A parkinson gait assessment method and system based on millimeter wave radar signals

Abstract

The application discloses a Parkinson gait evaluation method and system based on a millimeter wave radar signal. The method comprises the following steps: acquiring a millimeter wave radio frequency signal in an environment, performing data processing on the signal to obtain a distance spectrum of position change of a Parkinson patient's body and a Doppler spectrum reflecting the degree of change of the Parkinson patient's body; extracting a peak value at each moment from the distance spectrum through a filtering method, calculating a cross-correlation coefficient of the Doppler spectrum as an index of gait pattern matching, and dividing the Doppler spectrum according to each step; calculating gait features according to the filtered distance spectrum peak value and the divided Doppler spectrum; inputting the normalized gait features into a trained machine learning model for feature evaluation to obtain a UPDRS-III score of the patient. The application realizes non-invasive measurement and evaluation of Parkinson gait features in a low-cost manner, is convenient to operate, and is accurate in evaluation effect, thereby providing quantitative data and diagnosis reference for doctors.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of gait assessment, in particular to a Parkinson gait assessment method and system based on millimeter wave radar signals. BACKGROUND

[0002] In traditional Parkinson gait assessment, doctors observe and assess the patient's behavioral characteristics, motor ability, posture and other related symptoms through phenotypic testing, so as to assess the Parkinson gait. However, phenotypic testing relies on the subjective and qualitative assessment of Parkinson symptoms by professional doctors, which has the shortcomings of low accuracy, large error and easy to be affected by the doctor's psychological factors. Therefore, in recent years, the field has been seeking to introduce precise quantitative data collection and analysis mechanisms to assist doctors in gait assessment, and the number of application systems based on various sensors has rapidly grown, showing a booming trend. For example, the technology based on a three-dimensional motion capture system can track the position information of each joint of the Parkinson patient's body, and through the change of the position of the human body joint, the whole walking process can be quantified to realize the extraction of gait features. The technology based on an inertial sensor can track the speed, acceleration, angular velocity information of specific parts of the Parkinson patient's body such as feet, shanks, wrists and waists during walking, and through calculation, the step length, step speed and other gait features can be obtained. The technology based on a pressure sensor measures the change of plantar pressure of the patient during walking by installing a special pressure sensing mat or requiring the patient to wear a pressure sensing insole, and indirectly obtains gait features by analyzing the change of pressure.

[0003] Although the existing application systems can obtain relatively accurate gait features, these systems all have the problem of inconvenience of use: the use of a three-dimensional motion capture system requires the patient to wear a special marker tight-fitting clothes, the use of an inertial sensor requires the patient to wear related equipment on the joints of the body, and the use of a pressure sensing mat or a pressure sensing insole requires the patient to change into special shoes. Parkinson patients are usually old people, and wearing additional equipment is a burden and will affect their normal walking, and even some patients may develop fear. On the other hand, because of the high price and complex installation and use process, these devices are only suitable for one-time gait assessment in a hospital dedicated examination room, and cannot help doctors to conduct long-term assessment of the patient's gait.

[0004] Millimeter wave radar is a radar technology that uses millimeter wave frequency band radio frequency signals for detection and perception, with characteristics of high precision, high resolution and less environmental interference. In recent years, with the development of mobile computing and millimeter wave technology, millimeter wave radar has been widely used in commercial fields. However, the existing gait evaluation technology based on millimeter wave radar signals still stays in measuring the gait characteristics of ordinary users. Not only has it not considered the problem that the gait characteristics of Parkinson's patients are significantly different from normal people due to symptoms such as muscle stiffness, movement retardation, and unstable posture, but it has also not made further gait evaluation analysis based on the obtained gait characteristics. SUMMARY

[0005] In view of the weaknesses of the existing millimeter wave radar gait analysis, the present application proposes a Parkinson's gait evaluation method and system based on millimeter wave radar signals, which solves the problem of how to obtain accurate abnormal gait characteristics of Parkinson's patients through millimeter wave radar transmitting and receiving arrays, and evaluates the UPDRS-III gait score of Parkinson's patients using machine learning methods.

[0006] In order to achieve the above application purposes, the technical solutions of the present application are as follows:

[0007] In the first aspect, a Parkinson's gait evaluation method based on millimeter wave radar signals includes the following steps:

[0008] Obtain millimeter wave radio frequency signals in the environment, and process the signals to obtain a range spectrum reflecting the change in the position of the Parkinson's patient's body and a Doppler spectrum reflecting the change in the Parkinson's patient's body degree;

[0009] Extract the peak value of the range spectrum at each time through a filtering method, calculate the cross-correlation coefficient of the Doppler spectrum as an index for gait pattern matching, and divide the Doppler spectrum by each step;

[0010] Calculate the gait characteristics according to the filtered range spectrum peak value and the segmented Doppler spectrum;

[0011] Input the normalized gait characteristics into the trained machine learning model for feature evaluation, and obtain the UPDRS-III score of the patient.

[0012] Preferably, the millimeter wave radio frequency signals in the environment are frequency-modulated continuous wave signals emitted by multiple transmitting antennas of a transmitting antenna array, and the transmitted signals are received by a millimeter wave receiving antenna array.

[0013] Preferably, the processing of the signals to obtain the range spectrum reflecting the change in the position of the Parkinson's patient's body and the Doppler spectrum reflecting the change in the Parkinson's patient's body degree includes:

[0014] averaging the received signals to obtain a static component generated by objects in the environment, removing the static component from the received signals to obtain a signal containing only the body movement of the Parkinson's disease patient;

[0015] supplementing the signal from which the static component is removed with the phase difference between the signals received by the multiple antennas in the receiving antenna array to realize multi-channel signal enhancement;

[0016] performing fast Fourier transform on the enhanced signal to obtain a distance spectrum reflecting the position change of the Parkinson's disease patient and a Doppler spectrum reflecting the speed change of the Parkinson's disease patient.

[0017] Preferably, the distance spectrum is extracted for each time point by a filtering method, including:

[0018] The distance spectrum is replaced with the average value of the distance spectrum energy in a specified length time window by using average filtering;

[0019] The position corresponding to the point with the maximum distance spectrum energy at each time point is extracted to obtain a position change curve of the patient, the curve is smoothed and outliers are removed by using a Hamming filter and an average filter, and finally the filtered distance spectrum peak value representing the position information of the patient is obtained.

[0020] Preferably, the cross-correlation coefficient of the Doppler spectrum is calculated as an index for gait pattern matching, and the Doppler spectrum is segmented according to each step, including:

[0021] The Doppler spectrum is regarded as a two-dimensional image, and the cross-correlation coefficient of the image is calculated as the cross-correlation coefficient of the Doppler spectrum;

[0022] A time range is determined according to the reference gait cycle, and in the time range, the Doppler spectrum of the next step with the maximum cross-correlation coefficient with the current step Doppler spectrum is searched to determine the gait cycle of the current step;

[0023] The Doppler spectrum is segmented according to each step according to the determined gait cycle.

[0024] Preferably, gait feature calculation is performed according to the filtered distance spectrum peak value and the segmented Doppler spectrum, including:

[0025] The start and end time of each gait cycle is obtained according to the segmented Doppler spectrum, the distance of the patient to the radar at the start and end time of each gait cycle is obtained according to the filtered distance spectrum peak value, and the stride of each gait cycle is obtained.

[0026] The start time and end time of each gait cycle are subtracted to obtain step time information, and the average step speed in each gait cycle is obtained according to the stride and the step time.

[0027] Preferably, the machine learning model adopts an XGBoost model.

[0028] In a second aspect, a Parkinson gait assessment system based on millimeter wave radar signals is provided, comprising:

[0029] a transmitting antenna array for transmitting a frequency-modulated continuous wave signal;

[0030] a receiving antenna array for receiving millimeter wave radio frequency signals in the environment;

[0031] a computing device for implementing the Parkinson gait assessment of the Parkinson patient according to the Parkinson gait assessment method based on millimeter wave radar signals of the first aspect.

[0032] In a third aspect, a computing device is provided, comprising: one or more processors; memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the programs are executed by the processor to implement the steps of the Parkinson gait assessment method based on millimeter wave radar signals of the first aspect.

[0033] In a fourth aspect, the present application further provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by the processor to implement the steps of the Parkinson gait assessment method based on millimeter wave radar signals as described above.

[0034] Beneficial effects: The present application aims at the defects of the existing millimeter wave radar gait analysis technology, proposes to enhance the data of the frequency-modulated continuous wave signal through the millimeter wave radar array, then extracts the specific Parkinson gait features according to the micro-Doppler spectrum similarity, and evaluates the UPDRS-III gait score of the Parkinson patient using the machine learning method. The present application can realize the Parkinson gait feature measurement and UPDRS-III gait score evaluation on a general millimeter wave radar hardware platform through a simple software algorithm. The present application can realize the non-invasive measurement and evaluation of the Parkinson gait features in a low-cost manner, and provides the doctors with the quantitative data and diagnostic reference, while the whole data acquisition process is also very convenient for the Parkinson patients. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 is an application scenario diagram of Parkinson gait assessment based on millimeter wave radar signals.

[0036] Figure 2 is an implementation flowchart of the Parkinson gait assessment method based on millimeter wave radar signals.

[0037] Figure 3 is a method flowchart for generating data features.

[0038] Figure 4 is a method flowchart for evaluating gait features.

[0039] Figure 5 is an example of Doppler spectrum segmented by gait cycle.

[0040] Figure 6 is an example of UPDRS-III gait score evaluation of a Parkinson's patient in a day. DETAILED DESCRIPTION

[0041] The technical solutions of the present application will be further described below in conjunction with the drawings.

[0042] The existing gait evaluation technology based on millimeter wave radar signal still stays in measuring the gait features of ordinary users, without considering the abnormal gait of Parkinson's patients. The abnormal gait of Parkinson's patients such as dragging and lack of hand-leg coordination is also the fundamental problem of limiting the general millimeter wave radar signal gait evaluation scheme. The present application proposes a feature extraction algorithm for abnormal gait, and uses the extracted features to correctly evaluate the severity of Parkinson's gait. First, the multi-receiving antenna of the millimeter wave radar is used to enhance the signal and improve the accuracy of feature extraction, and an index GPMR (Gait Pattern Match Rate) for evaluating the similarity of micro-Doppler images is designed to realize the feature extraction of abnormal gait; then, a machine learning based method is used to evaluate the UPDRS-III gait score of Parkinson's patients through gait features.

[0043] Figure 1 is an application scenario of the present application. The scenario includes a millimeter wave transmitting antenna array 102. The millimeter wave transmitting antenna array 102 is composed of 3 antennas with a spacing of 5 millimeters, and the transmitted signal is a radio frequency signal of 77GHz to 81GHz, transmitted at 2000 times per second. The scenario includes a millimeter wave receiving antenna array 103. The millimeter wave receiving antenna array 103 is composed of 3 antennas with a spacing of 2.5 millimeters, and receives the signals transmitted by the millimeter wave transmitting antenna array 102. The millimeter wave receiving antenna array 103 transmits the received signals to the computing device 108, and the computing device 108 can obtain the distance spectrum and Doppler spectrum of the radio frequency signal through data processing, and further obtain the position and speed information of different objects. As shown in Figure 1As shown, there are two reflection paths from the millimeter wave transmitting antenna array 102 to the millimeter wave receiving antenna array 103: the environmental reflection path 104 is the path that reaches the millimeter wave receiving antenna array 103 after being reflected by the indoor object 101, and the target reflection path 105 is the path that reaches the millimeter wave receiving antenna array 103 after being reflected by the Parkinson patient's body 106 that is walking. The indoor object 101 is stationary, so the environmental reflection path 104 will not change, and the effect of the indoor object 101 on the received signal is constant. The Parkinson patient's body 106 is walking, so the target reflection path 105 is constantly changing, which in turn causes the received signal to change in signal strength distribution and phase as the Parkinson patient's body 106 moves. The present application removes the effect of the indoor object 101 by removing the static component of the received signal, and enhances the effect of the Parkinson patient's body 106 on the received signal by using multiple antenna arrays, extracts gait features through distance spectrum and Doppler spectrum, and evaluates them.

[0044] The present application is characterized in that the entire data acquisition process is not disturbed by stationary objects in the environment, and is not affected by light, noise, etc. The Parkinson patient's body 106 being detected does not need to wear any additional equipment. The algorithm for extracting gait features of Parkinson patients is robust to Parkinson patients with gait abnormalities. The transmitting antenna array 102 and the receiving antenna array 103 are both general commercial equipment, which only need to support the transmission of 77GHz to 81GHz radio frequency signals and support a sampling rate of 5000kps.

[0045] Figure 2 The present application is characterized in that the entire data acquisition process is not disturbed by stationary objects in the environment, and is not affected by light, noise, etc. The Parkinson patient's body 106 being detected does not need to wear any additional equipment. The algorithm for extracting gait features of Parkinson patients is robust to Parkinson patients with gait abnormalities. The transmitting antenna array 102 and the receiving antenna array 103 are both general commercial equipment, which only need to support the transmission of 77GHz to 81GHz radio frequency signals and support a sampling rate of 5000kps.

[0046] In step 201, the millimeter wave radio frequency signal transmission is realized by the transmitting antenna array 102, and the transmitted signal is a frequency modulated continuous wave (FMCW) emitted by multiple transmitting antennas. Specifically, a frequency modulated continuous wave with a frequency of 77 GHz to 81 GHz is transmitted, and the repetition period is 0.5 milliseconds, and there are 256 sampling points in each period. Of course, the frequency and bandwidth of the frequency modulated continuous wave used can be other values, which need to be set according to the desired accuracy of the final calculation.

[0047] After the data collection in step 202, the collected raw data needs to be processed (corresponding to step 203) to obtain data features (corresponding to step 204), including distance spectrum peak value and Doppler spectrum segmented according to gait. Figure 3 is the implementation process of data feature generation in the present application. As shown in Figure 3 shown, first, in step 301, the influence of the indoor object 101 on the received signal is considered. Since the indoor object 101 remains stationary, its influence on the received signal also remains unchanged. However, the Parkinson's patient's body 106 undergoes a process of static-motion-static during the collection of gait data, resulting in a total of 0 for the sum of the acceleration dynamic component and the deceleration dynamic component. Therefore, the received signal is averaged to obtain the static component generated by the indoor object 101, and the static component is removed from the received signal to obtain a signal containing only the action of the Parkinson's patient's body 106. At the same time, the removal of the influence of the object in the environment makes the device robust to different deployment environments. Then, in step 302, the signal after removing the static component is subjected to multi-channel signal enhancement, and the phase difference between the signals received by multiple antennas in the receiving antenna array 103 is used to supplement the signal received by a single antenna, thereby realizing signal enhancement without changing the hardware and detection conditions. Next, in step 303, the enhanced signal is subjected to fast Fourier transform (FFT) operation, and the range profile reflecting the position change of the Parkinson's patient's body 106 and the micro Doppler profile reflecting the speed change of the Parkinson's patient's body 106 can be obtained. After removing the static component, the Parkinson's patient's body 106 is the main reflection source, and in step 304, the peak value of the range profile at each time is extracted to obtain the position information of the target; in step 305, the GPMR (Gait Pattern Match Rate) matching is performed on the Doppler spectrum, and the two-dimensional cross-correlation coefficient of the Doppler spectrum is calculated, and the Doppler spectrum is divided according to each step. According to the peak value information of the range profile and the segmented Doppler spectrum, the data features are obtained.

[0048] The present application uses the phase difference between the signals received by multiple antennas to supplement the signal received by a single antenna in step 302. First, the interval between the receiving antennas is 2.5 mm, which results in a difference in the time of signal transmission received by each antenna; however, 2.5 mm is half a wavelength, which makes the signals received by each antenna substantially the same. Specifically, the amplitudes of the signals received by different antennas are almost the same, but the phases differ greatly. Considering the radio frequency signal as far field, the phase difference of the signals received by each antenna can reflect the delay of the signals reaching each antenna. By removing the delay and adding the signals received by multiple antennas, the multi-channel signal enhancement can be achieved.

[0049] The present application uses the distance spectrum to extract the peak value at each time in step 304, and the position change information of the Parkinson patient can be obtained. Because the position of each point on the distance spectrum corresponds to the radial distance from the object in reality to the radar, the position of the peak value on the distance spectrum corresponds to the position of the patient's body in reality. First, the average filter is used for the distance spectrum. For the distance spectrum at each time, the energy at each position is the average of the distance spectrum energy in a window with a length of 5. This can prevent the maximum energy from appearing at a non-trunk position due to the swinging arm of the patient when the patient is very close to the radar. Then, the position corresponding to the point with the maximum distance spectrum energy at each time is extracted, and the position change curve of the patient is obtained. The Hampel filter and the average filter are used to smooth the curve and exclude outliers, and the filtered distance spectrum peak value representing the position information of the patient is finally obtained. Using the filtered distance spectrum peak value can prevent the problem of misidentifying the position of the patient's hand as the position of the patient's body when the patient is very close to the radar.

[0050] The application uses GPMR matching to segment the Doppler spectrum in step 307, and the starting and ending time of each step of the patient and the speed change information of the patient in this period of time can be obtained. First, the walking process is a repeated action, and even if the patient has gait abnormalities, the speed of the patient is still periodically changed. Then, by regarding the Doppler spectrum as a two-dimensional image, and using the change of the cross-correlation coefficient of the image to reflect the periodicity of the patient's gait. This can prevent the problem that the speed cannot be accurately extracted when extracting the speed; at the same time, the cross-correlation coefficient is not sensitive to the starting speed of each step of the Doppler spectrum, and using the cross-correlation coefficient as an index of GPMR matching can ensure that it is equally applicable to the gait cycle with slow speed when the patient just starts walking and the gait cycle with fast speed when the patient walks at a constant speed. In the specific matching process, first, a reference gait cycle length of 0.5 seconds is determined, and then a search is performed in the range of 0.35 seconds to 0.9 seconds, and the Doppler spectrum of the next step with the maximum cross-correlation coefficient with the current Doppler spectrum is found, that is, the gait cycle of the current step is determined.

[0051] Figure 4 is the implementation process of gait feature evaluation in the application. As shown in Figure 4 , first, in step 401, the filtered distance spectrum peak value and the segmented Doppler spectrum are combined to calculate the gait features, and then the step length, step speed and other gait features of the detection target are obtained. Then, in step 402, the obtained gait features are normalized, and finally in step 403, the normalized features are input into the trained machine learning model for feature evaluation, and the UPDRS-III score of the patient is obtained.

[0052] The application performs gait feature calculation in step 401. First, according to the segmented Doppler spectrum, the starting time and ending time of each gait cycle can be obtained, and the step time information can be obtained by subtraction. Then, using the Hamper filter and the average filter, the curve is smoothed and the outliers are removed to obtain the filtered distance spectrum peak value, as described above. The actual distance of the patient to the radar can be obtained by the formula: the index of the distance spectrum peak value x light speed ÷ (2 bandwidth), combined with the starting time and ending time of each gait cycle of the patient, the step length of each gait cycle can be obtained. Finally, by (step length ÷ step time), the average step speed in each gait cycle can be obtained.

[0053] The feature normalization in step 402 can ensure that different gait features such as stride length and gait speed, although different in units, are distributed in the interval of 0-1, so that the machine learning model can correctly evaluate each feature. In step 403, the normalized gait features are used for feature evaluation. An XGBoost model pre-trained is used to evaluate the UPDRS-III gait score of the Parkinson's patient. Compared with the deep learning model, the XGBoost model has smaller demand for computing power, faster derivation speed, and does not need a large-scale gait feature data set for training, and is suitable for the scene of Parkinson's gait evaluation. At the same time, the XGBoost model can use the regression result as the output. In this way, the output result of the feature evaluation can be a number such as 2.3 or 1.6, which has higher accuracy than the 1 point, 2 point or 3 point obtained by the doctor observing the patient with the naked eye and referring to the UPDRS-III scale.

[0054] Figure 5 is an example of the Doppler spectrum segmented according to the gait cycle in the present application. The horizontal axis is time, and the vertical axis is velocity. The color depth of each pixel represents the size of the energy reflected by the patient's body at the corresponding position. From the figure, it can be seen that although the abnormal gait of the Parkinson's patient will cause the gait cycle to be unstable, and there is a phenomenon that the speed is faster in some cycles and slower in other cycles, the GPMR matching algorithm based on the cross-correlation coefficient can ensure correct segmentation of the abnormal Parkinson's gait.

[0055] Figure 6 is an example of the UPDRS-III gait score evaluation of a Parkinson's patient in a day. Through long-term and multiple gait feature measurements and feature evaluation, the UPDRS-III gait score fluctuation of the Parkinson's patient in a day is obtained. It can be seen that when the Parkinson's patient takes medicine, the result of the feature evaluation will immediately decrease; and as time goes on, the drug effect gradually disappears, and the result of the feature evaluation will gradually increase. It shows that the feature evaluation can correctly reflect the UPDRS-III gait of the Parkinson's patient.

[0056] The present application also provides a computing device, comprising: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the programs are implemented when executed by the processor to realize the steps of the Parkinson's gait evaluation method based on the millimeter wave radar signal as described in the first aspect.

[0057] The present application also provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by the processor to realize the steps of the Parkinson's gait evaluation method based on the millimeter wave radar signal as described above.

[0058] The above examples are only used to illustrate the technical solutions of the present application but not to limit the present application. Although the present application is described in detail with reference to the above examples, it should be understood by those skilled in the art that the specific embodiments of the present application can be modified or substituted equivalently without departing from the spirit and scope of the present application, and any modification or equivalent substitution without departing from the spirit and scope of the present application should be covered within the protection scope of the claims of the present application.

Claims

1. A method for parkinsonian gait assessment based on millimeter wave radar signals, characterized in that, The method comprises the following steps: Obtaining millimeter wave radio frequency signals in the environment, and performing data processing on the signals to obtain a distance spectrum of position change of the Parkinson's patient's body and a Doppler spectrum reflecting speed change of the Parkinson's patient's body; Extracting the peak value of the distance spectrum at each time point by filtering method, calculating the cross-correlation coefficient of the Doppler spectrum as the index of gait pattern matching, and dividing the Doppler spectrum according to each step; Calculating gait features according to the filtered distance spectrum peak value and the segmented Doppler spectrum; After normalizing the calculated gait features, inputting them into the trained machine learning model for feature evaluation to obtain the UPDRS-III score of the patient.

2. The method of claim 1, wherein, The millimeter wave radio frequency signals in the environment are frequency-modulated continuous wave signals emitted by multiple transmitting antennas of a transmitting antenna array, and the emitted signals are received by a millimeter wave receiving antenna array.

3. The method of claim 2, wherein, The data processing on the signals to obtain the distance spectrum of position change of the Parkinson's patient's body and the Doppler spectrum reflecting speed change of the Parkinson's patient's body comprises: Obtaining the static component generated by the objects in the environment by averaging the received signals, removing the static component from the received signals to obtain signals containing only the Parkinson's patient's body movements; Supplementing the signal received by a single antenna with the phase difference between the signals received by multiple antennas in the receiving antenna array to realize multi-channel signal enhancement; Performing fast Fourier transform on the enhanced signal to obtain the distance spectrum reflecting the position change of the Parkinson's patient's body and the Doppler spectrum reflecting the speed change of the Parkinson's patient's body.

4. The method of claim 1, wherein, The peak value of the distance spectrum at each time point is extracted by filtering method, which comprises: Using average filtering on the distance spectrum, replacing the distance spectrum at each time point with the average value of the distance spectrum energy in a specified length time window; Extracting the position corresponding to the point with the maximum distance spectrum energy at each time point to obtain the position change curve of the patient, smoothing the curve using the Hamming filter and the average filter and excluding outliers, and finally obtaining the filtered distance spectrum peak value representing the position information of the patient.

5. The method of claim 1, wherein, The cross-correlation coefficient of the Doppler spectrum is calculated as the index of gait pattern matching, and the Doppler spectrum is divided according to each step, which comprises: Regarding the Doppler spectrum as a two-dimensional image, calculating the cross-correlation coefficient of the image as the cross-correlation coefficient of the Doppler spectrum; Determining a time range according to the reference gait cycle, and in the time range, finding the Doppler spectrum of the next step with the maximum cross-correlation coefficient with the current step Doppler spectrum to determine the gait cycle of the current step; According to the determined gait cycle, the Doppler spectrum is divided according to each step.

6. The method of claim 1, wherein, The gait features are calculated according to the filtered distance spectrum peak value and the segmented Doppler spectrum, which comprises: According to the segmented Doppler spectrum, the start and end time of each gait cycle is obtained, and the distance from the patient to the radar at the start and end time of each gait cycle is obtained according to the filtered distance spectrum peak value, to obtain the stride of each gait cycle; Subtracting the start time from the end time of each gait cycle to obtain the step time information, and calculating the average speed in each gait cycle according to the stride and the step time.

7. The method of claim 1, wherein, The machine learning model adopts an XGBoost model.

8. A millimeter wave radar signal based Parkinsonian gait assessment system, characterized in that, It comprises: a transmitting antenna array for emitting frequency-modulated continuous wave signals; a receiving antenna array for receiving millimeter wave radio frequency signals in an environment; a computing device for implementing the method of parkinsonian gait assessment based on millimeter wave radar signals according to any one of claims 1-7 to assess gait of a parkinsonian patient.

9. A computing device, comprising: comprising: one or more processors; memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the programs, when executed by the processors, implement the steps of the method of parkinsonian gait assessment based on millimeter wave radar signals according to any one of claims 1-7.

10. A computer storage medium having stored thereon a computer program, characterized in that the computer programs, when executed by the processors, implement the steps of the method of parkinsonian gait assessment based on millimeter wave radar signals according to any one of claims 1-7.

Images