Life detection radar system with three-dimensional positioning function

Abstract

The invention relates to the technical field of radar, and discloses a three-dimensional positioning life detection radar system, and the radar system comprises a phased array antenna unit which is used for transmitting radar waves and dynamically adjusting the wave beam direction according to the target position; the signal receiving unit is used for receiving a radar signal reflected by a target; the Doppler frequency shift analysis module is used for analyzing the frequency change of the received radar signal and extracting a micro-motion signal of the target; the target motion analysis module is used for tracking the motion of the target through a Kalman filtering technology; the three-dimensional reconstruction module is used for reconstructing the three-dimensional position of the target through an electromagnetic field integral equation and multi-source sensor information fusion according to the received signals; and the deep learning module is used for identifying the micro-motion signal and optimizing target classification. By adopting the phased-array antenna unit and the beam forming technology, the technical effects of carrying out accurate three-dimensional scanning and rapid dynamic beam adjustment on the target are achieved.

Description

Technical Field

[0001] The present invention relates to the field of radar technology, and in particular to a three-dimensional positioning life detection radar system. Background Art

[0002] With the advancement of science and technology, life detection radar systems are widely used in disaster relief, military reconnaissance, building safety assessment and other fields. Especially in complex environments, such as ruins and underground tunnels, accurately locating the position and status of life is crucial for rescue operations. Traditional radar systems mainly rely on the time difference of signal echoes to determine the target position. However, these systems can usually only provide two-dimensional positioning information and are greatly affected by environmental factors (such as occlusion, multipath effect, noise, etc.), making it difficult to meet the needs of high precision and high efficiency.

[0003] Most of the existing life detection radar technologies are based on mechanical antennas and two-dimensional signal analysis methods. Traditional radar systems use rotating antennas for scanning, use reflected signals to measure distance, and combine simple signal processing algorithms to estimate the distance and direction of the target. However, due to the slow mechanical scanning speed and the inability to achieve fast and flexible beam adjustment, this system often finds it difficult to provide efficient and accurate positioning information when facing dynamic targets or complex environments. At the same time, most existing technologies rely on a single sensor, which is easily interfered by environmental noise, resulting in the loss of weak signals or false alarms.

[0004] In the prior art, the mechanical scanning method of the traditional radar system has a slow response speed and poor flexibility. Especially in complex environments, signal attenuation and multipath effects make it very difficult to accurately locate the target. In addition, the sensor fusion of the prior art is insufficient, and it is difficult to ensure stable positioning accuracy when the environment changes greatly or the target moves rapidly. Therefore, the prior art cannot effectively solve the problems of low positioning accuracy and slow real-time response in dynamic targets and complex environments. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a three-dimensional positioning life detection radar system, which solves the problems of low positioning accuracy, slow response speed and insufficient multi-source sensor data fusion of the existing radar system in complex environments.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a three-dimensional positioning life detection radar system, the radar system comprising: Phased array antenna unit, used to transmit radar waves and dynamically adjust the beam direction according to the target location to cover the target area; A signal receiving unit, used for receiving radar signals reflected by the target; Doppler frequency shift analysis module, used to analyze the frequency changes of received radar signals and extract the micro-motion signals of the target; The target motion analysis module tracks the motion of the target through Kalman filtering technology and accurately obtains the dynamic position of the target; A three-dimensional reconstruction module is used to reconstruct the three-dimensional position of the target according to the received signal through electromagnetic field integral equation and multi-source sensor information fusion; A deep learning module for identifying micro-motion signals and optimizing target classification; Positioning output module, used to output the precise three-dimensional coordinates and positioning reliability of the target; The system optimization module optimizes the operating parameters of the radar system in real time according to environmental changes and target dynamics.

[0007] Preferably, the phased array antenna unit includes multiple antenna elements, and the antenna elements dynamically adjust the beam direction by controlling their phase differences, thereby achieving accurate coverage of the target area, and the phase difference of the antenna elements is adjusted according to the azimuth and elevation information of the target.

[0008] Preferably, the phased array antenna unit adjusts the phase of the transmitted signal of each antenna element according to the calculated phase difference, so that the signals transmitted by all elements in the antenna array are superimposed in space and focused in the target direction to form a focused beam.

[0009] Preferably, when the target moves or the target position changes, the phased array antenna unit calculates the new direction angle of the target in real time, dynamically adjusts the phase difference of each antenna element in the antenna array, ensures that the beam always points in the direction of the target, and continuously tracks the target.

[0010] Preferably, the signal receiving unit includes a high-performance analog-to-digital converter for converting the received radar signal into a digital signal and performing real-time processing. The sampling rate of the analog-to-digital converter is 1 GHz or above to ensure that the reflected signal of weak life forms can be captured.

[0011] Preferably, the signal receiving unit further includes a gain control module for adjusting the gain of the signal to adapt to reception requirements of different signal strengths, thereby ensuring that the reflected signal of the target can be accurately received in the case of signal attenuation or strong interference.

[0012] Preferably, the Doppler frequency shift analysis module is used to receive the radar echo signal transmitted by the signal receiving unit, and extract the micro-motion signal of the target by analyzing the frequency change of the signal. The frequency shift analysis module includes a frequency analysis unit, which is used to calculate the frequency shift amount of the radar signal, identify the relative motion of the target and generate the speed information of the target.

[0013] Preferably, the Doppler frequency shift analysis module further includes a filter for removing noise components in the radar signal, enhancing the quality of the micro-motion signal, and ensuring that the target micro-motion information extracted by frequency shift analysis can accurately reflect the dynamic characteristics of the target.

[0014] Preferably, the target motion analysis module includes a Kalman filter, which is used to predict and update the position according to the motion data of the target, and the target motion analysis module includes: Target state prediction: predicting the target's next position through the target motion model; Target status update: correct the prediction results based on the actual measurement data and update the actual position of the target; Real-time updates, continuous prediction and updates, accurately obtain the dynamic position of the target.

[0015] Preferably, the three-dimensional reconstruction module reconstructs the three-dimensional position of the target by using electromagnetic field integral equations and multi-source sensor data fusion, and the three-dimensional reconstruction module includes: Signal preprocessing, denoising and filtering radar signals and sensor data; Calculate the electromagnetic field integral equation to estimate the target position; Data fusion combines sensor data with radar signals to output precise three-dimensional coordinates.

[0016] The present invention provides a three-dimensional positioning life detection radar system. It has the following beneficial effects: 1. The present invention achieves the technical effect of accurate three-dimensional scanning of the target and fast dynamic beam adjustment by adopting phased array antenna units and beamforming technology. Compared with the traditional mechanical radar antenna system in the prior art, the present invention can adjust the beam direction in real time, ensure high-precision positioning in complex environments, and solve the shortcomings of slow response speed and limited coverage of the mechanical scanning system.

[0017] 2. The present invention combines Doppler frequency shift analysis with deep learning algorithms to achieve accurate extraction and recognition of micro-motion signals of life. This technical effect enables the radar system to accurately identify micro-motion signals of life in strong noise and complex environments, solving the problem that it is difficult to effectively distinguish micro-motion signals in high-noise environments in the prior art.

[0018] 3. The present invention adopts Kalman filtering and multi-source sensor data fusion technology to effectively improve the accuracy of target dynamic tracking and three-dimensional positioning. By combining multiple sensor data, the present invention can still maintain high-precision positioning when the target moves and the environment changes greatly. Compared with the traditional system relying on a single sensor, it solves the impact of signal attenuation or occlusion on positioning accuracy.

[0019] 4. The present invention introduces adaptive control and real-time system optimization mechanisms to achieve dynamic adjustment of radar system operating parameters, ensuring that the system always maintains optimal performance under different environments. Compared with the solution of fixed operating parameters in the prior art, the present invention can automatically optimize system settings based on target dynamics and environmental feedback, avoiding the problem of reduced positioning accuracy under complex or unstable conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of the process flow of the phased array antenna unit of the present invention; Figure 2 It is a schematic diagram of the signal receiving unit flow of the present invention; Figure 3 It is a schematic diagram of a Doppler frequency shift analysis module of the present invention; Figure 4 It is a schematic diagram of a target motion analysis module of the present invention; Figure 5 It is a schematic diagram of a three-dimensional reconstruction module of the present invention; Figure 6 It is a schematic diagram of the deep learning module of the present invention; Figure 7 It is a schematic diagram of a positioning output module of the present invention; Figure 8 It is a schematic diagram of the system optimization module of the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings of the present invention specification to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] Please refer to the attached Figure 1 -Attached Figure 8 The embodiment of the present invention provides a three-dimensional positioning life detection radar system, the radar system comprising: Phased array antenna unit, used to transmit radar waves and dynamically adjust the beam direction according to the target location to cover the target area; A signal receiving unit, used for receiving radar signals reflected by the target; Doppler frequency shift analysis module, used to analyze the frequency changes of received radar signals and extract the micro-motion signals of the target; The target motion analysis module tracks the motion of the target through Kalman filtering technology and accurately obtains the dynamic position of the target; A three-dimensional reconstruction module is used to reconstruct the three-dimensional position of the target according to the received signal through the electromagnetic field integral equation and multi-source sensor information fusion; A deep learning module to identify micro-motion signals and optimize target classification; Positioning output module, used to output the precise three-dimensional coordinates and positioning reliability of the target; The system optimization module optimizes the operating parameters of the radar system in real time according to environmental changes and target dynamics.

[0023] The phased array antenna unit includes multiple antenna elements. The antenna elements dynamically adjust the beam direction by controlling their phase difference, thereby achieving accurate coverage of the target area. The phase difference of the antenna elements is adjusted according to the azimuth and elevation information of the target.

[0024] The phased array antenna unit adjusts the phase of the transmitted signal of each antenna element according to the calculated phase difference, so that the signals transmitted by all elements in the antenna array are superimposed in space and focused in the target direction to form a focused beam.

[0025] When the target moves or its position changes, the phased array antenna unit calculates the new direction angle of the target in real time and dynamically adjusts the phase difference of each antenna element in the antenna array to ensure that the beam always points in the direction of the target and continuously tracks the target.

[0026] The signal receiving unit includes a high-performance analog-to-digital converter, which is used to convert the received radar signal into a digital signal and perform real-time processing. The sampling rate of the analog-to-digital converter is 1GHz or above, ensuring that the reflected signals of weak life forms can be captured.

[0027] The signal receiving unit further includes a gain control module for adjusting the gain of the signal to adapt to the reception requirements of different signal strengths, ensuring that the reflected signal of the target can be accurately received in the case of signal attenuation or strong interference.

[0028] The Doppler frequency shift analysis module is used to receive the radar echo signal transmitted by the signal receiving unit, and extract the micro-motion signal of the target by analyzing the frequency change of the signal. The frequency shift analysis module includes a frequency analysis unit, which is used to calculate the frequency shift of the radar signal, identify the relative motion of the target and generate the speed information of the target.

[0029] The Doppler frequency shift analysis module further includes a filter for removing noise components in the radar signal, enhancing the quality of the micro-motion signal, and ensuring that the target micro-motion information extracted through frequency shift analysis can accurately reflect the dynamic characteristics of the target.

[0030] The target motion analysis module includes a Kalman filter for predicting and updating the position according to the motion data of the target. The target motion analysis module includes: Target state prediction: predicting the target's next position through the target motion model; Target status update: correct the prediction results based on the actual measurement data and update the actual position of the target; Real-time updates, continuous prediction and updates, accurately obtain the dynamic position of the target.

[0031] The three-dimensional reconstruction module reconstructs the three-dimensional position of the target through electromagnetic field integral equation and multi-source sensor data fusion, and the three-dimensional reconstruction module includes: Signal preprocessing, denoising and filtering radar signals and sensor data; Calculate the electromagnetic field integral equation to estimate the target position; Data fusion combines sensor data with radar signals to output precise three-dimensional coordinates.

[0032] In this embodiment, the phased array antenna unit combines a variety of technical means and algorithms to achieve real-time scanning and positioning of the target area. Its specific working process includes: transmitting signals, receiving reflected signals, and dynamically adjusting through beamforming technology, so that the radar can accurately detect the target and obtain signals. In this embodiment, the antenna unit works through phased array beamforming, using a digital method to quickly adjust the beam direction and scan different targets, thereby minimizing signal interference and improving the accuracy of target positioning.

[0033] Generally speaking, the basic structure of a phased array antenna consists of multiple unit antennas, each of which can individually control the phase of its transmitted signal, thereby changing the direction of the beam. In this way, fast direction adjustment and scanning can be achieved without moving the physical antenna. Each antenna unit works in coordination with other units to form an overall antenna array to achieve the purpose of adjusting the direction of the transmitted signal beam.

[0034] As an option, the phase adjustment of each antenna unit is achieved by controlling the time delay of the signal. Specifically, by adjusting the phase of each antenna unit, the signals emitted from different antenna units are superimposed in a specific direction, thereby forming a concentrated beam with strong directivity and accuracy. This method ensures that the signal can be concentrated to the target area, effectively enhancing the signal strength and reducing interference from other directions.

[0035] In the implementation process of the present invention, the phased array antenna achieves three-dimensional space coverage by adjusting the beam direction. The adjustment of the horizontal beam and the vertical beam is controlled by the phase difference of different units in the antenna array. The horizontal and vertical beam scanning angles are usually adjusted quickly by electronic control. This method has high accuracy and can dynamically adjust the angle of target positioning as needed, avoiding the limitations of the scanning range of traditional radar systems.

[0036] Specifically, through the electronic scanning technology of the phased array antenna, the radar system can achieve fast all-round scanning. During the target scanning process, the radar wave will be emitted within a predetermined angle range, and the beam direction will be adjusted at different time points for coverage. The system adjusts the emission angle of the radar wave according to the position of the target in space, so that the target can be accurately scanned in three-dimensional space, especially in complex environments, to ensure that the radar can capture effective reflected signals.

[0037] In one possible implementation, the phased array antenna unit can construct beam patterns (directivity patterns) at different angles by setting the phase difference of multiple antenna units. At this time, the phase difference of each antenna unit is adjusted through precise electronic control to ensure the precise alignment of the radar beam in space. For example, the horizontal angle of the radar wave and vertical angle Dynamic adjustment can be performed by controlling the phase difference of each antenna unit to ensure that the beam can scan the predetermined area.

[0038] Specifically, the beamforming can be calculated by the following formula:

[0039] in, For the antenna array in the direction The beam response on is the weighting coefficient of the nth antenna unit, indicating the power distribution of the antenna unit; is the component of the wave number in different directions, which determines the propagation direction of the beam; is the spatial position of the nth antenna element.

[0040] By adjusting the weighting coefficient of each antenna element and phase difference, directional beams at different angles can be formed, thereby achieving accurate scanning of the target area.

[0041] In some embodiments, the radar system can change the direction of the beam in real time by adjusting the phase difference of the antenna units in the array, thereby performing rapid scanning in different directions. This technology can support the radar system to quickly switch in different directions and adapt to the needs of target positioning in complex environments.

[0042] Through the above-mentioned technology, the use of phased array antenna units enables the radar system to achieve efficient and accurate three-dimensional scanning. Its dynamic beam control function can adapt to changes in targets in various environments, providing higher flexibility and response speed than traditional mechanical scanning radars. The system does not need to physically move the antenna, but only needs to adjust the direction of the electronic beam to achieve precise positioning and scanning. This not only improves the response speed of the system, but also enhances its adaptability in complex environments, especially in disaster relief, military reconnaissance and other scenes that require rapid response, which has significant advantages.

[0043] In addition, the phased array antenna unit improves the positioning accuracy and detection range of the system through rapid beamforming. In complex three-dimensional environments, flexible beam adjustment effectively reduces signal interference from different directions and improves the system's anti-interference ability.

[0044] The implementation of this antenna technology not only makes the radar system's scanning range wider, but also can quickly adjust the beam direction through electronic control, further improving the efficiency and accuracy of target positioning.

[0045] In this embodiment, the signal receiving unit is composed of multiple core modules, including a low noise amplifier (LNA), a mixer (Mixer), an analog-to-digital converter (ADC), and a bandpass filter (BPF), etc. Each module plays a key role in the signal receiving and processing process to ensure that the signal quality meets the system requirements.

[0046] Generally speaking, radar signals are affected by different environments during propagation, including path attenuation, multipath effects, and random noise interference. Therefore, necessary front-end processing must be performed before the signal enters the subsequent data processing module. The function of the signal receiving unit is to extract and enhance the reflected signal with high precision so that subsequent signal analysis can obtain more accurate results.

[0047] Specifically, after the radar signal reaches the receiving end, it first enters the low noise amplifier (LNA), which is mainly used to amplify the weak echo signal received while maintaining a low noise figure to avoid excessive interference during the amplification process. The gain of the LNA usually needs to be in the range of 20-40dB to ensure that the signal is strong enough without causing nonlinear distortion due to excessive gain.

[0048] In some embodiments, the noise figure of the LNA may be calculated by the following formula:

[0049] in, It represents the noise figure of the low noise amplifier (unit: dB). is the noise factor of the amplifier, is the gain of the amplifier in dB.

[0050] As an option, the LNA adopts a low-power and high-linearity design to meet the needs of the radar system for long-term operation while reducing the impact of thermal noise on the received signal. After the low-noise amplifier, the signal enters the bandpass filter (BPF), which is used to remove irrelevant frequency components to ensure that only signals in the target frequency band enter the subsequent processing unit.

[0051]

[0052] in, is the center frequency of the filter (in Hz), and are the lower and upper cutoff frequencies of the filter respectively (unit: Hz).

[0053] Bandpass filters can effectively suppress background noise and improve the signal-to-noise ratio (SNR). This system uses narrowband filtering to avoid the impact of multipath interference on the signal and ensure that the target signal has a high signal quality at the receiving end.

[0054] After the signal passes through the filter, it enters the mixer, which is used to convert the radio frequency signal (RF) into an intermediate frequency signal (IF). The working principle of the mixer is based on multiplication mixing, that is, the input signal is multiplied by the local oscillator signal (LO) to generate an intermediate frequency signal. .

[0055]

[0056] in, is the intermediate frequency signal frequency (unit: Hz), is the received RF signal frequency (unit: Hz), is the frequency of the local oscillation signal (unit: Hz).

[0057] The down-conversion operation of the mixer makes it easier for the system to process the signal and improves the dynamic range of the signal. The intermediate frequency signal is then sent to the analog-to-digital converter (ADC) for digitization.

[0058] In this embodiment, the analog-to-digital converter (ADC) is used to convert the analog signal into a digital signal for use by the subsequent Doppler analysis and target tracking modules. The sampling rate of the ADC needs to satisfy the Nyquist sampling theorem, that is:

[0059] in, is the sampling rate of the ADC (unit: Hz), is the highest frequency component of the signal (unit: Hz).

[0060] As an option, the sampling rate of the ADC in this system is set at 1 GHz or above to ensure that the micro-motion signals of the living body can be captured. A higher sampling rate can improve the time resolution, allowing the Doppler frequency shift analysis module to more accurately extract the motion information of the target.

[0061] In some embodiments, in order to further improve the quality of the received signal, the signal receiving unit is also equipped with an automatic gain control (AGC) module. The AGC can dynamically adjust the receiving gain according to the strength of the signal to ensure that the system can work stably and the signal can still be maintained within a suitable dynamic range even if the target is at different distances and environments.

[0062] In practical applications, radar systems may need to adapt to different environments, such as building ruins, underground structures, or open areas. In these situations, the strength and noise level of the received signal may change significantly, so the dynamic adjustment capability of the AGC module is crucial for the stable operation of the system.

[0063] Specifically, the AGC usually works by adjusting the gain of the amplifier. To maintain the amplitude of the received signal within an appropriate range, the control equation is as follows:

[0064] in, is the new amplifier gain (in dB), is the previous gain value (unit: dB), is the coefficient of control gain, is the target output voltage, is the current received signal voltage.

[0065] Through the above control method, AGC can effectively adjust the amplitude of the received signal, so that the radar system can adapt to the target detection needs in different environments.

[0066] Generally speaking, when a radar signal encounters a moving target, the target's movement will cause the frequency of the echo signal to change. This change is called Doppler shift. This phenomenon can be used to analyze the target's direction and speed. The Doppler shift analysis module is based on this principle and detects and analyzes the frequency shift of the reflected signal to obtain the target's movement information, especially the micro-motion information of living organisms.

[0067] Specifically, the Doppler frequency shift analysis module extracts the target's relative speed information by calculating the change in the frequency of the reflected signal. This speed information reflects the target's micro-motion characteristics, which in turn helps the system identify the existence and movement of living things. For the radar system, the target's relative speed Frequency Shift There is the following relationship between them:

[0068] in, is the frequency shift in Hertz (Hz), is the relative speed of the target in meters per second (m / s), is the speed of light, about Meters per second, is the original frequency of the radar signal, in Hertz (Hz).

[0069] In the above formula, the frequency shift It reflects the relative motion of the target. Based on this relationship, the relative speed of the target can be inferred by measuring the frequency change of the echo signal. This process is crucial for life detection, because the micro-motion signals of living organisms (such as heartbeat and breathing) usually cause very small frequency shifts. Accurate frequency shift analysis can enable the system to extract the micro-motion signals of the target from complex background noise.

[0070] As an option, the Doppler frequency shift analysis module in this embodiment can also be combined with a machine learning algorithm to further optimize the signal processing and analysis process. The machine learning algorithm can learn a more accurate feature extraction method by training a large number of sample signals, thereby improving the recognition accuracy of micro-motion signals. Especially in a multi-noise environment, the deep learning model can effectively reduce the interference of background noise on the frequency shift signal and improve the recognition ability and stability of the system.

[0071] Specifically, the Doppler frequency shift analysis module converts the signal into frequency domain information by performing Fourier transform on the frequency of the echo signal. By analyzing the frequency domain signal, the system can calculate the frequency shift of the target and further infer the relative movement speed of the target. When detecting the micro-motion signal of a living body, the module can accurately extract the frequency characteristics of the micro-motion of the target. For example, when the target is a person's heartbeat or breathing, these tiny movements will cause slight changes in the frequency shift. Through high-precision frequency shift analysis, the system can identify these micro-motion characteristics.

[0072] In some cases, when the target moves at a high speed, the system can further improve the accuracy of frequency shift detection by sampling multiple times. For example, in Doppler frequency shift analysis, by increasing the number of sampling points, high-frequency fluctuations and instantaneous frequency shifts can be captured more accurately, thereby improving the detection capability of fast-moving targets.

[0073] In one possible implementation, in order to improve the accuracy of the analysis, the frequency shift analysis module will also use adaptive filtering technology to suppress noise. In complex environments, radar signals are interfered by many factors, such as clutter and reflection interference. The adaptive filter can dynamically adjust the filtering parameters according to the changes in the signal, effectively remove irrelevant interference, and retain the effective frequency shift signal related to the target. This method can significantly improve the detection accuracy of micro-motion signals and reduce the false alarm rate.

[0074] In some embodiments, when the target is at a long distance or in an obstructed area, the frequency shift analysis module will also combine signal enhancement technology to enhance weak echo signals from a long distance or in a complex environment. These signal enhancement technologies can use signal averaging, time-frequency analysis and other methods to extract weak signals, further improving the sensitivity and reliability of the Doppler frequency shift analysis module.

[0075] As an implementation method, the Doppler frequency shift analysis module can also be combined with the target motion analysis module to form a closed-loop control. The target velocity information provided by the Doppler frequency shift analysis module will be passed as input to the Kalman filter module for target position correction and dynamic tracking. In this way, frequency shift analysis and target motion analysis complement each other to achieve accurate tracking of the target.

[0076] Generally speaking, in complex environments, the target's motion information is crucial for accurate positioning, especially for micro-motion monitoring of living organisms. When the target moves relative to the radar system, the target position changes continuously, and the traditional static positioning method cannot meet the needs. The target motion analysis module uses Kalman filtering technology to continuously correct the target's position information and improve positioning accuracy by accumulating and updating historical data. This module is combined with the Doppler frequency shift analysis module, and uses the target speed information extracted by the Doppler effect to provide a basis for subsequent dynamic target tracking and prediction.

[0077] Specifically, the target motion analysis module uses the Kalman filter algorithm to estimate the state of the target. Kalman filtering is a recursive algorithm that uses the error between the prediction model and the measured value to gradually update the state estimate of the system. The process consists of two main steps: prediction and update. First, the target state is predicted based on the previous state and motion model of the target; then, the state estimate is updated based on the actual observation data (such as the target speed obtained by Doppler shift analysis).

[0078] The core formula of Kalman filtering consists of two parts: prediction equation and update equation. The prediction equation predicts the next moment based on the state estimate of the previous moment, while the update equation corrects the previous estimate based on the new observation data.

[0079] In the prediction phase, the state of the target (including position and velocity) is predicted based on the motion model. The target state vector can be expressed as:

[0080] in, is the state vector of the target, including position and velocity information; is the state transfer matrix, which describes the motion law of the target; is the control input matrix, which represents the influence of external factors on the target motion; It is the control input and is usually used to describe the external influences on the system, such as wind speed, weather and other external factors.

[0081] Through the state equation, the system can predict the position and speed of the target at the current moment based on the state vector at the previous moment.

[0082]

[0083] in, is the observation vector, which usually contains the target velocity information provided by the Doppler frequency shift analysis module; is the observation matrix, which represents the relationship between the state vector and the observation; is the observation noise.

[0084] By comparing the differences between the predicted values ​​and the actual observed values, the Kalman filter algorithm is able to calculate the optimal state estimate, thereby providing real-time corrections to the target position and velocity.

[0085] As an option, the Kalman filter module can also combine multiple sensor information for state estimation. For example, in addition to the speed information obtained through Doppler shift analysis, the system can also combine data from other sensors, such as temperature sensors, infrared sensors, etc., to further improve the accuracy and robustness of target tracking. This multi-source data fusion strategy can effectively improve the reliability of target motion analysis.

[0086] In one possible implementation, in order to improve the robustness of the Kalman filter, the system can set an adaptive adjustment mechanism to automatically adjust the parameters of the Kalman filter according to environmental changes. For example, in an environment with strong multipath effects or large signal noise, the filter gain and noise covariance matrix can be dynamically adjusted to ensure the best performance of the system in different environments.

[0087] Specifically, the Kalman filter recursively corrects the target state at every moment, effectively avoiding positioning errors caused by external interference. By continuously fusing observation information from different sensors, the target motion analysis module can track the precise position of the target in real time, providing reliable data support for the subsequent 3D reconstruction module. This real-time update feature is particularly suitable for tracking and positioning fast-moving targets.

[0088] In some embodiments, the Kalman filter algorithm of the target motion analysis module can also be applied to target trajectory prediction. By learning the target's historical motion state, the system can predict the target's possible path in advance, so as to make corresponding adjustments in advance and optimize subsequent radar signal transmission and reception.

[0089] Generally speaking, the signal processing of radar systems is often affected by environmental interference, clutter and other factors, resulting in large errors in traditional two-dimensional positioning methods. The three-dimensional reconstruction module in this embodiment uses a more complex algorithm and multi-source sensor data fusion method to calculate the precise position of the target using radar signals and other sensor information (such as temperature, infrared, etc.), effectively improving the accuracy and robustness of positioning.

[0090] Specifically, the 3D reconstruction module calculates the electromagnetic field integral equation for the received radar signal and combines the multi-source sensor information to estimate the 3D coordinates of the target. In this process, the electromagnetic field integral equation is used to describe the propagation path of the signal, and the supplementary information provided by the multi-source sensor is used to further correct the calculation of the target position, thereby compensating for the possible limitations of the radar signal itself.

[0091] In this embodiment, the 3D reconstruction module combines two technologies: electromagnetic field integral equation and multi-source sensor data fusion. The electromagnetic field integral equation describes the signal propagation process from the radar signal source to the receiver, which can help estimate the coordinates of the target. By analyzing the radar echo signal, the propagation time and propagation path of the signal from the source to the target are calculated, and then the spatial position of the target is inferred.

[0092] In this process, the electromagnetic field integral equation involved can be expressed by the following formula:

[0093] in, is the field strength in the target area; is the Green function, which represents the propagation from the source point to the observation point; is the distribution function of the scattering source, usually related to the reflective properties of the target; is the surface integration element and represents the unit area at the receiver location.

[0094] As an option, the 3D reconstruction module can also combine sensor information fusion technology to fuse data obtained from other sensors (such as infrared sensors, temperature sensors or optical sensors) with radar signals. Sensor information fusion can provide additional information support for the radar system, especially in complex environments, by fusing data from different sources, it can effectively improve positioning accuracy.

[0095] In some embodiments, sensor data fusion is mainly achieved through algorithms such as weighted averaging, Kalman filtering or particle filtering. The measurement value of each sensor is adjusted by a weighting coefficient to make the fusion of multi-source data more reliable. Specifically, the three-dimensional coordinates of the target can be expressed as:

[0096] in, is the three-dimensional coordinate of the fused target; For the Coordinate estimates provided by the sensors; For the The weight coefficient of each sensor is usually determined by the accuracy and reliability of the sensor; is the total number of sensors.

[0097] In one possible implementation, the radar system uses a weighted fusion strategy to assign weights based on the signal-to-noise ratio and stability of different sensors. In this way, the radar signal and the sensor signal can complement each other and improve the accuracy of target positioning.

[0098] Specifically, the 3D reconstruction module combines electromagnetic field integral equations and sensor information fusion technology to provide accurate target positioning in different types of environments. In some embodiments, the system may face multipath effects (such as signal reflection, refraction, etc.), which may cause the propagation path of the target signal to deviate. Therefore, the 3D reconstruction module not only relies on the radar signal itself, but also can reduce the impact of multipath effects on positioning results by integrating sensor data.

[0099] In some embodiments, the 3D reconstruction module can also respond to changes in different environments through adaptive adjustment methods. For example, in a more complex environment (such as urban ruins or underground structures), the strength of the radar signal may be weak due to signal attenuation or reflection. At this time, the fusion of sensor data will be more important. By evaluating the signal quality of different sensors in real time, the system can dynamically adjust the weight of data fusion to ensure the accuracy of the target 3D positioning result.

[0100] In some embodiments, during the data fusion process, the 3D reconstruction module can also introduce a machine learning algorithm to further optimize the calculation of the target coordinates. By learning from historical data, the machine learning model can continuously optimize the weight distribution and adjust the data fusion strategy according to different environmental conditions. This optimization strategy can improve the accuracy and robustness of target reconstruction in complex environments.

[0101] As an option, the 3D reconstruction module can also work with other modules (such as the target motion analysis module) to share real-time target position and speed information. Through multiple reconstructions and predictions, the system can dynamically update the target's position and provide more accurate data support for subsequent target classification and positioning.

[0102] Generally, the radar system extracts the relative speed information of the target through the Doppler frequency shift analysis module, and this data can describe the motion characteristics of the target. However, radar signals are often interfered by a lot of noise during propagation, especially in complex environments, where the recognition of micro-motion signals is particularly difficult. At this time, traditional signal processing methods may not provide sufficient accuracy and stability. The deep learning module can effectively extract micro-motion signals from a lot of noise and improve the recognition rate of signals through algorithms such as convolutional neural networks (CNN).

[0103] Specifically, the deep learning module in this embodiment uses a convolutional neural network (CNN) to extract features and classify targets from radar echo signals. Convolutional neural networks are a deep learning method widely used in image processing and signal analysis, and are particularly suitable for extracting spatial and temporal features from complex data. The network processes the input signal through multiple convolutional layers, continuously extracts features at different levels, and outputs target categories or other analysis results through the final fully connected layer.

[0104] As an option, the deep learning module can use a standard convolutional neural network structure, which includes multiple convolutional layers, pooling layers, and fully connected layers. Each convolutional layer processes the input signal using convolution kernels of different sizes to extract different features. The pooling layer is used to reduce the dimension of the data, reduce the amount of calculation, and enhance the model's anti-interference ability. Finally, through the fully connected layer, the network maps the extracted features to the results of target classification or position prediction.

[0105] Specifically, the input of the convolutional neural network is the radar signal after Doppler frequency shift analysis, which contains the speed and micro-motion characteristics of the target. The network extracts the spatiotemporal features in the signal through convolution operations, and processes the features through pooling layers and fully connected layers, and finally obtains the classification result or accurate three-dimensional position of the target.

[0106] In one possible implementation, in order to further improve the recognition accuracy of the model, the deep learning module can use a multi-layer convolution structure, that is, to extract more fine-grained signal features through different convolution kernel sizes and step sizes. For example, the first layer of convolution can use a smaller convolution kernel to capture low-level features in the signal, such as small changes; while deeper convolutions use larger convolution kernels to extract complex spatiotemporal patterns, thereby providing more accurate feature information for subsequent target classification and 3D reconstruction.

[0107] In some embodiments, the deep learning module can also be combined with a recurrent neural network (RNN) or a long short-term memory network (LSTM) to further enhance the model's ability to process time series data. Since micro-motion signals of living organisms (such as heartbeats and breathing) usually have temporal continuity, the use of time series processing models such as LSTM can better capture the temporal dependencies in the signals, thereby improving the recognition accuracy of micro-motion signals.

[0108] In this case, the deep learning module not only processes the current signal, but also can more accurately predict and track the target based on the information of the historical signal. For example, the heartbeat signal has a strong periodicity, and the LSTM network can effectively capture this periodic feature and further optimize the recognition of the target.

[0109] In one possible implementation, the parameters of the convolutional layers and recurrent layers in the deep learning module can be trained using a back-propagation algorithm. This process involves using a large amount of labeled data to train the model. The back-propagation algorithm calculates the gradient of the loss function and adjusts the parameters in the network to continuously reduce the training error. The loss function used in the training process can be selected according to the specific task, for example, using a cross entropy loss function for target classification or a mean square error loss function for position prediction.

[0110] Specifically, after the micro-motion signal of the target is processed by the Doppler frequency shift analysis module, the obtained signal is input into the deep learning module for feature extraction. Through the convolutional neural network, the system can identify the micro-motion characteristics of the target and determine the type or location of the target through subsequent classification or regression networks. In this way, the deep learning module can effectively reduce the false alarm rate and improve the system's ability to detect micro-motion signals of living organisms, especially in a noisy environment. It can still maintain a high recognition accuracy.

[0111] In some embodiments, the deep learning module may also use data enhancement methods to further improve the generalization ability of the model. For example, by rotating, flipping, cropping, etc. the training data, more diverse training samples can be generated, so that the model can adapt to different signal interferences and environmental changes in actual applications.

[0112] As an option, the deep learning module can also perform data fusion with other modules (such as Doppler frequency shift analysis module, target motion analysis module). During the target positioning process, the deep learning module can combine the data of other sensors to further improve the recognition accuracy and robustness of the target. For example, in complex environments, in addition to radar signals, the deep learning module can also use the signals of temperature sensors or infrared sensors to enhance the recognition ability of living things.

[0113] Generally speaking, the target position in this system is gradually obtained through the processing of radar signals. The Doppler shift analysis module extracts the target's velocity information, the target motion analysis module uses the Kalman filter method to track the target in real time, and obtains the target's spatial coordinates through the 3D reconstruction module. The positioning output module summarizes the data from each module, performs the final calculation of the target's precise position, and outputs the target's 3D coordinates.

[0114] Specifically, the workflow of the positioning output module includes two main parts: data reception and position calculation. First, the positioning output module receives the motion data from the target motion analysis module and the spatial position information from the 3D reconstruction module, and combines the classification and recognition results of the target by the deep learning module to finally obtain the 3D coordinates of the target. During the calculation process, the positioning output module relies on multiple input sources to ensure that the generated 3D coordinates have high accuracy and credibility.

[0115] As an option, the positioning output module in this embodiment uses a weighted average method to fuse the output results of each module. According to the confidence of each module output, the positioning output module assigns a weight to each result, and then calculates the weighted average position. This method helps to improve the overall accuracy of the system, especially when the target moves faster or the signal noise is large. Through weighted fusion, the impact of a module output error on the overall result can be effectively reduced.

[0116] In one possible implementation, the positioning output module uses Kalman filtering for the final positioning calculation. Kalman filtering can not only predict the position of the target, but also update the position based on the sensor data. The positioning output module combines the position information output by each sensor and the dynamic position data from other modules to calculate the final coordinates of the target using the Kalman filtering algorithm. Through Kalman filtering, the system can adjust the prediction of the target position in real time to avoid deviations caused by sensor data errors.

[0117] The formula for positioning output can be expressed as:

[0118] in, is the final target location, is the target position predicted by the motion model, is the position measured by the sensor, is the Kalman gain, which depends on the confidence of the prediction and the accuracy of the measurement.

[0119] Specifically, the Kalman gain The calculation formula is:

[0120] in, is the prediction error covariance, which indicates the system’s confidence in the target position prediction; is the observation error covariance, which represents the noise level of the measurement data.

[0121] The Kalman filter continuously adjusts the gain , according to each updated target state and observation data, the estimated value of the target position is adjusted to optimize the positioning result.

[0122] In some embodiments, the positioning output module can also combine other supplementary information (such as the target's motion pattern or environmental data) to further optimize the calculation of the target's position. For example, when the target is in a dynamic environment, the system can consider the target's historical motion trajectory and dynamic characteristics, increase the weight of historical data, so that the system can more accurately predict and locate the target's current position.

[0123] As an option, the positioning output module can also give the positioning confidence each time the positioning result is output. The positioning confidence can be based on the estimation error in the Kalman filter, or based on the confidence output by other modules (such as the deep learning module) to indicate the reliability of the target position. The higher the confidence, the higher the accuracy of the positioning result.

[0124]

[0125] in, is the final error covariance, which indicates the reliability of the target position estimate.

[0126] Confidence The closer it is to 1, the higher the reliability of the positioning result.

[0127] In one possible implementation, the calculation results of the positioning output module can be displayed through a graphical user interface (GUI). Through the visual display, users can intuitively view the real-time location of the target, helping rescuers or operators make quick decisions. In the graphical interface, the three-dimensional coordinates of the target can be displayed in a map or three-dimensional model in the form of points, lines, graphics, etc., ensuring that the positioning results are easy to understand and operate.

[0128] Generally speaking, the performance of a radar system is often affected by environmental changes and target dynamics. For example, when the signal is weak or severely blocked, the system may not be able to effectively receive the reflected signal, resulting in reduced target positioning accuracy. Therefore, the system optimization module uses a real-time feedback mechanism to adjust the operating parameters of the radar system according to environmental changes and target status, ensuring that high positioning accuracy and response speed can be maintained under different conditions.

[0129] Specifically, the working principle of the system optimization module is based on the real-time analysis of radar signals and the prediction of target motion. First, the system optimization module receives the target position and speed information from the target motion analysis module, as well as the signal quality data from other modules. Through the comprehensive analysis of this information, the system optimization module can evaluate the current radar system performance in real time and adjust the relevant working parameters according to the analysis results.

[0130] As an option, the system optimization module can use an adaptive control algorithm to automatically adjust system parameters based on the quality of the radar signal and the dynamic behavior of the target. The adaptive control algorithm automatically adjusts the radar system's operating frequency, beam direction, and signal sampling rate by calculating the target's motion state, signal noise level, and other environmental factors in real time. For example, in the case of fast target movement, the system optimization module will automatically increase the signal sampling rate to ensure accurate positioning of fast-moving targets.

[0131] In one possible implementation, the system optimization module dynamically adjusts the radar's operating frequency and beam direction according to the target's movement rate. If the target is at a long distance or moves slowly, the system can reduce the sampling frequency, thereby reducing the system load and improving computing efficiency. When the target is close to the radar system or moves quickly, the system will increase the sampling frequency and adjust the beam direction in real time to ensure that the target is always within the radar's coverage.

[0132] Specifically, to achieve this goal, the system optimization module Perform real-time estimation and adjust sampling frequency based on sensor feedback data and beam adjustment rate The dynamic adjustment of these parameters can be calculated by the following formula:

[0133]

[0134] in, is the sampling frequency of the radar signal (unit: Hz), is the highest frequency component of the signal (unit: Hz), is the minimum sampling frequency (unit: Hz), is the adjustment rate of the beam direction (unit: degrees / second), is the beam angle change (unit: degree), is the time interval (unit: seconds).

[0135] In some embodiments, in order to further improve the adaptability of the system in complex environments, the system optimization module can also adjust the operating frequency of the radar system in combination with the intensity of the ambient noise. For example, in a strong noise environment, the system optimization module can choose to lower the operating frequency to avoid overlapping with the ambient noise frequency band, reduce interference, and improve signal quality.

[0136] As an option, the system optimization module can also use the Kalman filter algorithm to make real-time corrections to the dynamic position of the target. The Kalman filter algorithm is based on the predicted position of the target at the previous moment and the observed data at the current moment, and fuses multiple data sources through weight adjustment to obtain the optimal target position estimate. The system optimization module uses the correction mechanism of the Kalman filter to dynamically adjust the operating parameters of the radar system according to the real-time position and speed changes of the target to ensure that the positioning accuracy is not affected by the external environment.

[0137] In one possible implementation, the system optimization module also combines multi-source sensor data fusion, such as environmental data provided by other sensors such as temperature sensors and infrared sensors, to comprehensively optimize the operating parameters of the radar system. In this way, the system optimization module can make full use of information from multiple sensors to further improve the adaptability and performance of the system in complex environments.

[0138] In the process of multi-source sensor data fusion, the system optimization module performs weighted fusion of data from different sensors to obtain more accurate target location and environmental parameters. The weighted fusion method can assign weights according to the data quality of each sensor, thereby ensuring that the system can integrate information from multiple sensors to obtain the best target positioning results.

[0139] Specifically, this process can be achieved through the weighted average method. The data of each sensor is adjusted by a weighting coefficient to obtain a weighted target position estimate. The calculation formula for the fused target position can be expressed as:

[0140] in, is the target position estimation after fusion; For the The target position estimate provided by the sensors; For the The weight coefficient of each sensor indicates the reliability of the sensor data; is the total number of sensors.

[0141] Through the weighted average method, the system optimization module can dynamically adjust the contribution of each sensor to the target position according to its reliability, thereby achieving accurate target positioning in different environments.

[0142] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A three-dimensional positioning life detection radar system, characterized in that: The radar system comprises: Phased array antenna unit, used to transmit radar waves and dynamically adjust the beam direction according to the target location to cover the target area; A signal receiving unit, used for receiving radar signals reflected by the target; Doppler frequency shift analysis module, used to analyze the frequency changes of received radar signals and extract the micro-motion signals of the target; The target motion analysis module tracks the motion of the target through Kalman filtering technology and accurately obtains the dynamic position of the target; A three-dimensional reconstruction module is used to reconstruct the three-dimensional position of the target according to the received signal through electromagnetic field integral equation and multi-source sensor information fusion; A deep learning module for identifying micro-motion signals and optimizing target classification; Positioning output module, used to output the precise three-dimensional coordinates and positioning reliability of the target; The system optimization module optimizes the operating parameters of the radar system in real time according to environmental changes and target dynamics.

2. The three-dimensional positioning life detection radar system according to claim 1, characterized in that: The phased array antenna unit includes multiple antenna elements, which dynamically adjust the beam direction by controlling their phase differences, thereby achieving accurate coverage of the target area. The phase differences of the antenna elements are adjusted according to the directional angle and elevation angle information of the target.

3. The three-dimensional positioning life detection radar system according to claim 1, characterized in that: The phased array antenna unit adjusts the transmission signal phase of each antenna element according to the calculated phase difference, so that the signals transmitted by all elements in the antenna array are superimposed in space and focused in the target direction to form a focused beam.

4. The three-dimensional positioning life detection radar system according to claim 1, characterized in that: When the target moves or the target position changes, the phased array antenna unit calculates the new direction angle of the target in real time and dynamically adjusts the phase difference of each antenna element in the antenna array to ensure that the beam always points to the target direction and continuously tracks the target.

5. The three-dimensional positioning life detection radar system according to claim 1, characterized in that: The signal receiving unit includes a high-performance analog-to-digital converter for converting the received radar signal into a digital signal and performing real-time processing. The sampling rate of the analog-to-digital converter is 1 GHz or above to ensure that the reflected signal of weak life forms can be captured.

6. The three-dimensional positioning life detection radar system according to claim 1, characterized in that: The signal receiving unit further includes a gain control module for adjusting the gain of the signal to adapt to the reception requirements of different signal strengths, ensuring that the reflected signal of the target can be accurately received in the case of signal attenuation or strong interference.

7. The three-dimensional positioning life detection radar system according to claim 1, characterized in that: The Doppler frequency shift analysis module is used to receive the radar echo signal transmitted by the signal receiving unit, and extract the micro-motion signal of the target by analyzing the frequency change of the signal. The frequency shift analysis module includes a frequency analysis unit, which is used to calculate the frequency shift amount of the radar signal, identify the relative motion of the target and generate the speed information of the target.

8. The three-dimensional positioning life detection radar system according to claim 1, characterized in that: The Doppler frequency shift analysis module further includes a filter for removing noise components in the radar signal, enhancing the quality of the micro-motion signal, and ensuring that the target micro-motion information extracted through frequency shift analysis can accurately reflect the dynamic characteristics of the target.

9. The three-dimensional positioning life detection radar system according to claim 1, characterized in that: The target motion analysis module includes a Kalman filter, which is used to predict and update the position according to the motion data of the target. The target motion analysis module includes: Target state prediction: predicting the target's next position through the target motion model; Target status update: correct the prediction results based on the actual measurement data and update the actual position of the target; Real-time updates, continuous prediction and updates, accurately obtain the dynamic position of the target.

10. The three-dimensional positioning life detection radar system according to claim 1, characterized in that: The three-dimensional reconstruction module reconstructs the three-dimensional position of the target through electromagnetic field integral equation and multi-source sensor data fusion, and the three-dimensional reconstruction module includes: Signal preprocessing, denoising and filtering radar signals and sensor data; Calculate the electromagnetic field integral equation to estimate the target position; Data fusion combines sensor data with radar signals to output precise three-dimensional coordinates.

Images