A jitter removal method for UAV-borne radar based on cumulative minimum entropy criterion

Abstract

The present invention proposes a method for removing jitter from unmanned aerial vehicle (UAV) radars based on the cumulative minimum entropy criterion, belonging to the field of radar detection technology. The method first selects a reference frame and generates a reference signal using an accumulation method. The reference signal and the echo signal are then summed according to different time delays to generate an envelope synthesis vector, with only local cyclic shifts performed during the summation process. The waveform entropy is used to measure the envelope sharpness of the envelope synthesis vector at different time delays, thereby obtaining the jitter amount of the data in each frame of the echo signal. The jitter amount represents the UAV radar jitter amount extracted by the present invention using the minimum entropy criterion, and the data is corrected to ultimately obtain a vital signal. The method reduces the impact of mutations caused by bad envelopes, avoids the accidental errors caused by using random signals for calculations, saves calculation time, and substantially eliminates the jitter of the radar platform.

Description

Technical Field

[0001] The present invention belongs to the field of radar detection technology, and in particular relates to a method for removing jitter from an unmanned aerial vehicle (UAV)-borne radar based on a cumulative minimum entropy criterion. Background Art

[0002] In post-disaster search and rescue missions for survivors and injured personnel outdoors, survival rates decrease as search time increases. Therefore, rapid and efficient life detection technology has long been a research hotspot. In scenarios where rescuers cannot arrive quickly or work for extended periods, air-to-ground life detection technology is an effective solution for targeted search and rescue. Drones offer the advantages of simple operation, flexible mobility, and no risk of casualties. They can remotely acquire vital signs from injured individuals, making them crucial for search and rescue missions.

[0003] There are two main approaches to using drones in conjunction with radar for life detection. One involves first searching for human targets and obtaining their approximate location using other methods, such as a drone equipped with a high-definition camera. Another drone, equipped with a radar, then flies to the designated location and releases the radar for detection. However, this approach is subject to radar damage, low efficiency, and resource waste. Another approach involves using a drone-mounted radar platform to detect life signals. Because chest displacement caused by human breathing is weak, this requires the platform to have good temporal stability during detection. However, in practice, wind and communication quality limitations make it difficult for drone platforms to maintain stability for long periods of time in hovering mode, making it impossible to detect life forms. Platform motion greater than the radar's range resolution is defined as large platform motion, while platform motion less than the radar's range resolution is defined as small platform motion. Large platform motion can cause range shift, shifting the positions of all targets (including human and stationary targets) in the radar echo data. Stationary targets also generate low-frequency components in the same frequency band as the respiratory signal, drowning out the human respiratory motion signal. Eliminating the interference of large platform motion is crucial for effective detection of important signals by airborne radar.

[0004] To remove radar jitter, some researchers have achieved success by searching for the location (wall) corresponding to the maximum value point in the preprocessed echo signal to remove radar jitter and then perform signal correction. However, this method performs poorly when the radar jitter amplitude exceeds the radar's range resolution. To address the problem of UAV platform motion compensation, some researchers have further proposed a background residual method. This method selects a reference frame, correlates the other received data with it, and then shifts each frame based on the correlation results to remove interference from the UAV platform. However, in actual flight data, due to unknown factors such as internal disturbances (such as propellers), or when the echo signal-to-noise ratio is low, the scattering characteristics of the reference target can change dramatically, making the results less reliable. Generally speaking, the use of airborne radar to detect vital signs is relatively limited and is currently still in the laboratory research stage. Therefore, it is necessary to develop a method to remove UAV platform jitter and obtain human vital signs, which can address the interference of unstable UAV platforms. Summary of the Invention

[0005] To solve the above technical problems, the present invention proposes a jitter removal method for unmanned aerial vehicle (UAV) radar based on the cumulative minimum entropy criterion. First, a reference frame is selected and a reference signal is generated by the accumulation method, which reduces the impact of mutations caused by bad envelopes and avoids accidental errors caused by using random signals for calculations; then, the reference signal and the echo signal are summed according to different time delays to generate an envelope synthesis vector. During the summation process, only local cyclic shifts are performed to save calculation time; waveform entropy is used to measure the envelope sharpness of the envelope synthesis vector under different time delays, and then the jitter amount of the data in each frame of the echo signal is obtained; the jitter amount represents the UAV radar jitter amount extracted by us using the minimum entropy criterion, the data is corrected and the life signal is finally obtained.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] A method for removing jitter from unmanned aerial vehicle radar based on the cumulative minimum entropy criterion includes the following steps:

[0008] Step S1: using a drone equipped with a single-transmitter, single-receiver ultra-wideband radar to collect life signals, where the life signals are time-domain echo signals;

[0009] Step S2: pre-process the time domain echo signal to remove the fixed background, suppress the linear drift caused by temperature and radar system, and filter out high-frequency noise;

[0010] Step S3: generating a reference signal based on the preprocessed time domain echo signal, and determining a data interval and a data matrix corresponding to the preprocessed time domain echo signal;

[0011] Step S4: Perform cyclic delay summation on the reference signal and each column of the data matrix to generate an envelope synthesis vector, calculate the sequence entropy of the envelope synthesis vector, generate the number of data offset sampling points based on the minimum entropy value offset position in the calculation result, obtain the UAV platform jitter, and correct the time domain echo signal;

[0012] Step S5: Perform Fourier transform on the corrected time domain echo signal along the slow time to obtain the corresponding time-frequency signal, add a window to the time-frequency signal in the frequency domain, intercept the frequency band range of the target breathing frequency, and search and measure the distance of the vital signal in the time-frequency signal result after intercepting the frequency band range.

[0013] The beneficial effects of the present invention are:

[0014] A fast unmanned aerial vehicle (UAV) radar jitter removal method using the cumulative minimum entropy criterion reduces the impact of mutations caused by bad envelopes by determining the reference interval and solving the reference signal, avoiding the accidental errors caused by using random signals for calculations. The reference signal is used to perform local cyclic correction on the data frame by frame, saving calculation time compared to global cyclic correction. The amount of data jitter in the extracted callback signal for each frame is basically the same as the simulation result, that is, the jitter of the radar platform can be basically eliminated by the proposed scheme, making it possible to extract the breathing signal masked by jitter in the UAV case. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of a method for removing jitter from unmanned aerial vehicle radar using a cumulative minimum entropy criterion according to the present invention;

[0016] Figure 2 Schematic diagram of a simulation scenario of through-wall life detection according to the present invention;

[0017] Figure 3 This is the radar motion platform elevation information graph added to the simulation signal;

[0018] Figure 4a 、 Figure 4b 、 Figure 4c 、 Figure 4d 、 Figure 4e is a comparison chart of the life detection results of the present invention; wherein, Figure 4a This is the unprocessed life detection result diagram. Figure 4b This is the life detection result diagram after the radar motion platform elevation information is added to the simulation signal. Figure 4c This is the life detection result diagram corrected by the maximum value method. Figure 4d This is the life detection result diagram corrected by relevant methods. Figure 4e This is a diagram of life detection results corrected by the method proposed in the present invention. DETAILED DESCRIPTION

[0019] In order to further illustrate the technical solution, experimental results and advantages of the present invention, the following describes the implementation steps of the present invention in detail with reference to specific examples. In an exemplary embodiment of the present invention, a method for removing jitter from an unmanned aerial vehicle radar based on a cumulative minimum entropy criterion is provided, such as Figure 1 A flowchart of the method is given. The specific contents of the UAV radar jitter removal method used in this example based on the cumulative minimum entropy criterion include:

[0020] Step S1: Use a drone equipped with a single-transmitter, single-receiver ultra-wideband radar to collect life signals. The life signals are time-domain echo signals, also known as radar echoes. The time-domain echo signals collected by the radar are expressed as:

[0021]

[0022] in, Represents the time domain echo signal collected by the radar, Indicates human breathing signal, is the reflected waveform of a fixed target, is the echo noise, where the above signals are all disturbed by the motion of the UAV platform; and Represents the number of sampling points in the fast time and slow time dimensions, , , M and N are the total number of sampling points in the fast time and slow time dimensions, respectively, and both are natural numbers.

[0023] Step S2: Preprocess the time domain echo signal to remove the fixed background, suppress the linear drift caused by temperature and radar system, and filter out high-frequency noise. The preprocessing mainly includes the following two steps: (1) using the linear trend suppression method to remove the influence of the fixed background and suppress the linear drift caused by temperature and radar system in the slow time dimension; (2) designing a range filter to filter out high-frequency noise caused by oversampling and unnecessary low-frequency signals.

[0024] Step S3: Solve the reference signal and determine the data interval.

[0025] The pre-processed time domain echo signal is squared and then summed along the slow time direction to obtain a new signal , new signal The size is , search for new signals The maximum value of the fast time index value at the maximum value is recorded as ;

[0026] Search each column element in the preprocessed time domain echo signal, find the maximum absolute value of each column element, and record its corresponding fast time index value , , represents the number of columns of the preprocessed time domain echo signal; traverse each column of the preprocessed time domain echo signal and find of Column, q is the absolute value threshold of the difference, The column is the fast time index value corresponding to the maximum absolute value of each column element The subscript i is the center, and the length is selected The interval is accumulated and averaged along the slow time direction, and the result is recorded as the reference signal , the reference signal Is the length of vector of

[0027] Determine the data interval, refer to the human chest thickness, UAV platform jitter amplitude, signal sampling distance accuracy, and use the new signal The fast time index value of the maximum value of The selected point of the first frame of the preprocessed time domain echo signal is selected, and the length of the selected point is selected on both sides of the selected point. The interval is recorded as the data interval, and the selected point of the second frame of the preprocessed time domain echo signal is ,in Represents the interval between each selected point, which is obtained by the minimum entropy offset position of the previous frame data. Finally, after traversing the entire pre-processed time domain echo signal, a data matrix consisting of multiple data intervals is generated. , where the data matrix Each column of vector.

[0028] Among them, in this example , .

[0029] Step S4: The reference signal With the data matrix Each frame of the loop delay summation generates the envelope synthesis vector ,in The reference signal is the first The loop delays are summed across the columns. , data matrix number of rows , where each are all envelope synthesis vectors with length , the sequence entropy of each column of envelope synthesis vector is solved by the following formula:

[0030]

[0031] Among them, min represents the minimum value operation, H() represents the entropy operation, Represents the envelope shift represented by the minimum entropy of each column.

[0032] Obtain the data matrix based on the envelope translation obtained above With reference signal The distance offset, that is, the jitter of the UAV platform, is obtained. The pre-processed time domain echo signal is processed according to the obtained distance offset. The time domain echo signal is corrected frame by frame to remove the jitter of the UAV platform. The corrected time domain echo signal is recorded as .

[0033] Step S5: Perform Fourier transform and windowing on the time domain echo signal, including performing Fourier transform along slow time to obtain time-frequency signal, windowing in the frequency domain to cut out the frequency band range of the target breathing frequency, and searching and ranging for life signals in the results.

[0034] Corrected time domain echo signal Windowing after slow-time Fourier transform to frequency domain ,The windowed frequency band is designed according to the actual human respiratory signal frequency. The designed frequency band is limited to a narrow window, is the index of the frequency dimension, is the maximum frequency value of the narrow window.

[0035] In this embodiment, Figure 2 A schematic diagram of the simulation scenario for through-wall life detection is given, which includes a radar, a breathing model simulating human breathing, and the ground. The radar elevation direction changes according to the added radar motion platform elevation information map, as shown in the following figure. Figure 3 , which is the radar motion platform elevation information graph added to the simulation signal. The horizontal axis is the number of data channels, and the vertical axis is the number of sampling points. Figure 4a , Figure 4b , Figure 4c , Figure 4d , Figure 4e These are the unprocessed life detection results, the life detection results corrected by the simulation data results, the life detection results corrected by the maximum value method, the life detection results corrected by the correlation method, and the life detection results corrected by the method proposed in the present invention. The horizontal axis is the intercepted time-frequency signal band, in Hertz, and the vertical axis is the distance between the radar and the target, in meters. Through comparison, it can be seen that the method provided by the present invention can more effectively reduce the impact of signal mutations, reduce accidental errors, and recover the required life signals from the jitter of the UAV platform.

[0036] In summary, the present invention uses the method of cumulative minimum entropy criterion to reduce the amount of calculation by selecting reference frames and generating reference signals, avoids the influence of data mutation caused by bad envelopes, and avoids accidental errors caused by using random signals for calculations; uses reference signals with the help of waveform entropy to perform local correction of data frame by frame, saves calculation time, and basically eliminates the jitter caused by the drone platform, making it possible to extract the respiratory signal that was originally obscured by the jitter.

[0037] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above are only specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for removing jitter from unmanned aerial vehicle radar based on the cumulative minimum entropy criterion, characterized in that: The steps include: Step S1: using a drone equipped with a single-transmitter, single-receiver ultra-wideband radar to collect life signals, where the life signals are time-domain echo signals; Step S2: pre-process the time domain echo signal to remove the fixed background, suppress the linear drift caused by temperature and radar system, and filter out high-frequency noise; Step S3: generating a reference signal based on the preprocessed time domain echo signal, and determining a data interval and a data matrix corresponding to the preprocessed time domain echo signal; Step S4: Perform cyclic delay summation on the reference signal and each column of the data matrix to generate an envelope synthesis vector, calculate the sequence entropy of the envelope synthesis vector, generate the number of data offset sampling points based on the minimum entropy value offset position in the calculation result, obtain the UAV platform jitter, and correct the time domain echo signal; Step S5: Perform Fourier transform on the corrected time domain echo signal along the slow time to obtain the corresponding time-frequency signal, add a window to the time-frequency signal in the frequency domain, intercept the frequency band range of the target breathing frequency, and search and measure the distance of the vital signal in the time-frequency signal result after intercepting the frequency band range.

2. The method for removing jitter from unmanned aerial vehicle radar based on the cumulative minimum entropy criterion according to claim 1, characterized in that: The step S1 comprises: The life signal collected by the radar, that is, the time domain echo signal is expressed as: ， in, It represents the life signal collected by the radar, that is, the time domain echo signal. Indicates human breathing signal, represents the reflected waveform of a fixed target, Represents echo noise, where the time domain echo signal , human breathing signal , Reflection waveform of fixed target , echo noise All are disturbed by the motion of the UAV platform; and Represents the number of sampling points in the fast time and slow time dimensions, , , M and N are the total number of sampling points in the fast time and slow time dimensions, respectively, and both are natural numbers.

3. The method for removing jitter from unmanned aerial vehicle radar based on the cumulative minimum entropy criterion according to claim 2, characterized in that: The step S2 comprises: Step S2.1: Use the linear trend suppression method to remove the influence of the fixed background and suppress the linear drift caused by temperature and radar system; Step S2.2: Design a range filter to filter out high-frequency noise caused by oversampling and unwanted low-frequency signals.

4. The method for removing jitter from unmanned aerial vehicle radar based on the cumulative minimum entropy criterion according to claim 3, characterized in that: The step S3 comprises: The pre-processed time domain echo signal is squared and then summed along the slow time direction to obtain a new signal , new signal The size is , search for new signals The maximum value of the fast time index value at the maximum value is recorded as ; Search each column element in the preprocessed time domain echo signal, find the maximum absolute value of each column element, and record its corresponding fast time index value , , represents the number of columns of the preprocessed time domain echo signal; traverse each column of the preprocessed time domain echo signal and find of Column, q is the absolute value threshold of the difference, The column is the fast time index value corresponding to the maximum absolute value of each column element The subscript i is the center, and the length is selected The interval is accumulated and averaged along the slow time direction, and the result is recorded as the reference signal , the reference signal Is the length of vector of With reference to the thickness of human chest cavity, the vibration amplitude of UAV platform and the distance accuracy of signal sampling, the new signal The fast time index value of the maximum value of The selected point of the first frame of the preprocessed time domain echo signal is selected, and the length of the selected point is selected on both sides of the selected point. The interval is recorded as the data interval, and the selected point of the second frame of the preprocessed time domain echo signal is ,in Represents the interval between each selected point, which is obtained by the minimum entropy offset position of the previous frame data. Finally, after traversing the entire pre-processed time domain echo signal, a data matrix consisting of multiple data intervals is generated. , where the data matrix Each column of vector.

5. The method for removing jitter from unmanned aerial vehicle radar based on the cumulative minimum entropy criterion according to claim 4, characterized in that: The step S4 comprises: First, the reference signal With the data matrix Each column of cyclic delay summation generates the envelope synthesis vector ,in Represents the reference signal is the data matrix No. The loop delay summation performed by the column, , data matrix number of rows , where each envelope synthesis vector The length is , the sequence entropy of each column of envelope synthesis vector is solved by the following formula: ， Among them, min represents the minimum value operation, H() represents the entropy operation, represents the envelope shift represented by the minimum entropy of each column; Obtain the data matrix based on the obtained envelope translation With reference signal The distance offset of the two is the same in value, which is also the jitter of the UAV platform. According to the obtained distance offset, the pre-processed time domain echo signal is processed and the time domain echo signal is corrected frame by frame to remove the jitter of the UAV platform. The corrected time domain echo signal is recorded as .

6. The method for removing jitter from unmanned aerial vehicle radar based on the cumulative minimum entropy criterion according to claim 5, characterized in that: The step S5 comprises: Corrected time domain echo signal Windowing after slow-time Fourier transform to frequency domain , the windowed frequency segment is designed according to the actual human respiratory signal frequency. The designed windowed frequency segment is limited to a narrow window, and the index of the frequency dimension , is the maximum frequency value of the narrow window.

Images