Millimeter wave radar human body sensing detection method and system based on self-adaption low operation complexity

Abstract

The invention discloses a millimeter wave radar human body sensing detection method and system based on self-adaption low operation complexity, and the method comprises the steps: firstly carrying out the discrete Fourier transform through employing the obtained echo data, and generating a corresponding distance spectrum; and performing amplitude accumulation on the plurality of distance spectrums to obtain an accumulated distance spectrum. And calculating a cross correlation coefficient, a Pearson correlation coefficient and a structural similarity index between the distance spectrum of the current frame and the distance spectrum of the previous frame, generating a comprehensive judgment coefficient through a self-adaptive linear combination method, and detecting and judging the existence of a human body according to the comprehensive judgment coefficient. The method can dynamically adapt to human body detection requirements in different scenes by combining various correlation analysis technologies, has relatively low calculation complexity and relatively high robustness, is particularly suitable for human body sensing detection tasks in low-power-consumption application scenes, and can be applied to human body sensing detection. The human body sensing detection capability of the millimeter wave radar in the low-power-consumption application is remarkably improved, and meanwhile high detection precision and reliability are kept.

Description

Technical Field

[0001] The present invention relates to the fields of radar signal processing, millimeter wave technology and human body perception technology, and in particular to a method and system for human body perception detection based on adaptive low computational complexity millimeter wave radar. Background Art

[0002] Existing millimeter-wave radar technology has been widely used in fields such as autonomous driving, security monitoring and smart home. Its main advantages are strong penetration, good environmental adaptability and high resolution. However, traditional 24GHz millimeter-wave radar usually relies on complex signal processing algorithms such as fast Fourier transform (FFT) and high-order statistical filtering in human body detection. Although these algorithms can provide higher accuracy, they are accompanied by high computational complexity and power consumption, which limits their application in low-power scenarios.

[0003] In smart home and IoT devices, since most devices are battery-powered or run on power-constrained platforms, traditional high-complexity algorithms are difficult to meet power consumption requirements while ensuring detection accuracy. In addition, static clutter (stationary objects such as furniture or walls) interferes greatly with human detection results. Existing methods have a large amount of computation when removing clutter, which makes it difficult to meet the needs of real-time human perception. Many existing methods have a trade-off between detection accuracy and computational complexity, making it difficult to achieve efficient human detection under low power consumption conditions. This technical limitation poses a challenge to the widespread application of intelligent perception systems, and there is an urgent need to develop solutions that can balance low power consumption and efficient detection. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a method for human body perception detection based on adaptive low computational complexity millimeter wave radar, which is applicable to the human body perception detection method of low-power 24GHz millimeter wave radar, and combines adaptive algorithms and low computational complexity signal processing technology to achieve efficient human body detection and perception on a low-power hardware platform. This method is particularly suitable for intelligent monitoring, security equipment and intelligent perception systems in low-power scenarios in the Internet of Things (IoT).

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] The present invention provides a method for human body perception detection based on adaptive low computational complexity millimeter wave radar, comprising the following steps:

[0007] A single-channel millimeter-wave radar is used to obtain echo data of multiple chirp signals, and discrete Fourier transform is performed on them to generate the corresponding range spectrum;

[0008] Divide and process the distance spectra of different distance units;

[0009] Perform amplitude accumulation on multiple range spectra after static clutter removal processing, so as to obtain an accumulated range spectrum;

[0010] Calculate the mutual correlation coefficient, Pearson correlation coefficient and structural similarity index between the distance spectrum of the current frame and the previous frame, and generate a comprehensive decision coefficient through an adaptive linear combination method;

[0011] The presence of human body is detected and judged based on the comprehensive determination coefficient.

[0012] Furthermore, the mutual correlation coefficient ρ, the Pearson correlation coefficient p and the structural similarity index SSIM are calculated according to the following formulas:

[0013]

[0014] Among them, μ F and are the means of the current frame distance spectra F[n] and F1[n] respectively;

[0015] and are the variances of signals F[n] and F1[n] respectively;

[0016] is the covariance between signals F[n] and F1[n];

[0017] C1 and C2 are preset constants.

[0018] Furthermore, the comprehensive decision coefficient is calculated according to the following formula:

[0019] ∈(t)=a(t)*ρ+b(t)*p+c(t)*SSIM

[0020] Among them, ∈(t) represents the comprehensive decision coefficient; a(t), b(t), c(t) represent the adaptive weights of different correlation coefficient levels; ρ represents the mutual correlation coefficient; p represents the Pearson correlation coefficient; SSIM represents the structural similarity index.

[0021] Furthermore, the weights of the different levels of correlation coefficients are adaptively allocated and calculated according to the following formulas:

[0022]

[0023] c(t)=1―a(t)―b(t)

[0024] Among them, a(t), b(t), and c(t) represent the adaptively allocated weights, N(t) represents the current frame noise mean, and M(t) represents the amplitude value corresponding to the target.

[0025] Furthermore, the frequency and distance in the distance spectrum are calculated according to the following formula:

[0026]

[0027] Among them, f s represents the sampling rate of the radar echo signal ADC, c represents the speed of light, T represents the sweep time of a chirp signal, n represents the nth position in the range spectrum, B represents the sweep bandwidth of a chirp signal, N FFT Represents the number of discrete Fourier transform points.

[0028] Furthermore, the division of the regions is performed in the following manner:

[0029] Dividing the detection area into two areas according to the distance, the two areas including a presence sensing area and a motion sensing area;

[0030] The presence sensing area is used to detect stationary and moving human bodies through radar;

[0031] The motion sensing area is used to detect the motion state of a human body through radar.

[0032] Furthermore, the detection and judgment of the presence of a human body according to the comprehensive decision coefficient is specifically performed according to the following steps:

[0033] According to the calculated comprehensive decision coefficient ∈(t), a preset threshold is used to determine whether a human body exists in the detection area; when ∈(t) exceeds the preset threshold, it is considered that a human body exists, otherwise it is considered that no human body exists.

[0034] The millimeter-wave radar human perception detection system based on adaptive low computational complexity provided by the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The above method is implemented when the processor executes the program.

[0035] The beneficial effects of the present invention are:

[0036] The millimeter wave radar human perception detection method based on adaptive low computational complexity provided by the present invention is suitable for low-power millimeter wave radar, and has the characteristics of adaptive and low computational complexity. The method first uses a single-channel millimeter wave radar to obtain the echo data of multiple chirp signals, and performs a discrete Fourier transform (DFT) on it to generate a corresponding range spectrum (Rangebin Spectrum). Subsequently, the range spectrum of different distance units is divided and processed to remove static clutter interference. Then, the amplitude of multiple range spectra after static clutter removal is accumulated to obtain the accumulated range spectrum. On this basis, the mutual correlation coefficient, Pearson correlation coefficient and structural similarity index (SSIM) between the range spectrum of the current frame and the previous frame are calculated, and a comprehensive decision coefficient is generated by an adaptive linear combination method. Finally, the presence of the human body is detected and judged based on the comprehensive decision coefficient. By combining a variety of correlation analysis techniques, this method can dynamically adapt to the human detection needs in different scenarios, has low computational complexity and high robustness, and is particularly suitable for human perception detection tasks in low-power application scenarios.

[0037] This method has low computational complexity: the amount of complex calculations is reduced through adaptive linear combination, and the detection accuracy is improved through accumulation and correlation analysis. High robustness: this method can adaptively adjust parameters to adapt to human perception detection in various complex environments, reducing false detection and missed detection. Low power design: the algorithm is optimized to enable it to run efficiently on a low-power hardware platform, which is particularly suitable for applications in smart home and Internet of Things scenarios. The human perception detection capability of millimeter-wave radar in low-power applications has been significantly improved while maintaining high detection accuracy and reliability.

[0038] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to make the purpose, technical solution and beneficial effects of the present invention clearer, the present invention provides the following drawings for illustration:

[0040] Figure 1 2 is a system architecture diagram of this embodiment.

[0041] Figure 2 Schematic diagram of radar installation and area division in this embodiment.

[0042] Figure 3 Flow chart of the method in this embodiment.

[0043] Figure 4 : is the time domain data in this embodiment and its corresponding distance spectrum.

[0044] Figure 5 1 is the distance spectrum distribution in different empty scenes in this embodiment.

[0045] Figure 6 The change process of distance spectrum and three correlation coefficients from unmanned to manned.

[0046] Figure 7 This is the change process of the determination coefficient from no-man to man in this embodiment.

[0047] Figure 8 It is the detection result display of the host computer in this embodiment.

[0048] Fig. 9 It is the output result of human presence perception. DETAILED DESCRIPTION

[0049] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0050] Example 1

[0051] like Figure 1 As shown, the millimeter-wave radar human perception detection method based on adaptive low computational complexity provided in this embodiment has the characteristics of adaptability and low computational complexity; this method can dynamically adapt to human detection requirements in different scenarios by combining multiple correlation analysis technologies, has low computational complexity and high robustness, and is particularly suitable for human perception detection tasks in low-power application scenarios;

[0052] The millimeter wave radar human perception detection method based on adaptive low computational complexity provided in this embodiment includes the following steps:

[0053] S1. Use a single-channel millimeter-wave radar to transmit multiple chirp signals and receive their reflected echo signals;

[0054] S2. Perform discrete Fourier transform (DFT) on the received echo signal to generate a range spectrum (rangebin spectrum) of multiple range units;

[0055] S3, performing static clutter filtering on the distance spectrum of each distance unit to obtain a distance spectrum with clutter removed;

[0056] S4, performing amplitude accumulation on multiple distance spectra after clutter removal to generate a cumulative distance spectrum F(t) of the current frame;

[0057] S5, record the distance spectrum F(t-T) of the previous frame, and calculate the mutual correlation coefficient (ρ), Pearson correlation coefficient (p) and structural similarity index (SSIM) between the current frame and the previous frame;

[0058] S6. According to the changes of different correlation coefficients, weights a(t), b(t), and c(t) are adaptively allocated, where:

[0059]

[0060] c(t)=1―a(t)―b(t);

[0061] Where N(t) represents the current frame noise mean; M(t) represents the amplitude value corresponding to the target;

[0062] S7. Generate a comprehensive decision coefficient ∈(t) by linearly combining the above coefficients. The calculation formula is as follows:

[0063] ∈(t)=a(t)*ρ+b(t)*p+c(t)*SSIM;

[0064] S8. Determine whether there is a human body in the target area based on the comprehensive decision coefficient ∈(t).

[0065] The mutual correlation coefficient ρ in this embodiment is used to measure the time series similarity between the current frame and the previous frame; the Pearson correlation coefficient p is used to quantify the linear correlation between the current frame and the previous frame; the structural similarity index SSIM is used to describe the similarity between the current frame and the previous frame in visual structure;

[0066] In the adaptive weight allocation process of this embodiment, the weights a(t), b(t) and c(t) are updated using objective function values ​​based on multiple correlation levels N(t).

[0067] In this embodiment, the accumulation processing of the range spectrum includes weighted accumulation of the amplitude values ​​of the echo signals of each range unit to reduce the influence of signal noise.

[0068] In this embodiment, time filtering is performed on the decision coefficients a(t) of multiple frames to smooth the detection results and enhance the stability of human body detection.

[0069] The method in this embodiment can be applied to low-power 24GHz millimeter-wave radar for real-time detection of the presence of human bodies in the target area. It has the characteristics of low computational complexity and high robustness, and is particularly suitable for scenarios such as smart homes, security monitoring, and energy-saving control.

[0070] Example 2

[0071] The human body perception detection method for low-power 24GHz millimeter-wave radar provided in this embodiment aims to solve the problems of high computational complexity and high power consumption in the prior art. By effectively processing the millimeter-wave radar echo signal and combining it with an adaptive algorithm, this method achieves high-precision human body detection while filtering out static clutter, and maintains low computational complexity, which is suitable for low-power consumption scenarios; specifically, this method includes the following steps:

[0072] S1. Transmit multiple chirp signals through a single-channel millimeter-wave radar and receive their echo signals;

[0073] S2. Perform discrete Fourier transform on the received echo signal to generate a range spectrum (Rangebin Spectrum) of multiple range units;

[0074] S3. Divide the area into two regions (presence sensing region and motion sensing region) according to different distances, perform static clutter filtering on the distance spectrum in the motion sensing region to obtain a distance spectrum without clutter; directly calculate the distance spectrum of the signal in the presence sensing region.

[0075] S4, performing amplitude accumulation of the range spectra of the multiple chirp signal echoes in the two regions to obtain the accumulated current frame range spectrum F(t);

[0076] S5. Record the distance spectrum F(t-T) of the previous frame, recorded as F1(t), and calculate the mutual correlation coefficient (ρ), Pearson correlation coefficient (p) and structural similarity index (SSIM) between the current frame and the previous frame.

[0077] S6. Adaptively assign weights a(t), b(t), c(t) according to the levels of different correlation coefficients, where:

[0078]

[0079] c(t)=1―a(t)―b(t)

[0080] S7. Generate a comprehensive decision coefficient ∩(t) by linearly combining the above coefficients. The calculation formula is as follows:

[0081] ∈(t)=a(t)*ρ+b(t)*p+c(t)*SSIM

[0082] S8. Determine whether there is a human body in the target area based on the comprehensive decision coefficient ∈(t).

[0083] Example 3

[0084] In order to better understand the human perception detection method for low-power 24GHz millimeter-wave radar provided by the present invention, it is further described in conjunction with a specific embodiment. In the embodiment, a 24GHz low-power AT24MP1T1RS32A chip launched by the domestic company Gekong (Shanghai) Intelligent Technology Co., Ltd. is used, which contains a Cortex-M0+ processor core and a radio frequency transceiver. The system structure is as follows Figure 1 As shown. In this example, the radar is installed at a height of 1.8 meters, with a pitch angle of 10°, a horizontal field of view (FOV) of ±60°, and a coverage range of 5 meters. The detection area is divided into two sub-areas: presence sensing area (0.5m~2m) and motion sensing area (2m~5m).

[0085] For the presence sensing area, the radar needs to detect stationary and moving human bodies;

[0086] For the motion sensing area, it is only necessary to detect the motion state of the human body.

[0087] Figure 2 The schematic diagram of radar installation and detection is shown. Figure 3 The computational process of this method is demonstrated.

[0088] The processing steps of this method are as follows:

[0089] Step 1: Millimeter-wave radar signal acquisition and preprocessing

[0090] In this step, a 24GHz single-channel millimeter-wave radar is used for data acquisition. The radar transmits multiple chirp signals and receives echo signals reflected from the human body. The frequency of each chirp signal changes linearly, supporting accurate distance measurement. The received echo signal is mixed to generate an intermediate frequency signal. Subsequently, the analog-to-digital converter (ADC) samples the intermediate frequency signal to obtain discrete time domain signal data.

[0091] Figure 4 (a) in the figure shows the time domain echo patterns of multiple chirp signals.

[0092] The received time domain signal is converted to the frequency domain through discrete Fourier transform (DFT) to generate a range spectrum (Rangebin Spectrum) containing multiple range units. Each frequency component corresponds to the reflection distance and intensity characteristics of the target in space, characterizing the reflection characteristics of the human target. Through Fourier transform, the frequency characteristics of the echo signal can be effectively extracted, providing a basis for subsequent clutter filtering and feature extraction. The relationship between the frequency unit in the range spectrum and the echo range unit of the target is shown in the following formula:

[0093]

[0094] Where R represents the distance of the target; f s represents the sampling rate of the radar echo signal ADC, c represents the speed of light, T represents the sweep time of a chirp signal, n represents the nth position in the range spectrum, B represents the sweep bandwidth of a chirp signal, N FFT Represents the number of discrete Fourier transform points.

[0095] According to the above formula, the frequency units in the distance spectrum can be converted into their corresponding distance unit values. Figure 4 (b) in the figure shows the range spectrum amplitudes of multiple chirp signals.

[0096] Step 2: Static clutter filtering and range spectrum accumulation

[0097] In this preferred embodiment, different distance units are divided into two intervals:

[0098] There are two areas: presence sensing area (0.5m to 2m) and motion sensing area (2m to 5m). The advantage of the two areas is that different areas allow the system to adjust the sensitivity and algorithm according to demand, optimize resource allocation, and reduce false alarms and missed alarms. In addition, the area division makes the system better adapted to actual application scenarios, such as automatic control of equipment in smart homes or detection of personnel movement in security monitoring. Through this division, the system can optimize sensitivity and algorithm efficiency, make more accurate detection decisions based on signal characteristics at different distances, and improve overall detection performance and resource utilization efficiency.

[0099] For the motion sensing area, the frequency domain echo signal of each distance unit is clutter filtered, and the average value of each sampling point is subtracted to obtain the frequency domain echo signal after removing static clutter. Static clutter is usually generated by fixed objects (such as walls and furniture). Filtering these signals helps to enhance the dynamic reflection signal of human targets.

[0100] For the presence sensing area, all echo signals are retained, including those of stationary and moving targets.

[0101] After removing static clutter, multiple distance spectra are accumulated to enhance the strength of the reflected signal and suppress noise. The accumulation process can generate stable detection results in a short time, improving the robustness of the detection.

[0102] Figure 5 The distribution of distance spectra after accumulation in different environments is shown.

[0103] Step 3: Correlation Analysis

[0104] Based on the fact that the spectrum amplitude characteristics of stationary targets remain basically unchanged in the time dimension, this method detects the presence of human bodies through distance spectrum correlation analysis between adjacent frames. In this step, the radar system calculates the correlation coefficient (ρ), Pearson correlation coefficient (p), and structural similarity index (SSIM) between the current frame (F[n]) and the previous frame (F1[n]), and evaluates the inter-frame changes from three perspectives: signal similarity, amplitude change, and structural similarity. The calculation method is as follows:

[0105]

[0106] Among them, μ F and are the means of the current frame distance spectra F[n] and F1[n] respectively;

[0107] and are the variances of signals F[n] and F1[n] respectively;

[0108] is the covariance between signals F[n] and F1[n];

[0109] C1 and C2 are small constants introduced to avoid the denominator being zero;

[0110] N represents the sampling length of the signal;

[0111] F[i] represents the signal value at the i-th position in the current frame;

[0112] F1[i] represents the signal value at the i-th position in the previous frame;

[0113] F represents the current frame signal; F1 represents the previous frame signal;

[0114] Figure 6 The changes in the distance spectrum and three correlation coefficients from unmanned to manned are shown.

[0115] Step 4: Adaptive linear combination of correlation coefficients

[0116] The complexity of human perception detection stems from the diversity of environments and the changes in targets. Different scenarios (such as open environments or environments with occlusions) may rely differently on correlation indicators.

[0117] Therefore, this method adopts an adaptive linear combination strategy to perform a weighted combination of the three correlation coefficients to generate a comprehensive decision coefficient. The specific combination formula is as follows:

[0118] ∈(t)=a(t)*ρ+b(t)*p+c(t)*SSIM;

[0119] Among them, ρ, p and SSIM represent the cross-correlation coefficient, Pearson correlation coefficient and structural similarity index respectively, and a(t), b(t) and c(t) are adaptive weights.

[0120] The weight update function is an improvement based on the Sigmoid function. Its value depends on the noise level N(t) and the motion amplitude M(t) to dynamically allocate different correlation coefficient weights, thereby improving the detection performance of the radar system in different environments. The update process is as follows:

[0121]

[0122] c(t)=1―a(t)―b(t);

[0123] If N(t) is large, it means that the echo signal is subject to strong noise interference. At this time, the system should rely more on features with strong anti-interference ability, such as the cross-correlation coefficient.

[0124] a(t) will be smaller, and the weight of the cross-correlation coefficient will increase. If M(t) is larger, it means that the target is moving obviously, and the system can rely more on features that can capture dynamic changes, such as SSIM.

[0125] At this time, b(t) becomes smaller, the weight of SSIM increases, and the system can better detect moving targets.

[0126] c(t) provides a margin in the system for adjusting weights, ensuring that the system does not rely entirely on a single feature, but instead combines multiple correlation analyses to improve detection capabilities in complex environments.

[0127] Step 5: Human Presence Sensing

[0128] According to the calculated comprehensive decision coefficient ∈(t), by setting a suitable threshold, it is determined whether a human body exists in the detection area. When ∈(t) exceeds the preset threshold, it is considered that a human body exists, otherwise it is determined that no human body exists.

[0129] Figure 7 The change of the decision coefficient from no one to someone at different times is shown. The decision rule for ∈(t) is as follows:

[0130] area Judgment Rules Output Presence perception (0.5m~2m) ∈(t)>0.97 No one Motion perception (0.5m~2m) ∈(t)>0.95 No one

[0131] The human perception judgment result can be used in scenarios such as intelligent security monitoring and smart home control. The system triggers related actions based on the detection results, such as alarm, light switching, camera startup, etc.

[0132] The output result of human presence perception is as follows Figure 8 and Fig. 9 As shown. Through the above specific implementation, the low-power millimeter-wave radar human perception detection method of the present invention can effectively reduce the computational complexity, while maintaining high detection accuracy under low power consumption conditions. The method is particularly suitable for low-power consumption scenarios such as smart homes and security monitoring, and has good robustness and adaptability.

[0133] The low-power millimeter-wave radar in this embodiment adopts a low-power (21-25) GHz millimeter-wave radar, preferably a low-power 24 GHz millimeter-wave radar.

[0134] The above-described embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Equivalent substitutions or changes made by those skilled in the art based on the present invention are within the protection scope of the present invention. The protection scope of the present invention shall be subject to the claims.

Claims

1. A human body perception detection method based on millimeter wave radar with adaptive low computational complexity, characterized in that: The following steps are involved: A single-channel millimeter-wave radar is used to obtain echo data of multiple chirp signals, and discrete Fourier transform is performed on them to generate the corresponding range spectrum; The distance spectra of different distance units are divided and processed respectively; the amplitudes of multiple distance spectra are accumulated to obtain the accumulated distance spectrum; Calculate the mutual correlation coefficient, Pearson correlation coefficient and structural similarity index between the distance spectrum of the current frame and the previous frame, and generate a comprehensive decision coefficient through a linear combination method; The presence of human body is detected and judged based on the comprehensive determination coefficient.

2. The method for human body perception detection based on adaptive low computational complexity millimeter wave radar according to claim 1, characterized in that: The mutual correlation coefficient ρ, Pearson correlation coefficient p and structural similarity index SSIM are calculated according to the following formulas: Among them, μ F and are the means of the current frame distance spectra F[n] and F1[n] respectively; and are the variances of signals F[n] and F1[n] respectively; is the covariance between signals F[n] and F1[n]; C1 and C2 are preset constants.

3. The method for human body perception detection based on adaptive low computational complexity millimeter wave radar according to claim 1, characterized in that: The comprehensive decision coefficient is calculated according to the following formula: ∈(t)=a(t)*ρ+b(t)*p+c(t)*SSIM Among them, ∈(t) represents the comprehensive decision coefficient; a(t), b(t), c(t) represent the adaptive weights of different correlation coefficient levels; ρ represents the mutual correlation coefficient; p represents the Pearson correlation coefficient; SSIM represents the structural similarity index.

4. The method for human body perception detection based on adaptive low computational complexity millimeter wave radar according to claim 1, characterized in that: The weights of the different correlation coefficients are adaptively allocated according to the following formulas: c(t)=1-a(t)-b(t) Among them, a(t), b(t), and c(t) represent the adaptively allocated weights, N(t) represents the current frame noise mean, and M(t) represents the amplitude value corresponding to the target.

5. The method for human body perception detection based on adaptive low computational complexity millimeter wave radar according to claim 1, characterized in that: The frequency and distance in the distance spectrum are calculated according to the following formula: Among them, f s represents the sampling rate of the radar echo signal ADC, c represents the speed of light, T represents the sweep time of a chirp signal, n represents the nth position in the range spectrum, B represents the sweep bandwidth of a chirp signal, N FFT Represents the number of discrete Fourier transform points.

6. The method for human body perception detection based on adaptive low computational complexity millimeter wave radar according to claim 1, characterized in that: The division of the areas is carried out in the following manner: Dividing the detection area into two areas according to the distance, the two areas including a presence sensing area and a motion sensing area; The presence sensing area is used to detect stationary and moving human bodies through radar; The motion sensing area is used to detect the motion state of a human body through radar.

7. The method for human body perception detection based on adaptive low computational complexity millimeter wave radar according to claim 1, characterized in that: The detection and judgment of the presence of a human body according to the comprehensive determination coefficient is specifically performed in the following steps: According to the calculated comprehensive decision coefficient ∈(t), a preset threshold is used to determine whether a human body exists in the detection area; when ∈(t) exceeds the preset threshold, it is considered that a human body exists, otherwise it is considered that no human body exists.

8. A millimeter-wave radar human perception detection system based on adaptive low computational complexity, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

Images