improves the performance and accuracy of the classifier, and ensures the accuracy of human motion information detected.

4. This disclosure uses the established human body fall determination mechanism to determine whether a fall has occurred through joint calculation of the obtained human body movement information (speed, acceleration, distance, angle, etc.), avoiding the high rate of false alarms caused by single information calculation problems, improve the accuracy of fall detection and reduce the false alarm rate.

Description of drawings

The accompanying drawings constituting a part of the present disclosure are used to provide a further understanding of the present disclosure, and the exemplary embodiments and descriptions of the present disclosure are used to explain the present disclosure, and do not constitute improper limitations to the present disclosure.

Fig. 1 is the working flowchart of the embodiment of the present disclosure;

FIG. 2 is a flowchart of a fall detection mechanism in an embodiment of the present disclosure;

FIG. 3 is a VI-CFAR structural diagram of an embodiment of the present disclosure;

FIG. 4 is a structural diagram of an LSTM network according to an embodiment of the present disclosure.

Detailed ways:

The present disclosure will be further described below in conjunction with the accompanying drawings and embodiments.

It should be noted that the following detailed description is exemplary and intended to provide further explanation of the present disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.

It should be noted that the terminology used herein is only for describing specific embodiments, and is not intended to limit the exemplary embodiments according to the present disclosure. As used herein, unless the context clearly dictates otherwise, the singular is intended to include the plural, and it should also be understood that when the terms "comprising" and/or "comprising" are used in this specification, they mean There are features, steps, operations, means, components and/or combinations thereof.

Example 1

An embodiment of the present disclosure provides an indoor fall detection method based on millimeter-wave radar, as shown in FIG. 1 , including:

Step 1: Start the detection equipment, transmit and receive continuous frequency-modulated millimeter-wave radar signals, pass the millimeter-wave signals to the processor, and perform FFT operations on the millimeter-wave radar signals in three dimensions: speed, distance, and angle.

The specific steps of performing the FFT operation on the acquired millimeter-wave radar signal are:

Step 1.1: Perform FFT operation on each chirp dimension of the echo to obtain a distance-pulse diagram;

Step 1.2: Perform FFT operation on the result of the range FFT in the chirp dimension to obtain the range-Doppler map, and extract its peak value to obtain the beat frequency (f _I ) and Doppler frequency (f _D ) of the indoor target;

Step 1.3: Calculate the echo delay through the beat frequency, so as to calculate the distance of the indoor target; the beat frequency calculation formula is:

Among them, f _B is the beat frequency between the transmitted wave and the echo, f _D is the Doppler frequency, f _I is the frequency of the intermediate frequency signal, K represents the slope of the Chirp, R is the target distance, and c is the electromagnetic wave velocity.

Step 1.4: Calculate the speed of the moving target indoors through the Doppler frequency, and the Doppler frequency calculation formula is:

Among them, f _D is the Doppler frequency, v is the target velocity, and λ is the electromagnetic wave wavelength.

Step 1.5: Calculate the angle of the target using the phase difference of the echoes arriving at different RX antennas. Perform FFT operation on the FFT results of the distance and speed of multiple RX dimensions in the RX dimension to obtain the range-Doppler-azimuth map; the phase difference calculation formula between the two RXs is:

where ω is the phase difference between the two receiving antennas and θ is the angle of the target relative to the radar.

Step 2: the step of performing adaptive constant false alarm rate detection on the millimeter-wave radar signal after the FFT operation, and selecting effective target motion information is, as shown in Figure 3,

Step 2.1: Pass the signal into the square law detector and process it;

Step 2.2: Pass the signal output from the square-law detector to the VI-CFAR detector for processing, and obtain reference units A and B;

并求出参考单元平均值的平方

并求出A和B的VI和sum为单元数值的和sum；其中VI的计算公式为：Step 2.3: Compute the average of reference cells A and B

and find the square of the mean value of the reference cell

And find the VI and sum of A and B as the sum of the unit values; the calculation formula of VI is:

Among them, n is the number of reference units.

Step 2.4: Judgment of clutter background and CFAR strategy selection;

Step 2.5: Input the unit to be detected to the comparator for detection, and select effective target motion information.

Further, step 3: filter the effective target motion information, that is, use a Singerαβγ Kalman filter to filter the signal after VI-CFAR.

The steps for filtering are as follows:

Step 3.1: Define the indoor target as a random acceleration model, where the formula of the model is:

代表二阶目标随机加速度参数；w(n)是具有零均值、单位方差的高斯随机变量；σ_m是运动标准偏差；ρ_m为激动相关系数。in,

Represents the second-order target random acceleration parameter; w(n) is a Gaussian random variable with zero mean and unit variance; σ _m is the standard deviation of motion; ρ _m is the excitation correlation coefficient.

Among them, w(n) is a Gaussian random variable with zero mean and unit variance; σ _m is the motorized standard deviation; the formula of the activation correlation coefficient ρ _m is:

Among them, τ _m is the relative time of the target acceleration due to the target motion, and T is the target motion time.

Step 3.2: Define the autocorrelation function of the model, and its function formula is:

Among them, σ _a is the standard deviation.

Step 3.3: Define the transition matrix of the filter as:

in,

Step 3.4: The signal is then filtered using the model and matrix functions.

Step 4: Detect whether there is a person in the room. The method for detecting whether there is a person in the room is to use a sound sensor to judge whether there is a person in the room. When a voice conversation or other human behavior sounds are detected in the room, it is determined that the indoor target is a human body, otherwise Judging that no one is present in the room.

Step 5: Use a deep learning network to train a classifier, and use the classifier to obtain motion information of an effective human target in the room; the motion information of the human target includes speed, acceleration, distance, and angle.

As shown in Figure 4, the specific steps are:

Step 5.1: Set the structure of LSTM, including forget gate, input gate, cell state and output gate. Among them, the calculation formula of the forget gate is:

f _t ＝σ(W _f ·[h _t-1 ,x _t ]+b _f ) (9)

Among them, W _f is the weight, h _t-1 is the previous output, x _t is the current input, σ is the forget gate, and b _f is a set parameter.

The input gate calculation formula is:

i _t =σ(W _i ·[h _t-1 ,x _t ]+b _i ) (10)

是一个新的候选值向量，tanh为网络中的tanh层。W _i, like W _f, is also a weight, and b _i, like b _f, is a set parameter.

is a new candidate value vector, and tanh is the tanh layer in the network.

The cell state calculation formula is:

Among them, C _t is the new cell state, and C _t-1 is the previous cell state.

The output gate calculation formula is:

o _t =σ(W _o [h _t-1 ,x _t ]+b _o )(13)

h _t ＝o _t *tanh(C _t )(14)

Among them, W _o is the weight, b _o is a parameter. h _t is the new output.

Step 5.2: Input the data set for calculating the target motion information into the LSTM network training classifier;

Step 5.3: Input the distance, speed, and angle information obtained after the processed signal into the LSTM network as input data to obtain the motion information of the target;

Step 6: Use the human body motion information obtained by the classifier to determine whether there is a fall event;

As shown in Figure 2, the specific steps are:

Step 6.1: use the obtained human body velocity to obtain the acceleration;

Step 6.2: Set the acceleration threshold. If the acceleration of the human body is lower than the threshold, it is judged that there is no fall event, and if the acceleration exceeds the threshold, the next step is judged;

Step 6.3: If the acceleration of the human body exceeds the set acceleration threshold, continue to detect whether the human body has a rapid deceleration behavior (the speed of the human body will decrease sharply when it hits the ground). If there is no rapid deceleration behavior, it is determined that there is no fall event. If there is sudden deceleration in the human body, the detection will continue;

Step 6.4: Continue to judge whether the time from acceleration to deceleration of the human body is less than 3 seconds. If the time of the whole event is not less than 3 seconds, it is determined that there is no fall event, and if it is less than 3 seconds, continue to detect;

Step 6.5: Calculate the change in the height point of the center of gravity of the human body. If the change in the height of the human body from acceleration to deceleration is greater than 50cm, it is determined that there is a fall event.

Example 2

An embodiment of the present disclosure provides an indoor fall detection system based on millimeter wave radar, including:

The signal transmission and recovery module is used to transmit and receive continuous frequency modulation millimeter-wave radar signals; perform FFT calculation on the acquired millimeter-wave radar signals;

The signal calculation and processing module is used to perform FFT calculation on the acquired millimeter-wave radar signal; perform adaptive constant false alarm rate detection on the millimeter-wave radar signal after the FFT calculation, and select effective target motion information;

The model judgment and detection module is used to filter the effective target motion information to detect whether there are people in the room. If it detects the presence of people in the room, use the deep learning LSTM network to obtain the motion information of the human target, and judge according to the motion information of the human target Whether there is a fall incident.

The signal emission and recovery module includes detection equipment for transmitting and recovering continuous frequency-modulated millimeter-wave signals;

The signal calculation and processing module includes a processor, which is used to perform FFT calculation on the acquired millimeter wave radar signal; perform adaptive constant false alarm rate detection on the millimeter wave radar signal after the FFT calculation, and select effective target motion information;

The signal calculation and processing module also includes a sound sensor, which is used to judge whether there is a person in the room, and to obtain sound signals.

It also includes a square-law detector and a SingerαβγKalman filter, which are used to perform adaptive constant false alarm rate detection on the signal after FFT operation, select effective target motion information, and filter the signal.

Example 3

An embodiment of the present disclosure provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium is used to store computer instructions, and when the computer instructions are executed by a processor , realizing the steps of the indoor fall detection method based on the millimeter wave radar.

Example 4

An embodiment of the present disclosure provides an electronic device, including: a processor, a memory, and a computer program; wherein, the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device is running, the processing The computer executes the computer program stored in the memory, so that the electronic device executes the steps of implementing the method for detecting indoor falls based on millimeter wave radar.

The present disclosure is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present disclosure. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and a combination of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing equipment produce a An apparatus for realizing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby The instructions provide steps for implementing the functions specified in the flow chart or blocks of the flowchart and/or the block or blocks of the block diagrams.

Although the specific implementation of the present disclosure has been described above in conjunction with the accompanying drawings, it does not limit the protection scope of the present disclosure. Those skilled in the art should understand that on the basis of the technical solutions of the present disclosure, those skilled in the art do not need to pay creative work Various modifications or variations that can be made are still within the protection scope of the present disclosure.

Claims (10)

1. The indoor falling detection method based on the millimeter wave radar is characterized by comprising the following steps of:

transmitting and receiving continuous frequency modulation millimeter wave radar signals, and performing FFT operation on the obtained millimeter wave radar signals;

performing self-adaptive constant false alarm rate detection on the millimeter wave radar signals after FFT operation, and selecting effective target motion information;

filtering the effective target motion information, detecting whether a person exists in the room, if so, acquiring the motion information of the human target by using a deep learning LSTM network, and judging whether a falling event occurs according to the motion information of the human target.

2. An indoor fall detection method based on millimeter wave radar as defined in claim 1, wherein the process of performing FFT operation on the acquired millimeter wave radar signal comprises: performing FFT operation on millimeter wave radar signals in three dimensions of speed, distance and angle, and performing FFT operation on each chirp dimension of echo to obtain a distance-pulse diagram; and performing FFT operation on the distance FFT result in the chirp dimension to obtain a distance-Doppler graph, and extracting the peak value to obtain the beat frequency and Doppler frequency of the indoor target.

3. The indoor fall detection method based on millimeter wave radar according to claim 1, wherein the step of performing adaptive constant false alarm rate detection on the millimeter wave radar signal after FFT operation and selecting effective target motion information comprises the steps of:

s1: the signal is transmitted into a square law detector and processed;

s2: transmitting the signal output by the square law detector into a VI-CFAR detector for processing, and obtaining reference units A and B;

s3: calculating the reference units A and B, judging clutter background and selecting CFAR strategy;

s4: and inputting the unit to be detected into a comparator for detection, and selecting effective target motion information.

4. The indoor fall detection method based on millimeter wave radar as claimed in claim 1, wherein the filtering processing of the effective target motion information is performed by: and filtering the signals subjected to VI-CFAR by using a Singer alpha beta gamma Kalman filter.

5. The indoor fall detection method based on millimeter wave radar as set forth in claim 4, wherein the step of performing the filtering process is:

defining an indoor target as a random acceleration model and an autocorrelation function of the model, then defining a transfer matrix of a filter, and filtering signals by using the model and the matrix function.

6. The indoor fall detection method based on millimeter wave radar according to claim 1, wherein the method for detecting whether a person exists in the room is to use a sound sensor to determine whether a person exists in the room, and determine that the indoor target is a human body after detecting that a voice conversation or other human body behavioral sounds exist in the room, and otherwise determine that no person exists in the room.

7. An indoor fall detection method based on millimeter wave radar as claimed in claim 1, wherein the movement information of the human target includes speed, acceleration, distance and angle.

8. Indoor fall detecting system based on millimeter wave radar, its characterized in that includes:

the signal transmitting and recovering module is used for transmitting and receiving continuous frequency modulation millimeter wave radar signals; performing FFT operation on the obtained millimeter wave radar signals;

the signal calculation processing module is used for carrying out FFT operation on the obtained millimeter wave radar signals; performing self-adaptive constant false alarm rate detection on the millimeter wave radar signals after FFT operation, and selecting effective target motion information;

the model judging and detecting module is used for carrying out filtering processing on the effective target motion information, detecting whether a person exists in a room, if so, acquiring the motion information of the human target by using the deep learning LSTM network, and judging whether a falling event occurs according to the motion information of the human target.

9. A non-transitory computer-readable storage medium storing computer instructions that, when executed by a processor, implement the millimeter wave radar-based indoor fall detection method of any of claims 1-7.

10. An electronic device, comprising: a processor, a memory, and a computer program; wherein the processor is connected to the memory, and the computer program is stored in the memory, which processor executes the computer program stored in the memory when the electronic device is running, to cause the electronic device to perform the method of performing millimeter wave radar based indoor fall detection as claimed in any one of claims 1 to 7.

Images