Frequency modulation continuous wave radar interference elimination method based on adaptive dictionary matching

Abstract

The invention belongs to the technical field of frequency modulation continuous wave radars, and particularly relates to a frequency modulation continuous wave radar interference elimination method based on adaptive dictionary matching. The method comprises the following steps: establishing a mathematical model of radar sending signals, receiving signals and interference signals; selecting different time delays and frequency modulation slopes according to the mathematical model of the interference signals to construct a dictionary with enough precision; roughly and uniformly extracting dictionary data and receiving signals to calculate correlation; according to the correlation, finer dictionary data and received signals are uniformly extracted near parameters larger than a given threshold value, and the correlation is calculated; determining interference parameters according to the correlation after the expected precision is reached; calculating interference signals according to the determined parameters and eliminating the interference signals from the received signals; according to the method, the problems of calculation complexity and precision reduction caused by a traditional interference suppression method can be avoided, the interference recognition and elimination efficiency is improved, the method is suitable for scenes where interference signals are varied frequently and environment complexity is increased, and the robustness and detection performance of a system are enhanced.

Description

Technical Field

[0001] The invention belongs to the technical field of frequency modulation continuous wave radar, and in particular relates to a method for eliminating interference of frequency modulation continuous wave radar. Background Art

[0002] In recent years, frequency modulated continuous wave radar has been widely used in the fields of autonomous driving and intelligent transportation because it can still work normally under bad weather and lighting conditions. However, with the increase in the number of vehicles equipped with radars on the road, the deployment of multiple radar devices increases the probability of interference. Inter-radar interference can cause weak target signals to be obscured, detection accuracy to be reduced, and even false targets to be generated. Current interference suppression methods include zeroing, reconstruction, estimation and separation methods based on signal processing, as well as deep learning technology. However, traditional methods have high computational complexity and cannot meet the real-time requirements under complex road conditions. Deep learning methods usually require a large amount of data training and also have the problem of high computational overhead. Therefore, an efficient interference suppression solution is urgently needed.

[0003] Dictionary matching methods have been widely used as an efficient signal processing tool. This method uses a constructed dictionary to match the features in the received signal, and by gradually searching for the best matching parameters, the characteristics of the interference signal can be accurately estimated. Compared with traditional methods, the dictionary matching method can adaptively adjust the accuracy and quickly identify the parameters of different interference sources, thereby improving signal processing efficiency and reducing errors. In addition, since this method does not rely on specific hardware or feedback mechanisms, it can be directly applied in existing systems, significantly improving spectrum efficiency and system robustness. Summary of the invention

[0004] In order to overcome the problems of poor real-time performance and high complexity brought about by traditional methods in current radar systems, the present invention proposes a frequency modulated continuous wave radar interference elimination method based on adaptive dictionary matching.

[0005] The frequency modulated continuous wave radar interference elimination method based on adaptive dictionary matching proposed in the present invention constructs a two-dimensional dictionary containing multiple parameter combinations (including delay and slope), and uses the matching degree between the received signal and the signal in the dictionary to estimate the interference parameters, thereby realizing accurate detection and elimination of interference signals. Compared with traditional feedback methods, compressed sensing methods, and methods based on deep learning, this method avoids the problems of additional feedback overhead and high computational complexity, and can quickly and accurately achieve interference suppression without increasing the hardware burden. The specific steps are:

[0006] Step 1: Establish the mathematical model of the frequency modulated continuous wave radar sending signal, receiving signal and interference signal, including the following process:

[0007] 1.1, the signal sent by the frequency modulated continuous wave radar is:

[0008]

[0009] Among them, f c To send a carrier signal; is the modulation slope, B is the bandwidth, and T is the chirp storage time;

[0010] 1.2, define the target echo signal (received signal) as:

[0011]

[0012] Among them, τ r and P r are the time delay and received power of the rth target echo signal respectively;

[0013] 1.3, define the interference signal as:

[0014]

[0015] Among them, τ i , k i and P i are the time delay, modulation slope and received power of the i-th interference signal respectively;

[0016] Step 2: Construct a high-precision dictionary based on the above signal model, including the following process:

[0017] 2.1, define the column vector in the dictionary as:

[0018] a τk =[a τk (t t1 ),a τk (t t2 ),…,a τk (t tN )] T ∈R N*1 (4)

[0019] Where τ is the time delay of the target position corresponding to the distance converted from the speed of light within the radar working range, k is the slope of common radar models, and t t1 ,t t2 ,…t tN is the sampling time, N is the number of samples for each algorithm execution, and each element is defined as:

[0020]

[0021] 2.2, construct an adaptive dictionary covering radar operating parameters, specifically:

[0022] A=[a 11 …a 1k …a τ1 …a τk ]∈R N*(τ*k) (5)

[0023] Where a, τ and k are defined in 2.1;

[0024] Step 3: Sampling the received signal and calculating the power of the received signal, including the following process:

[0025] 3.1, define the received signal as:

[0026] y rx (t) = s R (t)+s I (t) (6)

[0027] 3.2, sampling the received signal:

[0028] y rx =[y rx (t t1 ),y rx (t t1 ),…,y rx (t tN )] (7)

[0029] Among them, t t1 ,t t2 ,…,t tN They are the samples at time t1, t2, …, tN respectively;

[0030] 3.3, calculate the received power:

[0031]

[0032] Step 4: When the received power is greater than the given threshold γ1, proceed to step 5, otherwise proceed to step 10, where γ1 = μ1 + b1σ1, μ1 is the mean value of the noise calculated from the received signal power measured without a target, σ1 is the standard deviation of the noise, b1 is a multiple set according to actual needs, and b1 is 3-6; usually b1 is 3-5; for safety-related applications such as autonomous driving and advanced driver assistance systems (ADAS), it is recommended to use a higher value, such as b1 of 4-6; to reduce the false alarm rate and improve the reliability of the system;

[0033] Step 5: Sparsely and discretely select column vectors from the dictionary constructed in step 2. It is recommended to use the delay converted from the speed of light corresponding to a distance of 1 meter as the interval, including the following process:

[0034] 5.1, uniformly sparsely select column vectors from the dictionary:

[0035]

[0036] Among them, τ1, τ2, …, τ m are sparse delay data respectively. It is recommended to use the delay converted from the speed of light corresponding to a distance of 1 meter as the interval, k1, k2, …, k n are the radar slopes for sparse selection, respectively;

[0037] 5.2, calculate the correlation as:

[0038]

[0039] Step 6: Analyze the correlation distribution between the received signal and the dictionary column vector, calculate the mean and standard deviation of the correlation value in the case of no target, set the threshold γ2 = μ2 + b2σ2, μ2 is the mean of the correlation value calculated by the received signal power measured in the case of no target, σ2 is the standard deviation, b2 is set to 1-3 as needed, when the correlation is greater than the given threshold γ2, go to step 7, otherwise go to step 10;

[0040] Step 7: When the distance interval of the selected dictionary vector delay conversion is less than or equal to 1 cm, proceed to step 9, otherwise proceed to step 8;

[0041] Step 8: Select a denser dictionary around the delay and slope parameters whose correlation is greater than a given threshold γ2, including the following process:

[0042] 8.1, reselect the dictionary column vector to become a denser dictionary:

[0043]

[0044] in, It can be expressed as a column vector determined by the parameters calculated in step 5 that are greater than the given threshold γ2. The dense dictionary suggests that the dense dictionary in the previous cycle Iterate at intervals;

[0045] 8.2, go to step 3.3;

[0046] Step 9: Select the delay and slope corresponding to the maximum correlation result, construct the interference signal, and eliminate it from the received signal, including the following process:

[0047] 9.1, construct the interference signal as:

[0048]

[0049] in, and The interference delay and slope estimated in step 8, respectively, is the interference power determined according to the time delay, which is specifically calculated as:

[0050]

[0051] P t ,G t ,G t and λ are the transmit power, transmit antenna gain, receive antenna gain and signal wavelength respectively, and R is the distance of delay conversion;

[0052] 9.2, subtracting the estimated interference from the received signal;

[0053] Step 10: After the above steps, the interference is eliminated and conventional radar signal processing is performed to calculate the distance and speed of the target.

[0054] The present invention aims at the interference suppression problem in the radar system, takes into account the diversity of different interference signals in time delay and slope, and uses the idea of ​​dictionary matching to construct a two-dimensional dictionary containing multiple parameter combinations. The dictionary can simultaneously search for the time delay and frequency modulation slope of the interference signal, avoiding the algorithm complexity and accuracy problems caused by spectrum changes or environmental complexity in traditional methods. By matching the received signal with the elements in the dictionary, the specific parameters of the interference signal can be quickly identified. In addition, in order to cope with the uncertainty caused by environmental changes and target movement, the present invention also designs a multi-level dictionary search mechanism, which gradually improves the matching accuracy from rough to precise, and finally realizes the accurate estimation and elimination of the interference signal, ensuring the robustness and efficiency of the system in practical applications. The present invention can quickly and accurately eliminate the interfering radar signal and improve the signal detection performance of the system.

[0055] The advantages of the present invention are: quickly and accurately eliminating interference signals in the radar system, avoiding the computational complexity and precision reduction problems caused by traditional interference suppression methods, improving the efficiency of interference identification and elimination, and being suitable for scenarios with multi-variable interference signals and increasing environmental complexity, thereby enhancing the robustness and detection performance of the system;

[0056] Compared with traditional feedback methods, compressed sensing methods and deep learning-based methods, this method avoids the problems of additional feedback overhead and high computational complexity, and can achieve interference suppression quickly and accurately without increasing the hardware burden. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 This is a radar signal processing flow chart of the present invention.

[0058] Figure 2 This is a diagram showing the interference elimination effect of the present invention.

[0059] Figure 3 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0060] The present invention will be further described below in conjunction with the accompanying drawings.

[0061] Reference Figure 1 – Figure 3 , a frequency modulated continuous wave radar interference elimination method based on adaptive dictionary matching, the specific steps are:

[0062] Step 1: Establish the mathematical model of the frequency modulated continuous wave radar sending signal, receiving signal and interference signal, including the following process:

[0063] 1.1 The signal sent by the frequency modulated continuous wave radar is:

[0064]

[0065] Among them, f c To send a carrier signal; is the modulation slope, B is the bandwidth, and T is the chirp storage time;

[0066] 1.2 Define the target echo signal as

[0067]

[0068] Among them, τ r and P r are the time delay and received power of the rth target echo signal respectively;

[0069] 1.3 Interference signal is defined as

[0070]

[0071] Among them, τ i and P i are the time delay and received power of the i-th interference signal respectively;

[0072] Step 2: Construct a high-precision dictionary based on the above signal model, including the following process:

[0073] 2.1 Define the column vector in the dictionary as:

[0074] a τk =[a τk (t t1 ),a τk (t t2 ),…,a τk (t tN )] T ∈R N*1 (4)

[0075] Where τ is the time delay converted within the radar working distance. The accuracy of this distance affects the algorithm complexity and accuracy. It is recommended to use an interval of 0.01 meters. k is the slope of common radar models. t t1 ,t t2 ,…t tN is the sampling time;

[0076]

[0077] 2.2 Construct an adaptive dictionary covering radar operating parameters:

[0078] A=[a 11 …a 1k …a τ1 …a τk ]∈R N*(τ*k) (5)

[0079] Where a, τ and k are defined in 2.1;

[0080] Step 3: Sampling the received signal and calculating the power of the received signal, including the following process:

[0081] 3.1 The received signal is defined as:

[0082] y rx (t) = s R (t)+s I (t) (6)

[0083] 3.2 Sampling the received signal:

[0084] y rx =[y rx (t t1 ),y rx (t t1 ),…,y rx (t tN )] (7)

[0085] Among them, t t1 ,t t2 ,…,t tN They are the samples at time t1, t2, …, tN respectively;

[0086] 3.3 Calculate the received power:

[0087]

[0088] Step 4: When the received power is greater than the given threshold γ1, proceed to step 5, otherwise proceed to step 10;

[0089] Step 5: Sparsely and discretely select column vectors from the dictionary constructed in step 2. It is recommended to use the delay converted from the speed of light corresponding to a distance of 1 meter as the interval, including the following process:

[0090] 5.1 Uniformly sparsely select column vectors from the dictionary:

[0091]

[0092] Among them, τ1, τ2, …, τ m are sparse delay data respectively. It is recommended to use the delay converted from the speed of light corresponding to a distance of 1 meter as the interval, k1, k2, …, k n are the radar slopes for sparse selection, respectively;

[0093] 5.2 Calculate the correlation as:

[0094]

[0095] Step 6: Analyze the correlation distribution between the received signal and the dictionary column vector, calculate the mean and standard deviation of the correlation value in the absence of a target, set the threshold γ2 = μ2 + k2σ2, k2 can be set to a value between 1 and 3 as needed, when the correlation is greater than the given threshold γ2, go to step 7, otherwise go to step 10;

[0096] Step 7: When the distance interval of the selected dictionary vector delay conversion is less than or equal to 1 cm, proceed to step 9, otherwise proceed to step 8;

[0097] Step 8: Select a denser dictionary around the delay and slope parameters whose correlation is greater than a given threshold γ2, including the following process:

[0098] 8.1 Reselect the dictionary column vector to become a denser dictionary:

[0099]

[0100] where a τmkn It can be expressed as a column vector determined by the parameters calculated in step 5 that are greater than the given threshold γ2. The dense dictionary suggests that the dense dictionary in the previous cycle Iterate at intervals;

[0101] 8.2 Go to step 3.3;

[0102] Step 9: Select the delay and slope corresponding to the maximum correlation result, construct the interference signal, and eliminate it from the received signal, including the following process:

[0103] 9.1 Construct the interference signal as:

[0104]

[0105] in, and The interference delay and slope estimated in step 8, respectively, is the interference power determined according to the time delay, which is specifically calculated as follows;

[0106]

[0107] P t ,G t ,G t and λ are the transmit power, transmit antenna gain, receive antenna gain and signal wavelength respectively, and R is the distance of delay conversion;

[0108] 9.2 Subtract the estimated interference from the received signal;

[0109] Step 10: After the above steps, the interference is eliminated, and conventional radar signal processing is performed to calculate the distance and speed of the target.

[0110] Aiming at the problem of interference signal suppression in radar systems, the present invention proposes an interference elimination method based on interference features based on a dictionary matching algorithm, which avoids the problems of high computational complexity and reduced accuracy in traditional algorithms and achieves rapid and accurate identification and elimination of interference signals.

[0111] In order to verify the effectiveness of the proposed method, the present invention simulates the results of radar signal restoration after interference suppression. In the simulation, the specific parameters are set as follows:

[0112] Specific setting of delay τ: According to the radar working range and target distance, the delay range is set to 0 to 60 meters, and the corresponding delay τ is calculated as: Among them, R is the target distance, c is the speed of light, and the sparse selection of delay is selected every 1 meter interval to cover the entire delay range;

[0113] Specific setting of threshold γ1: Under given simulation conditions, by statistically analyzing the power distribution of received signals without interference, the mean is calculated to be 0.5 and the standard deviation is 0.1. Therefore, the threshold γ1 is set to 0.5+5*0.1=1. This setting can effectively distinguish normal signals from interference signals and reduce false alarms.

[0114] Specific setting of threshold γ2: By analyzing the correlation distribution in different scenarios, it is determined that the mean of the correlation in the case of no target is 0.3 and the standard deviation is 0.05. Therefore, the threshold γ2=0.3+1.5*0.05=0.375 is set. This threshold ensures that the interference signal can still be effectively detected under the condition of high signal-to-noise ratio;

[0115] The results are as follows Figure 2As shown, when the radar echo signal is interfered, the target signal is drowned by the interference signal and noise, and the target signal cannot be accurately identified. However, after the anti-interference processing of the method of the present invention, the target signal is accurately identified. Therefore, the present invention provides a frequency modulation millimeter wave radar anti-interference algorithm with high feasibility and high accuracy, which overcomes the problems of poor real-time performance and other problems associated with compressed sensing and other methods, and realizes accurate radar judgment of targets.

[0116] The above describes the excellent optimization effect shown by an embodiment of the present invention. Obviously, the present invention is not limited to the above embodiment, and various modifications can be made to it without departing from the basic spirit of the present invention and without exceeding the scope of the essential content of the present invention.

Claims

1. A frequency modulated continuous wave radar interference elimination method based on adaptive dictionary matching, characterized in that a two-dimensional dictionary containing multiple parameter combinations is constructed, and the interference parameters are estimated by using the matching degree between the received signal and the signal in the dictionary, so as to achieve accurate detection and elimination of the interference signal; the specific steps are: Step 1: Establish the mathematical model of the frequency modulated continuous wave radar sending signal, receiving signal and interference signal, specifically: 1.1, the signal sent by the frequency modulated continuous wave radar is: in, f c To send a carrier signal; is the modulation slope, B is the bandwidth, and T is the chirp storage time; 1.2, define the received signal as: Among them, τ r and P r are the time delay and received power of the rth target echo signal respectively; 1.3, define the interference signal as: Among them, τ i , k i and P i are the time delay, modulation slope and received power of the i-th interference signal respectively; Step 2: Construct a high-precision dictionary based on the above signal model. The specific process is as follows: 2.1, define the column vector in the dictionary as: a τk =[a τk (t t1 ),a τk (t t2 ),…,a τk (t tN )] T ∈R N*1 (4) in, τ is the time delay of the target position corresponding to the distance converted from the speed of light within the radar working range, k is the slope of the radar model, t t1 ,t t2 ,…,t tN is the sampling time, N is the number of samples for each algorithm execution; each element is defined as: 2.2, construct an adaptive dictionary covering radar operating parameters, specifically: A=[a 11 … a 1k … a τ1 … a τk ]∈R N*(τ*k) (5) Where a, τ and k are defined in 2.1; Step 3: Sample the received signal and calculate the power of the received signal. The specific process is as follows: 3.1, define the received signal as: y rx (t)=s R (t)+s I (t) (6) 3.2, sampling the received signal: and rx =[and rx (t t1 ),and rx (t t1 ),…,and rx (t tN )] (7) in, t t1 ,t t2 ,…,t tN They are the samples at time t1, t2, …, tN respectively; 3.3, calculate the received power: Step 4: When the received power is greater than the given threshold γ1, proceed to step 5, otherwise proceed to step 10, where γ1 = μ1 + b1σ1, μ1 is the mean value of the noise calculated from the received signal power measured without a target, σ1 is the standard deviation of the noise, and b1 is a multiple set according to actual needs, and the value of b1 is 3-6; Step 5: Sparsely and discretely select column vectors from the dictionary constructed in step 2, with the delay corresponding to a distance of 1 meter converted from the speed of light as the interval. The specific process is as follows: 5.1, uniformly sparsely select column vectors from the dictionary: in, τ1,τ2,…,τ m are sparse delay data, with the delay corresponding to a distance of 1 meter converted from the speed of light as the interval, k1, k2, …, k n are the sparsely selected radar slopes, respectively; 5.2, calculate the correlation as: Step 6: Analyze the correlation distribution between the received signal and the dictionary column vector, calculate the mean and standard deviation of the correlation value in the case of no target, set the threshold γ2 = μ2 + b2σ2, μ2 is the mean of the correlation value calculated by the received signal power measured in the case of no target, σ2 is the standard deviation, b2 is set to a value between 1 and 3 as needed, when the correlation is greater than the given threshold γ2, go to step 7, otherwise go to step 10; Step 7: When the distance interval of the selected dictionary vector delay conversion is less than or equal to 1 cm, proceed to step 9, otherwise proceed to step 8; Step 8: Select a denser dictionary around the delay and slope parameters whose correlation is greater than a given threshold γ2. The specific process is: 8.1, reselect the dictionary column vector to become a denser dictionary: in, It can be expressed as a column vector determined by the parameters calculated in step 5 that are greater than the given threshold γ2. The dense dictionary is the dense dictionary in the previous cycle. Iterate at intervals; 8.2, go to step 3.3; Step 9: Select the delay and slope corresponding to the maximum correlation result, construct the interference signal, and eliminate it from the received signal. The specific process is as follows: 9.1, construct the interference signal as: in, and are the interference delay and slope estimated in step 8, respectively. is the interference power determined according to the time delay, which is specifically calculated as: P t ,G t ,G t and λ are the transmit power, transmit antenna gain, receive antenna gain and signal wavelength respectively, and R is the distance of delay conversion; 9.2, subtracting the estimated interference from the received signal; Step 10: After the above steps, the interference is eliminated and conventional radar signal processing is performed to calculate the distance and speed of the target.

Images