CZT-based target motion state estimation system and method

Abstract

The invention discloses a CZT-based target motion state estimation system and method, and the method comprises the steps: generating a time-frequency graph according to an IQ signal, carrying out the ground noise estimation and image segmentation processing in a preset frequency band, and detecting a target signal; doppler frequency estimation based on CZT is carried out on the target signal; then, performing interpolation processing on the Doppler frequency, and eliminating singular value points; and finally, the Doppler frequency curve of the target signal in unit time is classified, so that the motion state of the target is judged, the Doppler frequency estimation algorithm based on the CZT is higher in precision, does not need to depend on a large number of annotation samples, is lower in cost, is high in judgment condition interpretability, and is higher in motion state judgment accuracy under the condition that only IQ signals are obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of target state estimation, and in particular to a CZT-based target motion state estimation system and method. Background Art

[0002] Target motion state estimation uses sensors such as radar, cameras, and inertial measurement units to obtain relevant information such as the target's position, speed, acceleration, etc., and predicts or estimates the state of the target at a certain point in the future or over a period of time based on past and current observation data. It is widely used in video surveillance, intelligent transportation systems, unmanned driving, security management and other fields.

[0003] Detecting whether a target is stationary can help improve the accuracy and robustness of further target detection and tracking; however, target motion state detection is usually based on the target's real motion image or continuous radar echo, targeting a single echo data or a signal from a moving target. However, the effect of target motion on the Doppler frequency is very small in the shortwave band and is difficult to measure accurately. Faced with a small amount of data and the classification results are not affected by the content of the signal itself, the robustness and interpretability of deep learning algorithms are low. Summary of the invention

[0004] The main purpose of the present invention is to provide a target motion state estimation method based on CZT, aiming to solve the problems of low accuracy of target motion state estimation and low robustness and interpretability of deep learning algorithms.

[0005] In order to achieve the above object, the present invention proposes a target motion state estimation method based on CZT, and the target motion state estimation method based on CZT includes:

[0006] Collecting an IQ signal, and performing a short-time Fourier transform with an FFT point number of N on the IQ signal to generate a time-frequency diagram;

[0007] Perform background noise estimation and image segmentation processing within a preset frequency band, detect the target signal within the preset frequency band, and obtain the index point of the signal segment;

[0008] Performing Doppler frequency estimation on the target signal based on the CZT adjacent phase difference frequency offset estimation algorithm, and drawing a corresponding Doppler frequency curve according to the Doppler frequency value;

[0009] Interpolate the Doppler frequency and eliminate singular value points;

[0010] The Doppler frequency curve is classified to determine the motion state of the target.

[0011] In one embodiment, the specific steps of performing Doppler frequency estimation on the target signal using the CZT-based adjacent phase difference frequency offset estimation algorithm are as follows:

[0012] The signal segment is first processed by sampling point shift and conjugation, and then multiplied with itself;

[0013] The Doppler frequency value is calculated using a chirp Z transform.

[0014] In one embodiment, the specific steps of performing background noise estimation and image segmentation processing within a preset frequency band and detecting a target signal within the preset frequency band are:

[0015] Estimating the noise floor within a preset frequency band to obtain a noise floor reference value;

[0016] Detect a target signal within a preset frequency band whose amplitude is higher than the background noise reference value.

[0017] In one embodiment, the interpolation process is a cubic spline interpolation method.

[0018] The present invention further proposes a CZT-based target motion state estimation system, which is used to implement the above-mentioned CZT-based target motion state estimation method. The CZT-based target motion state estimation system includes:

[0019] A signal detection module, the signal detection module is used to collect IQ signals, and perform short-time Fourier transform with an FFT point number of N on the IQ signals to generate a time-frequency diagram; perform background noise estimation and image segmentation processing within a preset frequency band, and detect the target signal within the preset frequency band to obtain the index point of the signal segment;

[0020] A Doppler frequency estimation module, wherein the Doppler frequency estimation module is used to perform Doppler frequency estimation on the target signal based on the adjacent phase difference frequency deviation estimation algorithm of CZT, and draw a corresponding Doppler frequency curve according to the Doppler frequency value;

[0021] A classifier is used to classify the Doppler frequency curve to determine the motion state of the target.

[0022] The technical solution of the present invention first generates a time-frequency diagram according to the IQ signal, then performs background noise estimation and image segmentation processing within a preset frequency band and detects the target signal; then performs Doppler frequency estimation based on CZT on the target signal; then, interpolation processing is performed on the Doppler frequency and singular value points are eliminated; finally, the Doppler frequency curve of the target signal within a unit time is classified to determine the motion state of the target. The Doppler frequency estimation algorithm based on CZT has higher accuracy, does not need to rely on a large number of annotated samples, has lower cost, and has strong interpretability of the judgment conditions. The accuracy of the motion state judgment is higher when only the IQ signal is obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a schematic diagram of the overall process of the target motion state estimation method based on CZT of the present invention;

[0024] Figure 2 It is a flow chart of a Doppler frequency estimation algorithm of a target motion state estimation method based on CZT of the present invention;

[0025] In the figure: 1-signal detection module, 2-Doppler frequency estimation module, 3-classifier. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0027] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. 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.

[0028] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0029] In the description of the present invention, it should be understood that the terms "upper", "lower", "inside", "outside", "left", "right", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings, or are directions or positional relationships in which the product of the invention is usually placed when in use, or are directions or positional relationships commonly understood by those skilled in the art. These directions or positional relationships are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as a limitation on the present invention.

[0030] Furthermore, the terms “first”, “second”, etc. are merely used for distinguishing descriptions and should not be understood as indicating or implying relative importance.

[0031] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms such as "setting" and "connection" should be understood in a broad sense. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0032] Existing target motion state detection is usually based on the target's real motion image or continuous radar echo, targeting single echo data or signals emitted by moving targets. However, the impact of target motion on Doppler frequency is very small in the shortwave band and is difficult to measure accurately. Faced with a small amount of data and the classification results are not affected by the content of the signal itself, the robustness and interpretability of deep learning algorithms are low.

[0033] In order to solve the above problems, the present invention proposes a target motion state estimation method based on CZT. The specific implementation of the present invention is described in detail below with reference to the accompanying drawings.

[0034] like Figure 1-2 As shown, the target motion state estimation method based on CZT includes the following steps:

[0035] Collecting an IQ signal, and performing a short-time Fourier transform with an FFT point number of N on the IQ signal to generate a time-frequency diagram;

[0036] Perform background noise estimation and image segmentation processing within a preset frequency band, detect the target signal within the preset frequency band, and obtain the index point of the signal segment;

[0037] Performing Doppler frequency estimation on the target signal based on the CZT adjacent phase difference frequency offset estimation algorithm, and drawing a corresponding Doppler frequency curve according to the Doppler frequency value;

[0038] Interpolate the Doppler frequency and eliminate singular value points;

[0039] The Doppler frequency curve is classified to determine the motion state of the target.

[0040] In one embodiment, the specific steps of performing Doppler frequency estimation on the target signal using the CZT-based adjacent phase difference frequency offset estimation algorithm are as follows:

[0041] The signal segment is first processed by sampling point shift and conjugation, and then multiplied with itself;

[0042] The Doppler frequency value is calculated using a chirp Z transform.

[0043] In one embodiment, the specific steps of performing background noise estimation and image segmentation processing within a preset frequency band and detecting a target signal within the preset frequency band are:

[0044] Estimating the noise floor within a preset frequency band to obtain a noise floor reference value;

[0045] Detect a target signal within a preset frequency band whose amplitude is higher than the background noise reference value.

[0046] Furthermore, the interpolation process is a cubic spline interpolation method.

[0047] In this embodiment, a short-time Fourier transform with an FFT point number of N is first performed on the IQ signal collected within a unit time to obtain a time-frequency diagram of the signal; then, a noise floor estimation is performed within a preset frequency band of the target signal to obtain a noise floor reference value, that is, the average value of the frequency spectrum at a location where there is no signal in the preset frequency band; then, a detection threshold is set to intercept a signal segment whose signal amplitude is higher than the noise floor reference value; in order to prevent the beginning and end signals from being missed, this embodiment adds a time point with an all-zero frequency band at the beginning and end of the time-frequency diagram; finally, the signals whose phase difference before and after is less than the tolerance are merged and regarded as a signal segment, and then the signal segment is input into a Doppler frequency shift estimation algorithm based on CZT.

[0048] Figure 2 This is a flow chart of the Doppler frequency estimation algorithm. In this embodiment, the signal segment is shifted and conjugated by a sampling point and then multiplied by itself. Then, CZT (Chirp-Z Transform) is used to calculate its fourth spectrum line, and the distance between the peak spectrum line and the zero frequency is calculated, that is, the precise Doppler frequency value of the signal segment. Compared with the existing FFT (fast Fourier transform), the Doppler frequency estimation algorithm based on CZT has the characteristics of high precision. Therefore, this embodiment uses the adjacent phase difference frequency deviation estimation algorithm based on CZT to accurately estimate the Doppler frequency of the target signal to improve the accuracy of motion state determination.

[0049] In this embodiment, after calculating the Doppler frequencies of all signal segments, the index points of the signal segments are further mapped to real time; specifically, for any signal segment, the ms time Doppler frequency that completely belongs to the signal segment is first set as the Doppler frequency of the signal segment; if it does not completely belong to the signal segment, but the signal proportion is greater than the threshold , the Doppler frequency of the MS signal is also set to the Doppler frequency of the signal segment; for other unassigned time points, cubic spline interpolation is used to obtain a complete Doppler frequency shift curve; then, the Doppler frequency shift curve is singular value detected and reassigned to the most recent Doppler frequency value, that is, the most recent non-singular value is assigned to the time point.

[0050] In this implementation, the Doppler frequency of the moving target Micro Doppler The Doppler frequency shift caused by the radial translation motion of the target It consists of two parts, namely It should be noted that the speed of the target in the radial direction is a continuous function of time and will not change suddenly. If the target is moving, then in a short period of time, the target has a constant positive or negative speed, that is, a constant Therefore, the Doppler frequency curve within the target sampling time can be transformed by Fourier transform. If the amplitude at the 0 frequency point is close to 0, it can be considered that the target has a radial speed of 0, and the target is considered to be stationary; otherwise, the target is considered to be moving. According to the discrete Fourier transform formula The value of the Doppler curve at 0 frequency can be calculated, that is, the Doppler frequency curve sequence Average.

[0051] The target motion state estimation method based on CZT of the present invention first generates a time-frequency diagram according to the IQ signal, then performs background noise estimation and image segmentation processing in a preset frequency band and detects the target signal; then performs Doppler frequency estimation based on CZT on the target signal; then, interpolation processing is performed on the Doppler frequency and singular value points are eliminated; finally, the Doppler frequency curve of the target signal within a unit time is classified to determine the motion state of the target. The Doppler frequency estimation algorithm based on CZT has higher accuracy, does not need to rely on a large number of annotated samples, has lower cost, and has strong interpretability of the judgment conditions. The accuracy of the motion state judgment under the condition of only obtaining the IQ signal is higher.

[0052] The present invention also proposes a CZT-based target motion state estimation system, which is used to implement the above-mentioned CZT-based target motion state estimation method; the specific steps of the CZT-based target motion state estimation method refer to the above-mentioned embodiments. Since the CZT-based target motion state estimation system adopts all the technical solutions of all the above-mentioned embodiments, it at least has all the beneficial effects brought by the technical solutions of the above-mentioned embodiments, which are not repeated here one by one.

[0053] In one embodiment, the target motion state estimation system based on CZT includes:

[0054] A signal detection module 1, wherein the signal detection module 1 is used to collect an IQ signal, and perform a short-time Fourier transform with an FFT point number of N on the IQ signal to generate a time-frequency diagram; perform background noise estimation and image segmentation processing within a preset frequency band, and detect a target signal within the preset frequency band to obtain an index point of a signal segment;

[0055] A Doppler frequency estimation module 2, wherein the Doppler frequency estimation module 2 is used to perform Doppler frequency estimation on the target signal based on the adjacent phase difference frequency deviation estimation algorithm of CZT, and draw a corresponding Doppler frequency curve according to the Doppler frequency value;

[0056] The classifier 3 is used to classify the Doppler frequency curve to determine the motion state of the target.

[0057] In this embodiment, the CZT-based target motion state estimation system first detects the target signal according to the time-frequency diagram of the IQ signal through the signal detection module 1, then estimates the Doppler frequency of the target signal through the Doppler frequency estimation module 2, and then classifies the Doppler frequency curve of the target signal in unit time through the classifier 3, so as to determine the motion state of the target.

[0058] In summary, the target motion state estimation method based on CZT of the present invention has a wide range of applications, strong versatility, and low requirements on the computing power of equipment hardware; the Doppler frequency estimation algorithm based on CZT has higher accuracy; this method does not need to rely on a large number of annotated samples, has low cost, and is easy to migrate and deploy; and the judgment condition of the motion state classifier 3 is based on the Doppler frequency curve of the target motion, and the judgment accuracy of the motion state is higher and the interpretability is stronger when other information is limited.

[0059] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A target motion state estimation method based on CZT, characterized in that: The target motion state estimation method based on CZT comprises the following steps: Collecting an IQ signal, and performing a short-time Fourier transform with an FFT point number of N on the IQ signal to generate a time-frequency diagram; Perform background noise estimation and image segmentation processing within a preset frequency band, detect the target signal within the preset frequency band, and obtain the index point of the signal segment; Performing Doppler frequency estimation on the target signal based on the CZT adjacent phase difference frequency offset estimation algorithm, and drawing a corresponding Doppler frequency curve according to the Doppler frequency value; Interpolate the Doppler frequency and eliminate singular value points; The Doppler frequency curve is classified to determine the motion state of the target.

2. The target motion state estimation method based on CZT according to claim 1, characterized in that: The specific steps of performing Doppler frequency estimation on the target signal using the CZT-based adjacent phase difference frequency offset estimation algorithm are as follows: The signal segment is first processed by sampling point shift and conjugation, and then multiplied with itself; The Doppler frequency value is calculated using a chirp Z transform.

3. The target motion state estimation method based on CZT according to claim 1, characterized in that: The specific steps of performing background noise estimation and image segmentation processing within the preset frequency band and detecting the target signal within the preset frequency band are: Estimating the noise floor within a preset frequency band to obtain a noise floor reference value; Detect a target signal within a preset frequency band whose amplitude is higher than the background noise reference value.

4. The target motion state estimation method based on CZT according to claim 1, characterized in that: The interpolation process is a cubic spline interpolation method.

5. A target motion state estimation system based on CZT, characterized in that: The CZT-based target motion state estimation system is used to implement the CZT-based target motion state estimation method according to any one of claims 1 to 4, and the CZT-based target motion state estimation system includes: A signal detection module, the signal detection module is used to collect IQ signals, and perform short-time Fourier transform with an FFT point number of N on the IQ signals to generate a time-frequency diagram; perform background noise estimation and image segmentation processing within a preset frequency band, and detect the target signal within the preset frequency band to obtain the index point of the signal segment; A Doppler frequency estimation module, wherein the Doppler frequency estimation module is used to perform Doppler frequency estimation on the target signal based on the adjacent phase difference frequency deviation estimation algorithm of CZT, and draw a corresponding Doppler frequency curve according to the Doppler frequency value; A classifier is used to classify the Doppler frequency curve to determine the motion state of the target.

Images