CN113970762A - Method and system for positioning multistage interference source - Google Patents

Abstract

The invention discloses a method and a system for positioning a multi-stage interference source, which are characterized in that signals sampled by all receivers are transmitted to a processing center for Fourier transformation, the average power of the signals received by each receiver is calculated, a plurality of groups of training data sets are generated by randomly generating interference source positions in different ranges, and then the training data sets are used for training an FNN neural network corresponding to each group of data; respectively inputting the power vectors of the received signals of the receiver obtained by the processing center into a plurality of FNN neural networks, and reducing the current positioning area through the output of the neural networks until the range of the current positioning area is smaller than the lower limit of the range of the sub-area; and gridding the final sub-region, calculating a cost function corresponding to each grid central point through the receiving vectors of the receivers after Fourier transformation and the grid central points, and taking the central point of the grid with the maximum cost function as an estimated value of the position of the interference source. The problem that the traditional precision is limited by parameter estimation precision and the calculation complexity of a direct positioning technology is too high is solved.

Description

A method and system for locating multi-level interference sources

technical field

The invention belongs to the technical field of target positioning, and in particular relates to a multi-level interference source positioning method and system.

Background technique

Global Navigation Satellite System (GNSS) is a general term for satellite navigation systems that provide continuous positioning, navigation and timing services (Position, Velocity and Time, PVT) for ground users through space satellite constellations, and is now widely used. It has become a key infrastructure in various fields. The importance of satellite navigation and positioning technology in national defense and military is self-evident. At the same time, it has also shown huge application prospects and broad commercial market in the civilian field. However, the GNSS system is extremely susceptible to interference. The interference caused by the signal only includes intra-satellite (multipath), intra-system (near-far problem), inter-system (such as Galileo system interfering with the global positioning system) and extra-system interference. Are naturally occurring (eg weather) or man-made. In addition, its application field and environment are increasingly complex, making GNSS vulnerable to radio frequency interference from intentional or unintentional sources, so its anti-jamming technology has become a hotspot in related fields of research and application. The main concern of this paper is the identification of the interference location in the anti-jamming technology, that is, the location of the interference source.

The positioning of the interference source for the navigation satellite signal receiver belongs to passive positioning, which usually requires the cooperation of multiple sensors or receivers. The positioned target neither sends a positioning request nor communicates with the positioning system to exchange information. Traditional two-step positioning methods usually use Received Signal Strength (RSS), Direction of Arrival (AOA), Time Different of Arrival (TDOA) and Frequency Different of Arrival (FDOA) ) measurement to estimate the position of the interference source, but its positioning accuracy is limited by the parameter estimation error, and the positioning accuracy is not ideal under low signal-to-noise ratio. Different from the traditional two-step positioning method, the direct positioning method (Direct Position Determination, DPD) uses the original signal sampled by the receiver to directly perform a grid search on the positioning range to find the grid center point that maximizes the cost function as the interference source position estimate. , without first estimating intermediate parameters, the DPD method performs better than the two-step method at low signal-to-noise ratios, but the DPD method uses grid search to find the global optimal solution is too complex.

In addition, at present, deep learning technology has become a research hotspot in the field of anti-jamming due to its powerful feature mining and data processing and analysis capabilities, and there have been studies using it for GNSS receiver interference detection and interference identification, but deep learning is used in navigation satellite systems. There are few applications in the location of outdoor interference sources.

Facing the static interference source scenario for GNSS receivers, we conduct research on the interference source positioning technology based on deep learning, and propose a multi-level interference source positioning method and system based on neural network and direct positioning method. The positioning process can be divided into Two steps: The first step is to use multiple FNN neural networks trained in advance to perform preliminary positioning to obtain an area with interference sources of a certain size; the second step is to use the DPD method to perform grid search in the preliminary positioned area.

SUMMARY OF THE INVENTION

The technical problem to be solved by the present invention is to provide a multi-level interference source positioning method and system for the above-mentioned deficiencies in the prior art, so as to solve the problem that the accuracy of the traditional two-step method is limited by the accuracy of parameter estimation and the calculation complexity of the direct positioning technology is too high. The problem.

The present invention adopts following technical scheme:

A method for locating a multi-level interference source, comprising the following steps:

S1. Model the interference source positioning problem in the satellite navigation system, use the signal power received by all receivers as the input of the neural network, and use the sub-area number where the interference source is located as the training label, and establish a plurality of different area sizes. FNN neural network and training;

S2. Divide the current positioning range into two sub-regions, input the signal power of all receivers into the FNN neural network trained in step S1, obtain the sub-region where the interference source is located, and reduce the current positioning range to the sub-region;

S3, determine the size of the current positioning range obtained in step S2, if the size of the obtained sub-region is less than the set lower limit T, the region at this time is called the final sub-region;

S4, dividing grids in the final sub-region determined in step S3, performing Fourier transform on the sampled signals of all receivers, and then calculating the objective function corresponding to each grid;

S5. Comparing all the objective functions obtained in step S4 to obtain the maximum value, using the grid center point corresponding to the maximum value of the objective function as the estimated value of the position of the interference source, and obtaining the position of the interference source according to the estimated value.

Specifically, step S1 is specifically:

S101. Model the interference source location problem in the satellite navigation system, there are L stationary receivers synchronized in frequency and time, the sampling frequencies are all f _s , and the position of the lth receiver is p _l =(x _l , y _l ), the position of the static interference source is q=(x, y), determine the mixed signal r _l (t) model received by the lth receiver, and determine that the real satellite signal is contained in the mixed signal;

S102, fix the receiver position, randomly generate the position of the interference source in different ranges, divide the current area into two sub-regions, and generate multiple sets of training data sets in different ranges, where the training data is the received signal strength of the receiver, The training label is the number of the sub-area where the interference source is located;

S103 , train a plurality of FNN neural networks corresponding to different range regions, and then select a network structure that can accurately distinguish the number of the sub-region where the interference source is located according to the training result.

Further, in step S101, the real satellite signal model is expressed as:

表示第l个接收机接收到的真实卫星信号功率，C(t)表示扩频码，D(t)表示导航电文数据，

表示真实卫星信号到第l个接收机的时延，f_c表示真实卫星信号的载波频率，f_D,l表示多普勒频移，

表示载波初相。in,

represents the real satellite signal power received by the lth receiver, C(t) represents the spreading code, D(t) represents the navigation message data,

represents the delay from the real satellite signal to the lth receiver, f _c represents the carrier frequency of the real satellite signal, f _D,l represents the Doppler frequency shift,

Indicates the initial phase of the carrier.

Further, in step S102, the initial area is gradually divided into areas of different ranges, and the sub-areas after each division are numbered i, i=1, 2, and the interference sources are randomly set multiple times in the areas of different sizes. position, and obtain the received signal power vector P and sub-region number of the receivers in one-to-one correspondence, and obtain training data sets in different range regions.

Further, in step S103, the input of the FNN neural network is the received signal strength power vector P of the L receivers, and the output is the sub-region number i, i=1, 2, and the FNN neural network includes an input layer, an output layer and a hidden layer. The number of layers, the number of hidden layer nodes and the number of layers are continuously adjusted manually during training.

Specifically, in step S3, if the size of the current positioning range obtained in step S2 is greater than the set lower limit T of the size of the sub-region, step S2 is repeated.

Specifically, in step S4, the objective function L(r; p) related to the received signal vector r and the interference source position p is:

为频域接收信号，N为接收信号采样点，b_l为未知的路径衰减系数，T为时间间隔。Among them, the maximization of B is equivalent to the maximum eigenvalue of B, L is the number of receivers, r _l is the superimposed complex signal,

is the received signal in the frequency domain, N is the sampling point of the received signal, b _l is the unknown path attenuation coefficient, and T is the time interval.

为：Specifically, in step S5, the estimated value of the interference source position

for:

Among them, the maximization of B is equivalent to the maximum eigenvalue of B, and p is the coordinate of the center point of each grid.

Another technical solution of the present invention is a multi-level interference source positioning system, comprising:

The signal processing module models the problem of positioning the interference source in the satellite navigation system. The signal power received by all receivers is used as the input of the neural network, and the sub-area number where the interference source is located is used as the training label to establish a number of different area sizes. The corresponding FNN neural network is trained;

The offline training module divides the current positioning range into two sub-regions, inputs the signal power of all receivers into the FNN neural network trained by the signal processing module, obtains the sub-region where the interference source is located, and reduces the current positioning range to the sub-region ;

The online positioning module judges the size of the current positioning range obtained by the offline training module. If the size of the obtained sub-region is smaller than the set lower limit T, the region at this time is called the final sub-region, and the sampling signals of all receivers are subjected to the Fourier transform. Lie transform, and then calculate the objective function corresponding to each grid;

The direct positioning module compares all the objective functions obtained by the online positioning module to obtain the maximum value, takes the grid center point corresponding to the maximum value of the objective function as the estimated value of the position of the interference source, and obtains the position of the interference source according to the estimated value.

Compared with the prior art, the present invention at least has the following beneficial effects:

The multi-level interference source positioning method of the present invention has a large range of the entire positioning area, and it is too complicated to directly use the direct positioning method for grid search during online positioning. Therefore, the trained neural network is used to gradually reduce the scope of the positioning area. , the direct positioning method is used in the final sub-region, which not only obtains the positioning accuracy close to the direct positioning method, but also reduces the complexity of online positioning.

Further, model the system scenario of the interference source positioning problem in the satellite navigation system, determine that the input of the neural network is the signal power received by all receivers, the training label is the number of the sub-region where the interference source is located, and establish multiple different regions. Size the corresponding FNN neural network and train it.

Further, in order to fit the reality, because the model of the real satellite signal is the background of the whole system, simulation is performed according to this model to generate training data.

Further, randomly generate interference source positions in areas of different ranges, and divide the current area into two sub-areas, and generate multiple sets of training data sets under the system model described in step S101 under different ranges, and the training data is received. The received signal strength of the machine, and the training label is the number of the sub-area where the interference source is located.

Further, the training data set and training labels generated in S102 are used to train FNN neural networks corresponding to a plurality of different range regions, and then a network structure that can accurately distinguish the number of the sub-region where the interference source is located is selected according to the training result.

Further, if the size of the current positioning range obtained in step S2 is greater than the set sub-region size lower limit T, it means that the current positioning range can be further reduced, and the reduction of the positioning range means that the complexity of the direct positioning method in the final sub-region is reduced, Therefore, the current positioning range is continuously reduced by repeating step S2, and the current positioning range has a lower limit, that is, the lower limit T of the size of the sub-region.

Further, the grid is divided in the final sub-region, and the objective function corresponding to each grid in the final sub-region is calculated, and the center of the grid point with the largest objective function is the estimated interference source position.

Further, the estimated value is the result of locating the interference source, and the mean square error between the estimated value and the actual value is used to evaluate the positioning performance of the scheme.

To sum up, for the static interference source scenario for GNSS receivers, the present invention proposes a multi-level interference source positioning method and system based on neural network and direct positioning method. The positioning process can be divided into two steps: the first step Use multiple FNN neural networks trained in advance for preliminary positioning to obtain a small area with interference sources; the second step is to use DPD method to perform grid search in the preliminary positioned area. The present invention achieves a high-precision positioning effect close to the direct positioning method, and at the same time, the computational complexity is greatly reduced compared with the direct positioning method for large-scale grid search, which is beneficial to the rapid and high-precision positioning in the online positioning stage. accomplish.

The technical solutions of the present invention will be further described in detail below through the accompanying drawings and embodiments.

Description of drawings

1 is a system scene diagram of the present invention;

Fig. 2 is the flow chart of the online positioning stage of the present invention;

3 is a schematic diagram of sub-region division of the present invention;

Fig. 4 is a schematic diagram of re-partitioning sub-regions;

Fig. 5 is the structural diagram of the FNN neural network used in the present invention;

6 is a graph showing the variation of positioning error with signal-to-noise ratio of different positioning schemes in an embodiment of the present invention;

7 is a graph showing the variation of the positioning error with the signal-to-noise ratio of different grid side lengths in the final sub-region in an embodiment of the present invention;

FIG. 8 is a graph showing the influence of the FNN neural network sub-region number determination error on the positioning accuracy of the subsequent direct positioning method in the embodiment of the present invention.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

In the description of the present invention, it is to be understood that the terms "comprising" and "comprising" indicate the presence of the described features, integers, steps, operations, elements and/or components, but do not exclude one or more other features, The existence or addition of a whole, step, operation, element, component, and/or a collection thereof.

It should also be understood that the terminology used in the present specification is for the purpose of describing particular embodiments only and is not intended to limit the present invention. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural unless the context clearly dictates otherwise.

It should further be understood that, as used in this specification and the appended claims, the term "and/or" refers to and including any and all possible combinations of one or more of the associated listed items .

Various structural schematic diagrams according to the disclosed embodiments of the present invention are shown in the accompanying drawings. The figures are not to scale, some details have been exaggerated for clarity, and some details may have been omitted. The shapes of various regions and layers shown in the figures and their relative sizes and positional relationships are only exemplary, and in practice, there may be deviations due to manufacturing tolerances or technical limitations, and those skilled in the art should Regions/layers with different shapes, sizes, relative positions can be additionally designed as desired.

The invention provides a multi-level interference source locating method. First, a neural network is used to perform preliminary positioning to obtain an area with interference sources of a certain size; The performance close to the DPD method is obtained, while the complexity of online calculation is reduced, and the positioning accuracy is also higher than the traditional two-step method. A new idea of using neural network to locate outdoor interference sources is proposed.

A method for locating a multi-level interference source of the present invention comprises the following steps:

S1. Model the interference source positioning problem in the satellite navigation system, use the signal power received by all receivers as the input of the neural network, and use the interference source position as the training label to establish multiple FNN neural networks corresponding to different area sizes;

S101. First, model the interference source location problem in the satellite navigation system. Suppose there are L stationary receivers synchronized in frequency and time, the sampling frequency is _fs , their positions are known, and the lth receiver The position of is represented by p _l =(x _l , y _l ). The location of the stationary interferer is represented by q=(x, y), and its location is known. Ignoring the data information and directivity in the navigation signal, the expression of the mixed signal r _l (t) received by the lth receiver is:

r _l (t)=S _l (t)+J _l (t)+n _l (t)

Among them, the subscript l represents the number of the receiver, the mixed signal contains the real satellite signal S _l (t), the interference signal J _l (t) and the noise n _l (t) three parts, n _l (t) is an additive Gaussian white noise with variance

The real satellite signal model is expressed as:

表示第l个接收机接收到的真实卫星信号功率，C(t)表示扩频码，D(t)表示导航电文数据，

表示真实卫星信号到第l个接收机的时延，f_c表示真实卫星信号的载波频率，f_D,l表示多普勒频移，

表示载波初相。in,

represents the real satellite signal power received by the lth receiver, C(t) represents the spreading code, D(t) represents the navigation message data,

represents the delay from the real satellite signal to the lth receiver, f _c represents the carrier frequency of the real satellite signal, f _D,l represents the Doppler frequency shift,

Indicates the initial phase of the carrier.

In practice, because the satellite is far away from the ground, the power of the interference source is much larger than the power of the real satellite signal, and therefore the influence of the real satellite signal on the positioning of the interference source can be ignored in the experiment. Set the interference source signal form, carrier frequency, number of receivers, sampling frequency, interference signal attenuation model, noise variance and other parameters according to the above model. The power of the interfering signal is shown in Figure 1.

S102, generating training data and training labels;

Referring to Figure 3, the initial area is gradually divided into areas of different sizes, and the sub-areas after each division are numbered i, i=1, 2. The received power vectors P and sub-area numbers of multiple sets of receivers corresponding to one-to-one can be obtained, and at this time, training data sets in areas of different sizes can be obtained.

S103. Train a plurality of FNN neural networks corresponding to regions of different sizes. The input of the neural network is the received power vector P of the L receivers, and the output is the sub-region number i, i=1, 2. The neural network includes an input layer and an output layer. And the hidden layer, the number of hidden layer nodes and layers are continuously adjusted manually during the training, and then the network structure that can accurately distinguish the cell number where the interference source is located is selected according to the training results.

Consider a situation where an interference source appears near the boundary of two sub-regions, as shown in the left figure of Figure 4. For this situation, we propose a scheme to re-divide the sub-regions, as shown in Figure 4. After the re-division, the interference source does not Closer to the junction of the two sub-regions, the FNN neural network can accurately distinguish the sub-region where the interference source is located.

At the same time, in the training process of the above-mentioned FNN neural network, as the size of the area gradually decreases, the lower limit T of the sub-area size is selected according to the training result, that is, the accuracy rate of the FNN neural network in distinguishing the cell number where the interference source is located. When the lower limit of the area size is T, the accuracy of the FNN neural network in distinguishing the cell number where the interference source is located drops below 99.9%.

At this point, the offline training phase is completed, and the next step is the online positioning phase.

S2. Divide the current positioning range into two sub-regions, input the signal power of all receivers into the trained FNN neural network, obtain the sub-region where the interference source is located, and reduce the positioning range to the sub-region;

Please refer to Figure 2, the position of the interference source in the online positioning stage, determine the position of the interference source according to the interference source signal sampled by the receiver, divide the current positioning range into two sub-areas, and input the interference signal power of all receivers to train well The FNN neural network corresponding to the size of the current positioning range is obtained, and the sub-region number where the interference source is located is obtained, the positioning range is reduced to the sub-region where the interference source is located, and then goes to step S3.

S3. Determine whether the size of the current positioning range obtained in step S2 is smaller than the pre-determined lower limit T of the sub-region size, otherwise step S2 is repeated. If the obtained sub-region size is smaller than the pre-determined lower limit T, the area at this time is called is the final sub-area, and step S4 is performed in the final sub-area;

S4. Divide grids in the final sub-region, perform Fourier transform on the sampled signals of all receivers, and then calculate the objective function corresponding to each grid;

Considering only the interference signal, the superimposed complex signal observed by the lth receiving station at time t is

r _l (t)=b _l s _l (t-τ _l )+w _l (t), 0≤t≤T

为传播时延。Among them, s _l (t) represents the interference signal, b _l is the unknown path attenuation coefficient,

is the propagation delay.

Sampling N points for the received signal and the time interval is T, then there are

Then the superimposed complex signal is written as

r _l =b _l s _l +w _l

The discrete Fourier transform is performed on the above formula, and the received signal expression in the frequency domain is obtained as

代表零均值、相关矩阵R_l＝σ²I的高斯白噪声向量，且各个接收站处噪声相互独立，对于只有噪声的假设H₀和目标存在的假设H₁，其概率密度函数分别为Further, suppose

Represents a white Gaussian noise vector with zero mean and correlation matrix R _l =σ ² I, and the noise at each receiving station is independent of each other. For the hypothesis H ₀ with only noise and the hypothesis H ₁ with the existence of the target, the probability density functions are respectively

Among them, K ₀ and K ₁ are constants independent of the target position.

The log-likelihood ratio can be calculated as

make

得In order to further obtain the expression form of position estimation, the position-independent attenuation coefficient β _l is solved according to the generalized likelihood ratio maximum criterion, let

have to

得到substitute

get

则有Assumption

then there are

Finally got:

In practice, the interference source is non-cooperative, and most of its transmitted signal forms are unknown. At this time, it can be made:

where w _k =2πk/(NT).

The cost function L(r; p) related to the location of the received signal and the interferer is

Calculate the cost function corresponding to all grid center points, that is, the objective function.

S5. Compare all the objective functions obtained in step S4 to obtain the maximum value thereof, and use the grid center point corresponding to the maximum value of the objective function as the estimated value of the position of the interference source.

Taking the maximization of L(r; p) as the goal, compare all the objective functions obtained in step S4 to obtain the maximum value, and use the grid center point corresponding to the maximum value of the objective function as the estimated value of the position of the interference source, that is, traverse all the The grid finds the position p where the largest eigenvalue of B is the largest, which is the estimated value of the interference source position. which is

In still another embodiment of the present invention, a multi-level interference source positioning system is provided, and the system can be used to realize the above-mentioned multi-level interference source positioning method. Specifically, the multi-level interference source positioning system includes a signal processing module and an offline training module. , online positioning module and direct positioning module.

Among them, the signal processing module transmits the signals sampled by all receivers to the processing center, performs Fourier transform, and calculates the average power of the signals received by each receiver;

The offline training module generates multiple sets of training data sets by randomly generating the positions of interference sources in different ranges, and then uses them to train the FNN neural network corresponding to each set of data;

The online positioning module inputs the received signal power vector of the receiver obtained by the processing center into multiple FNN neural networks respectively, and gradually narrows the current positioning area through the output of the neural network until the current positioning area is smaller than the lower limit of the sub-area range obtained in advance;

The direct positioning module grids the final sub-region, calculates the cost function corresponding to each grid center point through the acceptance vectors of multiple receivers after Fourier transformation and the grid center point, and calculates the grid with the largest cost function. The center point is used as an estimate of the location of the interference source.

In yet another embodiment of the present invention, a terminal device is provided, the terminal device includes a processor and a memory, the memory is used for storing a computer program, the computer program includes program instructions, and the processor is used for executing the computer Program instructions stored in the storage medium. The processor may be a central processing unit (Central Processing Unit, CPU), or other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gates Field-Programmable GateArray (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., are the computing core and control core of the terminal, and are suitable for implementing one or more instructions. Loading and executing one or more instructions to implement corresponding method processes or corresponding functions; the processor according to the embodiment of the present invention can be used for the operation of the multi-level interference source locating method, including:

Model the interference source location problem in the satellite navigation system, use the signal power received by all receivers as the input of the neural network, and use the sub-region number where the interference source is located as the training label, and establish multiple FNN neural networks corresponding to different region sizes. Network and train; divide the current positioning range into two sub-regions, input the signal power of all receivers into the trained FNN neural network, obtain the sub-region where the interference source is located, and reduce the current positioning range to the sub-region; judge The size of the current positioning range, if the size of the obtained sub-area is smaller than the set lower limit T, the area at this time is called the final sub-area; the grid is divided into the final sub-area, and the sampled signals of all receivers are subjected to Fourier analysis. Leaf transformation, and then calculate the objective function corresponding to each grid; compare all objective functions to get the maximum value, take the grid center point corresponding to the maximum value of the objective function as the estimated value of the position of the interference source, and obtain the interference source according to the estimated value Location.

In yet another embodiment of the present invention, the present invention further provides a storage medium, specifically a computer-readable storage medium (Memory), where the computer-readable storage medium is a memory device in a terminal device for storing programs and data . It can be understood that, the computer-readable storage medium here may include both a built-in storage medium in the terminal device, and certainly also an extended storage medium supported by the terminal device. The computer-readable storage medium provides storage space in which the operating system of the terminal is stored. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions may be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory, or a non-volatile memory (non-volatile memory), such as at least one disk memory.

One or more instructions stored in the computer-readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the method for locating the multi-level interference source in the above-mentioned embodiment; one or more instructions in the computer-readable storage medium are composed of The processor loads and performs the following steps:

.

In order to make the purposes, 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 with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments These are some embodiments of the present invention, but not all embodiments. The components of the embodiments of the invention generally described and illustrated in the drawings herein may be arranged and designed in a variety of different configurations. Thus, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but is merely representative of selected embodiments of the invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

Example

Consider a stationary interferer randomly appearing within the location range and apply the previously described scheme to locate this interferer. The initial positioning range is set as a square with a side length of 4km, and the four receivers are respectively arranged at the four vertices [0,0], [0,4000], [4000,0], [4000,4000] of the square area. The interference source signal is a QPSK signal, the carrier frequency is 1.25MHz, the number of sampling points is N=200, and the sampling frequency is 5MHz.

Comparison scheme

Comparison scheme 1: first estimate the TDOA parameters and then use the chan method to solve the position.

Contrast scheme 2: First perform wavelet denoising, then estimate the TDOA parameters, and finally use the chan method to solve the position.

Comparison scheme 3: DPD method is directly used for grid search-type position calculation in the entire positioning range.

According to the experimental settings, the FNN has four input nodes and one output node. The structure of the hidden layer of the neural network is determined as described in step S103. According to the results of multiple experiments, the final FNN structure is set to (4, 6, 18, 1), as shown in Figure 5.

In order to determine the lower limit T of the side length of the final sub-area, the dichotomy method is used to gradually reduce the side length of the sub-area, and the accuracy rate 1 of the FNN neural network in distinguishing the cell number where the interference source is located under the corresponding side-length sub-area is obtained. The method of re-dividing the sub-regions is used to correct the number of interference sources near the junction of the two sub-regions, and the accuracy rate is 2, as shown in Table 1. It can be seen from the table that our method is effective and the sub-region edge The lower limit of length T is set at 250m.

Table 1 The accuracy rate of FNN network in distinguishing the cell number where the interference source is located under different side length sub-regions

其中P_J为干扰的功率，P_S为噪声的功率，定位误差采用均方根误差

衡量(K为实验次数)，定位区域和接收机位置设置同上，分别仿真了大尺度衰落下先进行小波去噪再利用TDOA进行chan定位结算方案，大尺度衰落和瑞利衰落下先进行小波去噪再利用TDOA进行chan定位结算方案、直接利用DPD大范围搜索方案、最终子区域分别为200m和400m时利用本文所提方案的定位误差随信噪比变化曲线图，结果如图6示，从图中可以看出，基于TDOA的对比方案1和对比方案2性能较差且严重依赖于参数估计精度，尽管进行了小波去噪，性能还是低于使用了DPD的对比方案3以及本文所提方案。而本文所提出的方案在信噪比>10dB时得到了与直接使用DPD方法进行搜索相近的性能，且计算复杂度低于在整个定位范围内直接用DPD方法进行网格搜索式的位置解算。The signal-to-noise ratio is defined as

Among them, P _J is the power of interference, P _S is the power of noise, and the positioning error adopts the root mean square error

Measure (K is the number of experiments), the positioning area and receiver position are set as above, respectively simulate the large-scale fading and then use the TDOA to perform wavelet denoising and then use TDOA to perform the chan positioning settlement scheme, and first perform the wavelet de-noising under large-scale fading and Rayleigh fading. Noise and then use TDOA to carry out the chan positioning settlement scheme, directly use the DPD large-scale search scheme, and use the proposed scheme when the final sub-area is 200m and 400m respectively. The curve diagram of the variation of the positioning error with the signal-to-noise ratio, the results are shown in Figure 6, from It can be seen from the figure that the comparison scheme 1 and comparison scheme 2 based on TDOA have poor performance and are heavily dependent on the accuracy of parameter estimation. Despite the wavelet denoising, the performance is still lower than that of the comparison scheme 3 using DPD and the scheme proposed in this paper. . However, when the signal-to-noise ratio is greater than 10dB, the scheme proposed in this paper obtains similar performance to the direct use of the DPD method for search, and the computational complexity is lower than that of the grid search-style position solution directly using the DPD method in the entire positioning range. .

In order to explore the relationship between the DPD search performance and the grid side length in the final sub-region, the grid side lengths were set to 40m, 20m, 10m, and 5m, respectively. In order to be divisible, the final sub-region size was set to 400m. The simulation results are shown in Figure 7. , it can be seen from the figure that the smaller the grid side length is, the higher the positioning accuracy is, but at the same time the computational complexity is also higher, which is in line with our expectations. Select the appropriate grid size for the specific situation.

The influence of the cell number determination error in the first stage on the positioning error accuracy of the grid search in the final cell by the direct positioning method in the subsequent stage is also considered, and the final cells are simulated respectively, as shown in Figure 8. , it can be seen from the figure that the error caused by the error in the determination of the cell number is about half of the side length of the final cell, which indicates that the determination of the cell number in the first stage has a great influence on the positioning accuracy, so the neural network The training of , and the selection of the lower limit T of the side length of the subregion are very important.

Finally, the time complexity of the method proposed in this paper and the traditional DPD method is analyzed. The total online positioning time = online positioning time + DPD search time in online sub-regions ≈ DPD search time in online sub-regions. Suppose there are L receivers in total, the side length of the positioning range is a, the side length of the grid is b, and there are G=(a/b) 2 grids in the whole range. The time of the eigenvalue is T1, the time to find the largest eigenvalue of the two eigenvalues is T2, and finally n smallest sub-regions are divided. The computational complexity of our proposed method is ((G/n)*T1+(G/n-1)*T2)/(G*T1+(G-1)*T2)≈1/n of the traditional DPD method. In this experiment n is between 256-400, so the proposed method effectively reduces the computational complexity of large-area DPD searches.

To sum up, the present invention provides a method and system for locating a multi-level interference source. In the first step, a neural network is used to perform preliminary positioning to obtain an area with interference sources of a certain size; in the second step, DPD is used in the preliminary positioning area. method to perform a grid search. The simulation results show that the present invention obtains the performance close to the DPD method while reducing the complexity of online calculation, and the positioning accuracy is also higher than that of the traditional two-step method.

As will be appreciated by those skilled in the art, the embodiments of the present application may be provided as a method, a system, or a computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It will be understood that each flow and/or block in the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to the processor of a general purpose computer, special purpose computer, embedded processor or other programmable data processing device to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing device produce Means for implementing the functions specified in a flow or flow of a flowchart and/or a block or blocks of a block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory result in an article of manufacture comprising instruction means, the instructions The apparatus implements the functions specified in the flow or flow of the flowcharts and/or the block or blocks of the block diagrams.

These computer program instructions can also be loaded on a computer or other programmable data processing device to cause a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process such that The instructions provide steps for implementing the functions specified in the flow or blocks of the flowcharts and/or the block or blocks of the block diagrams.

The above content is only to illustrate the technical idea of the present invention, and cannot limit the protection scope of the present invention. Any changes made on the basis of the technical solution according to the technical idea proposed by the present invention all fall within the scope of the claims of the present invention. within the scope of protection.

Claims (9)

1. a multi-level interference source positioning method, is characterized in that, comprises the following steps: S1. Model the interference source positioning problem in the satellite navigation system, use the signal power received by all receivers as the input of the neural network, and use the sub-area number where the interference source is located as the training label, and establish a plurality of different area sizes. FNN neural network and training; S2. Divide the current positioning range into two sub-regions, input the signal power of all receivers into the FNN neural network trained in step S1, obtain the sub-region where the interference source is located, and reduce the current positioning range to the sub-region; S3, determine the size of the current positioning range obtained in step S2, if the size of the obtained sub-region is less than the set lower limit T, the region at this time is called the final sub-region; S4, dividing grids in the final sub-region determined in step S3, performing Fourier transform on the sampled signals of all receivers, and then calculating the objective function corresponding to each grid; S5. Comparing all the objective functions obtained in step S4 to obtain the maximum value, using the grid center point corresponding to the maximum value of the objective function as the estimated value of the position of the interference source, and obtaining the position of the interference source according to the estimated value. 2. method according to claim 1, is characterized in that, step S1 is specifically: S101. Model the interference source location problem in the satellite navigation system, there are L stationary receivers synchronized in frequency and time, the sampling frequencies are all f _s , and the position of the lth receiver is p _l =(x _l , y _l ), the position of the static interference source is q=(x, y), determine the mixed signal r _l (t) model received by the lth receiver, and determine that the real satellite signal is contained in the mixed signal; S102, fix the receiver position, randomly generate the position of the interference source in different ranges, divide the current area into two sub-regions, and generate multiple sets of training data sets in different ranges, where the training data is the received signal strength of the receiver, The training label is the number of the sub-area where the interference source is located; S103 , train a plurality of FNN neural networks corresponding to different range regions, and then select a network structure that can accurately distinguish the number of the sub-region where the interference source is located according to the training result. 3. The method according to claim 2, wherein in step S101, the real satellite signal model is expressed as:

Among them, P _l ^s represents the real satellite signal power received by the lth receiver, C(t) represents the spreading code, D(t) represents the navigation message data,

represents the delay from the real satellite signal to the lth receiver, f _c represents the carrier frequency of the real satellite signal, f _D,l represents the Doppler frequency shift,

Indicates the initial phase of the carrier. 4. The method according to claim 2, characterized in that, in step S102, the initial area is gradually divided into areas of different ranges, and the sub-areas after each division are numbered i, i=1, 2, by The positions of the interference sources are randomly set for many times in areas of different sizes, and the received signal power vectors P and sub-area numbers of the receivers corresponding to one-to-one groups are obtained, and the training data sets in different areas are obtained. 5. The method according to claim 2, wherein in step S103, the input of the FNN neural network is the received signal strength power vector P of L receivers, and the output is the sub-region number i, i=1, 2, FNN neural network includes input layer, output layer and hidden layer, and the number of hidden layer nodes and layers is continuously adjusted manually during training. 6 . The method according to claim 1 , wherein, in step S3 , if the size of the current positioning range obtained in step S2 is greater than the set lower limit T of the size of the sub-region, step S2 is repeated. 7 . 7. The method according to claim 1, wherein in step S4, the objective function L(r; p) related to the received signal vector r and the interference source position p is:

Among them, the maximization of B is equivalent to the maximum eigenvalue of B, L is the number of receivers, r _l is the superimposed complex signal,

is the received signal in the frequency domain, N is the sampling point of the received signal, b _l is the unknown path attenuation coefficient, and T is the time interval. 8. The method according to claim 1, wherein in step S5, the estimated value of the interference source position

for:

Among them, the maximization of B is equivalent to the maximum eigenvalue of B, and p is the coordinate of the center point of each grid. 9. A multi-level interference source positioning system is characterized in that, comprising: The signal processing module models the problem of positioning the interference source in the satellite navigation system. The signal power received by all receivers is used as the input of the neural network, and the sub-area number where the interference source is located is used as the training label to establish a number of different area sizes. The corresponding FNN neural network is trained; The offline training module divides the current positioning range into two sub-regions, and inputs the signal power of all receivers into the FNN neural network trained by the signal processing module to obtain the sub-region where the interference source is located, and reduces the current positioning range to the sub-region. ; The online positioning module judges the size of the current positioning range obtained by the offline training module. If the size of the obtained sub-region is smaller than the set lower limit T, the region at this time is called the final sub-region, and the sampling signals of all receivers are subjected to four-dimensional analysis. Lie transform, and then calculate the objective function corresponding to each grid; The direct positioning module compares all the objective functions obtained by the online positioning module to obtain the maximum value, takes the grid center point corresponding to the maximum value of the objective function as the estimated value of the position of the interference source, and obtains the position of the interference source according to the estimated value.

Images