A method for detecting deceptive signals based on adaptive immune algorithm

Abstract

The disclosed embodiment is about a deception signal detection method based on an adaptive immune algorithm. The method includes: using an immune algorithm to obtain a real monitor and a deception monitor in a real signal detection mode and a deception signal detection mode respectively; wherein the adaptive algorithm performs nonlinear adaptive adjustment on the crossover probability and the mutation probability in the immune algorithm; calculating the data to be detected with each real monitor and each deception monitor respectively, and obtaining the real signal affinity index of the data to be detected with each real monitor and the deception signal affinity index of each deception monitor respectively; according to the real signal affinity index and the deception signal affinity index, calculating the probability that the data to be detected belongs to the real signal detection mode and the probability of the deception signal detection mode. The disclosed method utilizes the difference in the time domain changes of the carrier phase double difference observation values ​​of the real signal and the deception signal, and uses an adaptive immune algorithm to realize the deception detection of satellite navigation signals.

Description

Technical Field

[0001] The disclosed embodiments relate to the technical field of satellite navigation deception detection, and in particular to a deception signal detection method based on an adaptive immune algorithm. Background Art

[0002] In recent years, the BeiDou Navigation Satellite System (BDS) has been widely used in civil and military fields. However, due to the fragility of the Global Navigation Satellite System (GNSS) signal itself, it is extremely susceptible to artificial deception interference during the reception process. Common deception interference methods include generation deception and forwarding deception. Among them, forwarding deception interference has the characteristics of strong concealment and easy implementation, which seriously threatens the security of navigation and positioning. Therefore, it is necessary to study effective deception detection methods based on the characteristics of forwarding deception interference. Most of the existing forwarding deception interference technologies transmit multiple deception signals to the outside world through a single antenna, so the directions of the incoming waves received by the receiver are the same; while the real satellite signals have different signal directions when they reach the receiver. In view of this feature, the carrier phase difference containing the arrival angle information can be used to detect deception signals.

[0003] Therefore, it is necessary to improve one or more problems existing in the above-mentioned related technical solutions.

[0004] It should be noted that this section is intended to provide background or context for the technical solutions of the present disclosure stated in the claims. The description herein is not admitted to be prior art by virtue of being included in this section. Summary of the invention

[0005] The purpose of the embodiments of the present disclosure is to provide a method for detecting spoofing signals based on an adaptive immune algorithm, thereby overcoming one or more problems caused by limitations and defects of related technologies at least to a certain extent.

[0006] According to an embodiment of the present disclosure, a method for detecting spoofing signals based on an adaptive immune algorithm is provided, the method comprising:

[0007] Using an immune algorithm to obtain real monitors and deception monitors in a real signal detection mode and a deception signal detection mode, respectively, wherein the number of the real monitors and the number of the deception monitors are both preset values;

[0008] Wherein, an adaptive algorithm is used to perform nonlinear adaptive adjustment on the crossover probability and the mutation probability in the immune algorithm;

[0009] Calculate the data to be detected with each of the real monitors and each of the deception monitors to obtain the affinity index of the data to be detected with the real signal of each of the real monitors and the affinity index of the deception signal of each of the deception monitors;

[0010] According to the real signal affinity index and the spoof signal affinity index, the probability that the data to be tested belongs to the real signal detection mode and the probability that the spoof signal detection mode are calculated.

[0011] In one embodiment of the present disclosure, the immune algorithm includes:

[0012] Setting a fault mode, and setting the number of monitors generated in the real signal detection mode and the spoof signal detection mode to a preset value;

[0013] The carrier phase double difference observation values ​​of the real signal and the spoofing signal are used as antigen inputs, and selection, crossover, and mutation evolution are performed in the real signal detection mode and the spoofing signal detection mode respectively to obtain a preset value of initial real monitors and a preset value of initial spoofing monitors;

[0014] The fitness values ​​of each of the initial true monitors and each of the initial deception monitors are calculated respectively, and the maximum individual of the true fitness value and its true fitness value, the maximum individual of the deception fitness value and its deception fitness value among all the initial true monitors, all the initial deception monitors are obtained respectively.

[0015] In one embodiment of the present disclosure, the immune algorithm further includes:

[0016] When the true fitness value is less than the required fitness threshold, the evolution is stopped and the true monitor is obtained;

[0017] When the deception fitness value is less than a required fitness threshold, the evolution is stopped and the deception detector is obtained.

[0018] In one embodiment of the present disclosure, the immune algorithm further includes:

[0019] When the true fitness value is greater than or equal to the fitness threshold, the individual with the largest true fitness value is used as an antigen for evolution, and a new initial true monitor is generated;

[0020] When the deceptive fitness value is greater than or equal to the fitness threshold, the individual with the largest deceptive fitness value is used as an antigen for evolution, and a new initial deceptive monitor is generated.

[0021] In one embodiment of the present disclosure, in the fault mode, the real signal and the spoof signal are further stimulated and clustered for learning, thereby gradually maturing the generated initial real monitor and the initial spoof monitor.

[0022] In one embodiment of the present disclosure, the crossover probability and the mutation probability are adaptively adjusted to improve the operation speed and the accuracy of the detection result.

[0023] In one embodiment of the present disclosure, obtaining a preset value of real signal affinity indexes and a preset value of spoof signal affinity indexes includes:

[0024] The Euclidean distances between the data to be detected and the real monitor and the deception monitor are calculated respectively, and the real signal affinity index and the deception signal affinity index are obtained.

[0025] In one embodiment of the present disclosure, the calculating the probability that the data to be tested belongs to the real signal detection mode and the spoof signal detection mode includes:

[0026] Comparing each of the real signal affinity indexes and each of the spoof signal affinity indexes with a similarity threshold respectively;

[0027] The number of real signals with affinity index less than the similarity threshold obtained by statistics is a first number, and the number of spoofed signals with affinity index less than the similarity threshold is a second number;

[0028] According to the first number and the second number, the probability that the data to be tested belongs to the real signal detection mode and the probability that the spoof signal detection mode are respectively calculated.

[0029] In one embodiment of the present disclosure, the method further includes:

[0030] The real signal detection mode and the spoof signal detection mode are constructed by utilizing the difference in the time domain variation of the carrier phase double difference observation values ​​of the real signal and the spoof signal.

[0031] In one embodiment of the present disclosure, ignoring noise, the carrier phase double difference observation value of the real signal is affected by satellite motion and will produce significant changes in the time domain, while the carrier phase observation double difference value of the spoof signal is 0.

[0032] The technical solution provided by the embodiments of the present disclosure may have the following beneficial effects:

[0033] In the embodiments of the present disclosure, through the above-mentioned spoofing signal detection method based on the adaptive immune algorithm, on the one hand, the difference in the time domain change of the carrier phase double difference observation values ​​of the real signal and the spoofing signal is used to implement the spoofing detection of the satellite navigation signal using the immune algorithm. On the other hand, in view of the problem that the time domain change of the carrier phase double difference observation values ​​of some real signals is not significant in the actual detection process, the adaptive algorithm is introduced for optimization, thereby accelerating the population convergence speed and improving the detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification are used to explain the principles of the present disclosure. Obviously, the accompanying drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0035] Figure 1 A step diagram of a fraud signal detection method based on an adaptive immune algorithm disclosed in the present invention is shown;

[0036] Figure 2 A schematic diagram showing the basic principle of forwarding deception interference in an exemplary embodiment of the present disclosure is shown;

[0037] Figure 3 A basic flow chart of the immune algorithm in an exemplary embodiment of the present disclosure is shown;

[0038] Figure 4 A schematic diagram of a fault detection model based on an immune algorithm in an exemplary embodiment of the present disclosure is shown;

[0039] Figure 5 The double difference observation value of the carrier phase of the real signal in the exemplary embodiment of the present disclosure is shown;

[0040] Figure 6 The double difference observation value of the carrier phase of the spoofing signal in the exemplary embodiment of the present disclosure is shown;

[0041] Figure 7 The following is a diagram showing the arrangement of an experimental scene in an exemplary embodiment of the present disclosure;

[0042] Figure 8 A diagram showing a trend of fitness changes of real signal populations in an exemplary embodiment of the present disclosure is shown;

[0043] Fig. 9 A diagram showing a fitness change trend of a spoofing signal population in an exemplary embodiment of the present disclosure is shown;

[0044] Fig.10 The double difference observation value of the carrier phase of the real signal in the actual detection in the exemplary embodiment of the present disclosure is shown;

[0045] Fig.11 A trend diagram of the fitness change of the real signal population before and after the improvement in the exemplary embodiment of the present disclosure is shown;

[0046] Fig.12 A graph showing the fitness change trend of the spoof signal population before and after improvement in an exemplary embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0047] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that the disclosure will be more comprehensive and complete and to fully convey the concepts of the example embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0048] In addition, the accompanying drawings are only schematic illustrations of the embodiments of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated descriptions will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities.

[0049] In this example implementation, a method for detecting fraudulent signals based on an adaptive immune algorithm is first provided. Figure 1 As shown in , the spoofing signal detection method based on the adaptive immune algorithm may include: steps S101 to S103.

[0050] Step S101: using an immune algorithm to obtain real monitors and deception monitors in a real signal detection mode and a deception signal detection mode respectively, wherein the number of the real monitors and the number of the deception monitors are both preset values;

[0051] Wherein, an adaptive algorithm is used to perform nonlinear adaptive adjustment on the crossover probability and the mutation probability in the immune algorithm;

[0052] Step S102: Calculating the data to be detected with each of the real monitors and each of the deception monitors to obtain the real signal affinity index of the data to be detected and each of the real monitors and the deception signal affinity index of each of the deception monitors;

[0053] Step S103: Calculating the probability that the data to be tested belongs to the real signal detection mode and the probability that the data to be tested belongs to the spoof signal detection mode according to the real signal affinity index and the spoof signal affinity index.

[0054] Through the above-mentioned spoofing signal detection method based on the adaptive immune algorithm, on the one hand, the difference in the time domain changes of the carrier phase double difference observation values ​​of the real signal and the spoofing signal is used to realize the spoofing detection of the satellite navigation signal by using the immune algorithm. On the other hand, in view of the problem that the time domain changes of the carrier phase double difference observation values ​​of some real signals are not significant in the actual detection process, the adaptive algorithm is introduced for optimization, thereby accelerating the population convergence speed and improving the detection accuracy.

[0055] Next, we will refer to Figures 1 to 4 Each step of the above-mentioned spoofing signal detection method based on the adaptive immune algorithm in this example implementation is described in more detail.

[0056] Step S101: using an immune algorithm to obtain real monitors and deception monitors in a real signal detection mode and a deception signal detection mode respectively, wherein the number of the real monitors and the number of the deception monitors are both preset values; wherein an adaptive algorithm is used to perform nonlinear adaptive adjustment on the crossover probability and the mutation probability in the immune algorithm.

[0057] Under normal circumstances, the carrier phase observation equation of receiver j for the mth satellite is:

[0058]

[0059] In the formula, λ represents the signal carrier wavelength; c is the speed of light; represents the geometric distance from the receiver antenna j to the mth satellite; represents the ionospheric and tropospheric delays at the target receiver; δt1 represents the clock error of the target receiver; δt m represents the clock error of satellite m; represents the integer ambiguity of receiver j for the mth satellite; It represents the carrier phase observation error of the signal received by antenna j.

[0060] Similarly, the carrier phase observation equation of receiver k for the mth satellite is:

[0061]

[0062] By differentiating equation (1) and equation (2), we can obtain the single difference equation of the carrier phase of the target receiver j and k for the mth satellite:

[0063]

[0064] The single-difference equation for the carrier phase of the nth satellite is:

[0065]

[0066] The above single difference operation can eliminate some spatially related errors. In order to further eliminate the receiver clock error, the difference operation is performed on equation (3) and equation (4) to obtain the carrier phase double difference observation equation:

[0067]

[0068] like Figure 2 As shown in Figure 1, when a satellite is subjected to forwarding deception interference, the carrier phase observation equation of the receiver antenna j for the mth satellite is:

[0069]

[0070] In formula (6), t sj It includes the delay from repeater s to receiver j, as well as the artificially set delay for deceptive interference.

[0071] Similarly, the single-difference observation equation of the carrier phase of the mth satellite by receiver k is:

[0072]

[0073] After double difference processing, the carrier phase double difference observation equation is obtained as follows:

[0074]

[0075] By comparing equation (5) and equation (8), it can be found that, when the noise error is ignored, the carrier phase double difference observation value of the real signal will change significantly in the time domain due to the influence of satellite motion; while the carrier phase double difference observation value of the spoofing signal is 0. Based on this characteristic, spoofing interference detection can effectively identify the signal type and thus improve the reliability of satellite navigation.

[0076] The immune algorithm is an intelligent information search algorithm introduced based on the mechanism of the human immune system. Based on the genetic algorithm, it extracts the characteristics of the problem to be solved, while ensuring that its own mechanism remains unchanged during the search process, it increases the diversity of the population and solves the degradation problem that occurs when the genetic algorithm searches for the optimal solution in multiple peaks (referring to the fact that some of the newly generated individuals will have fitness that does not meet the requirements and cannot maintain the characteristics of survival of the fittest). The process is as follows: Figure 3 shown.

[0077] The immune algorithm achieves fault detection by comparing the similarity relationship between antigens and antibodies. The specific detection process is: first, the antigen (original input data) is identified (learning memory, feature extraction process) to generate an initial fault monitor (i.e., initial antibody); the antigen further stimulates the initial antibody, so that it gradually matures after undergoing processes such as cross-mutation, and can meet the requirements of fault diagnosis. During the continuous stimulation of the antigen, the antibody records the data features of the signal under fault conditions. After the new antigen enters the detector, fault detection can be achieved by analyzing the similarity between the data features of the antigen and the data features recorded during the antibody maturation process. Figure 4 It is a fault detection model based on immune algorithm.

[0078] The model is divided into two layers: fault detection and fault diagnosis. The fault detection part corresponding to the first layer is responsible for monitoring whether the system is in an abnormal state; the fault diagnosis part corresponding to the second layer can confirm the type of fault and the location of the fault. In terms of antigen selection, the first type of antigen is used to produce initial antibodies, so it should characterize the data characteristics of the system under normal working conditions, that is, autologous input; while the second type of antigen is used to further stimulate the antibody and make it cluster and learn, and gradually mature, so the characteristic data used should be when the system is in an abnormal state, that is, non-autologous input. The mature antibody can detect the fault type and fault location, and is widely used in system troubleshooting.

[0079] Crossover probability P in immune algorithm c and mutation probability P m These are two key indicators that affect the performance and convergence of the algorithm. By observing the concentration of fitness during the population evolution process, the crossover probability P is c and mutation probability P m Perform nonlinear adaptive adjustment to achieve the purpose of optimizing algorithm performance.

[0080] The crossover probability P is given below c and mutation probability P m The adaptive adjustment expression for the value of is:

[0081]

[0082]

[0083] In the formula, is the mean of the population fitness, f max is the maximum fitness of the population, k1 and k2 are constants and 0<k1, k2≤1.

[0084] when When the average fitness of the population is significantly different from the maximum fitness value, the fitness value distribution of the population is relatively dispersed, and the samples are rich. At this time, the crossover probability P of the population is increased. c , which can make the individual genes of the population fully exchanged, evolve better individuals, and reduce the mutation probability P of the population m , accelerate the convergence of population fitness and prevent the destruction of excellent individuals.

[0085] when When , the average fitness of the population gradually approaches the maximum fitness value. At this time, the fitness distribution of the population is relatively concentrated, and the difference between individuals is very small. At this time, the exchange of genes between populations has little effect on the gene quality of individuals, but will increase the amount of calculation. Therefore, the crossover probability P of the population should be reduced. c At the same time, in order to avoid falling into the problem of local optimal solution, the mutation probability P of the population is increased m , enabling it to search globally for the optimal solution.

[0086] From the above analysis, it can be concluded that the key to using immune algorithms for deception detection lies in the setting of the monitor. The literature (Wen Zheng, Sun Huake. MATLAB Intelligent Algorithm [M]. Beijing: Tsinghua University Press, 2017: 283-290.) gives the fitness threshold H that meets the detection requirements of the monitor. max =10, on this basis, the adaptive immune algorithm is used to detect deception using the difference in the time domain and space changes of the double difference observation values ​​of the carrier phase of the real signal and the deception signal.

[0087] In one embodiment, the method further comprises: constructing the real signal detection mode and the spoof signal detection mode by utilizing the difference in time domain variation of the carrier phase double difference observation values ​​of the real signal and the spoof signal.

[0088] Specifically, most of the existing forwarding deception jamming technologies transmit multiple deception signals to the outside world through a single antenna, so the directions of the waves received by the receiver are the same; while the real satellite signals have different signal directions when they reach the receiver. In view of this characteristic, deception signal detection can be performed using the carrier phase difference containing the arrival angle information. Therefore, the real signal detection mode and the deception signal detection mode are constructed, and the carrier phase double difference observation values ​​of the real signal and the deception signal respectively change in the time domain.

[0089] In one embodiment, the carrier phase double difference observation value of the real signal is affected by the satellite motion and will produce significant changes in the time domain, while the carrier phase double difference observation value of the spoof signal is 0.

[0090] Specifically, when noise errors are ignored, the double difference observation value of the carrier phase of the real signal will change significantly in the time domain due to the influence of satellite motion; while the double difference observation value of the carrier phase of the spoofing signal is 0. Deception interference detection based on this characteristic can effectively identify the signal type, thereby improving the reliability of satellite navigation.

[0091] In one embodiment, the immune algorithm includes: setting a fault mode, and setting the number of monitors generated in the real signal detection mode and the spoof signal detection mode to preset values; using the carrier phase double difference observation values ​​of the real signal and the spoof signal as antigen inputs, performing selection, crossover, and mutation evolution in the real signal detection mode and the spoof signal detection mode, respectively, to obtain a preset value of initial real monitors and a preset value of initial spoof monitors; calculating the fitness value of each of the initial real monitors and each of the initial spoof monitors, respectively, to obtain the maximum individual of the real fitness value of all the initial spoof monitors and its real fitness value, and the maximum individual of the spoof fitness value and its spoof fitness value.

[0092] Specifically, the first step is to use the deception jamming platform to perform soft simulation to obtain the double difference observation values ​​of the carrier phase of the real signal and the deception signal as antigen input;

[0093] Step 2: Set the fault mode, and each mode generates a preset value monitor (antibody);

[0094] In the fault mode, the antibodies generated are the second type of antigens. The second type of antigens are used to further stimulate the antibodies and make them cluster and learn, and gradually mature. Therefore, the characteristic data used should be the abnormal state of the system, that is, non-self-input. The mature antibodies can detect the fault type and fault location, and are widely used in system troubleshooting.

[0095] The number of monitors generated in the real signal detection mode and the spoof signal detection mode reaches a preset value. Assume that the number of monitors generated in each mode is X. For example, each mode generates X=150 monitors. No specific limitation is made here.

[0096] Step 3: Decode the encoding of each mode after performing immune operations such as selection, crossover, and mutation;

[0097] Selection, crossover, and mutation evolution are performed in the real signal detection mode and the deception signal detection mode respectively to obtain a preset value of initial real monitors and a preset value of initial deception monitors. Assuming that the number of initial real monitors and initial deception monitors are X1 and X2 respectively, and X1=150 and X2=150, X1=150 initial real monitors and X2=150 initial deception monitors are obtained.

[0098] Among them, in order to address the problem that some real signal carrier phase double difference observations do not change significantly in the time domain during the actual detection process, the crossover probability and the mutation probability are adaptively adjusted, thereby accelerating the population convergence speed and improving the detection accuracy, so as to achieve the purpose of optimizing the algorithm performance.

[0099] Step 4: Find the individual with the largest fitness value in the group and its fitness value.

[0100] The fitness values ​​of each initial true monitor and each initial deception monitor are calculated respectively, and the individuals with the maximum fitness values ​​among X1=150 initial true monitors and X2=150 initial deception monitors and their fitness values ​​are obtained respectively.

[0101] In one embodiment of the present disclosure, the immune algorithm also includes: when the true fitness value is less than the required fitness threshold, stopping the evolution and obtaining the true monitor; when the deception fitness value is less than the required fitness threshold, stopping the evolution and obtaining the deception monitor.

[0102] Specifically, it is known that the fitness threshold H that meets the detection requirements of the monitor max =10, on this basis, the adaptive immune algorithm is used to detect deception by using the difference in the time domain and space changes of the double difference observation values ​​of the carrier phase of the real signal and the deception signal. That is, when the fitness value is less than 10, the evolution is stopped and the real monitor and the deception monitor are obtained.

[0103] In one embodiment of the present disclosure, the immune algorithm also includes: when the true fitness value is greater than or equal to the fitness threshold, the individual with the largest true fitness value is evolved as an antigen, and a new initial true monitor is generated; when the deception fitness value is greater than or equal to the fitness threshold, the individual with the largest deceived fitness value is evolved as an antigen, and a new initial deception monitor is generated.

[0104] Specifically, when the true fitness value is greater than or equal to 10, the individual with the largest true fitness value is evolved as an antigen, and a new initial true monitor is generated; when the deception fitness value is greater than or equal to 10, the individual with the largest deceived fitness value is evolved as an antigen, and a new initial deception monitor is generated.

[0105] Step S102: Calculate the data to be detected respectively with each of the real monitors and each of the deception monitors to obtain the real signal affinity index of the data to be detected and each of the real monitors and the deception signal affinity index of each of the deception monitors.

[0106] For encoding in the real number domain, the affinity index aff can be obtained by calculating the Euclidean distance between antibodies:

[0107]

[0108] Where L is the total dimension of antibody encoding, and The antibody η and antibody γ are Dimensional encoding.

[0109] Step S103: Calculating the probability that the data to be tested belongs to the real signal detection mode and the probability that the data to be tested belongs to the spoof signal detection mode according to the real signal affinity index and the spoof signal affinity index.

[0110] The similarity between antibodies can be expressed as:

[0111]

[0112] In this paper, we choose δ r =30, the algorithm has better global search capability and better convergence effect.

[0113] In one embodiment, the calculation of the probability that the data to be tested belongs to the real signal detection mode and the spoof signal detection mode includes: comparing each of the real signal affinity indexes and each of the spoof signal affinity indexes with a similarity threshold, respectively; obtaining statistically the number of real signal affinity indexes less than the similarity threshold as a first number, and the number of spoof signal affinity indexes less than the similarity threshold as a second number; and calculating, based on the first number and the second number, the probability that the data to be tested belongs to the real signal detection mode and the probability that the spoof signal detection mode are respectively obtained.

[0114] Specifically, each of the true signal affinity indexes is compared with the similarity threshold. If it is less than the similarity threshold, it is accumulated once, and the cumulative number of all the true signal affinity indexes that are less than the similarity threshold is N1; similarly, the cumulative number of all the true signal affinity indexes that are less than the similarity threshold is N2; then the probability that the detected data belongs to this fault type is N / X*100%.

[0115] Through the above-mentioned spoofing signal detection method based on the adaptive immune algorithm, on the one hand, the difference in the time domain changes of the carrier phase double difference observation values ​​of the real signal and the spoofing signal is used to realize the spoofing detection of the satellite navigation signal by using the immune algorithm. On the other hand, in view of the problem that the time domain changes of the carrier phase double difference observation values ​​of some real signals are not significant in the actual detection process, the adaptive algorithm is introduced for optimization, thereby accelerating the population convergence speed and improving the detection accuracy.

[0116] This embodiment is further described below in conjunction with a specific simulation example.

[0117] 1. Construction of two mode monitors

[0118] The double-difference observations of the carrier phase of the real signal and the spoofing signal have obvious differences in the time domain. By utilizing this feature, a spoofing interference platform is built to obtain the antigen data (original input data) in each mode. Then, the antigens are used to continuously train the antibodies (monitors) to make them mature, thereby constructing two mode monitors: normal signal and spoofing signal.

[0119] The connected devices are the deception source, signal source, and receiver, which are connected to the computer software platform through a combiner. The deception mode is set for soft simulation to obtain the double difference observation values ​​of the normal signal and the deception signal carrier phase.

[0120] from Figure 5 and Figure 6 It can be seen that due to the influence of noise factors in the process of data collection, the real signal shows an oscillating upward trend in the time domain, while the deceptive signal oscillates back and forth within a certain range. Multiple sets of data are taken to continuously stimulate the initial data to make it mature, thereby constructing two mode monitors for real signals and deceptive signals.

[0121] 2. Data Collection

[0122] In order to verify the effectiveness of the algorithm proposed in this paper, a forwarding deception detection experiment was conducted on the roof. The forwarding deception interference equipment includes a deception interference source, an interference source receiving antenna and a transmitting antenna. The signal acquisition equipment includes two NovAtel-OEM628 boards and two NovAtel702 receiving antennas. The collected data will be transmitted to the software processing platform in the computer for decoding. The experimental scene is as follows: Figure 7 shown.

[0123] In this experiment, the distance between the receiver antenna and the transmitting antenna of the forwarding deception interference source is 5 meters, the transmitting antenna of the forwarding deception interference source is 20 meters away from the receiving antenna, and the distance between the two receiving antennas of the receiver is 3 meters. The experiment includes two steps: (1) Turn off the forwarding deception interference source and turn on the receiving device normally. At this time, the received signal is the real Beidou satellite signal. (2) Keep the receiver working normally and turn on the forwarding deception interference source. At this time, the received signal is the deception signal.

[0124] The experiment lasted 60 minutes, with a sampling frequency of 1 Hz. The collected data was transmitted to the software processing platform for decoding to facilitate the next step of analysis.

[0125] 3. Experimental Analysis

[0126] To verify the effectiveness of this algorithm, the reference (Wen Zheng, Sun Huake. MATLAB Intelligent Algorithm [M]. Beijing: Tsinghua University Press, 2017: 283-290.) sets the number of individual chromosome segments, mutation probability and crossover probability. The simulation parameter settings are shown in Table 1:

[0127] Table 1 Simulation parameter settings

[0128]

[0129] First, the collected data is processed by carrier phase difference to obtain three groups of normal signals a, b, and c and three groups of deception signals d, e, and f. Then the processed data is used for deception detection by the aforementioned algorithm. Since the generation of each initial population and the probability of subsequent mutation and crossover are random, the final detection result should be obtained under as many populations as possible. In this experiment, each group is set to pass m = 1500 simulations and then take the average value to obtain the final probability. At the same time, the number of errors in each group of experimental detection is counted to obtain the detection error rate. The detection results of each group are shown in Table 2:

[0130] Table 2 Real signal spoofing detection results

[0131]

[0132] The final detection accuracy ω is:

[0133]

[0134] From formula (13), we can get the detection accuracy of the three groups of normal signals to be 95.13%. The fitness change trend of each group is as follows: Figure 8 shown.

[0135] By analyzing the above detection results and the changing trends of the individual fitness of each experimental population, it can be seen that when the input signal is a real signal, the immune algorithm can accurately distinguish the difference between it and the deception signal, and the final output probability is also relatively ideal. With the increase in the number of iterations, the individual fitness gradually converges to around 5. This shows that after 200 iterations, the population samples trained by the monitor can meet the purpose of deception detection, and the stability and accuracy are relatively high.

[0136] The next step is to input the deceptive signal for detection. Under the condition of ensuring the same experimental conditions as the real signal input, the detection results of each group are shown in Table 3:

[0137] Table 3 Deception signal deception detection results

[0138]

[0139] Similarly,

[0140]

[0141] From formula (14), we can get that the detection accuracy of the three groups of deception signals reached 93.02%. The fitness change trend of each group is as follows: Fig. 9 shown.

[0142] Compared with the probability results when the real signal is input, the false alarm rate increases significantly after the deception signal is input. The reason is that the forwarding deception interference is a slow-changing process. When the GNSS repeater starts working, the delayed satellite signal is transmitted to the target receiver by the repeater. At this time, the receiver has not been deceptively interfered, resulting in the real satellite signal mixed with the satellite signal received by the receiver, so the false alarm rate will increase significantly. However, as time goes by, the receiver is gradually deceived by the deception signal, and the final output probability result can also accurately detect the existence of the deception signal. At this time, the fitness change trend of the population individuals also converges to around 5, indicating that the population sample stability of the monitor meets the requirements of the experimental simulation.

[0143] 4. Adaptive algorithm improved deception detection method

[0144] According to the above parameters, the accuracy of real signal detection using immune algorithm reached 95.13%, and the accuracy of spoof signal detection reached 93.02%. However, in the actual spoof signal detection process, the carrier phase double difference observation value of the real signal is not strictly in accordance with the time domain. Figure 5 Significant changes in trends may also occur, such as Fig.10 The situation shown in the figure will cause false alarm. At this time, higher requirements are placed on the detection performance of the monitor.

[0145] Based on the original deception detection, this section proposes a deception detection method with an improved adaptive algorithm. c and mutation probability P m The values ​​are all fixed, which weakens the optimization function of the algorithm. For some real signals that do not change significantly in the time domain, the detection effect is not good. At this time, an adaptive algorithm is introduced to adjust the crossover probability P in real time according to the concentration and dispersion degree of the population. c and mutation probability P m By performing nonlinear adaptive adjustment, the purpose of improving detection accuracy can be achieved.

[0146] In this paper, k1=1, k2=0.5, the deception signal detection method based on the immune algorithm mentioned above is M1, and the deception signal detection method with adaptive algorithm optimization is M2. The effectiveness of the improved algorithm is verified by comparing the detection performance and population convergence speed of M1 and M2.

[0147] The experimental steps and parameters are consistent with the above. The detection results of the real signal and the spoofing signal of the algorithm before and after the improvement are obtained through simulation:

[0148] Table 4 Real signal deception detection results before and after improvement

[0149]

[0150] From formula (13), we can get that the detection accuracy before improvement is 95.13%, and the detection accuracy after introducing the adaptive algorithm is 99.28%, which is 4.15% higher than before improvement. We compare the population fitness change trends of a group of experiments before and after improvement. Fig.11 shown.

[0151] Similarly, when the deceptive signal is detected, the detection results are shown in Table 5.

[0152] Table 5 Deception signal deception detection results before and after improvement

[0153]

[0154] From formula (14), we can get that the detection accuracy before improvement is 93.02%, and the detection accuracy after introducing the adaptive algorithm is 98.8%, which is 5.78% higher than before improvement. The change trend of population fitness is as follows: Fig.12 shown.

[0155] By analyzing the deception detection results before and after the improvement and the trend of population fitness changes, it can be found that after the introduction of the adaptive algorithm, the crossover probability P c and mutation probability P m As the concentration of the population changes in real time, the convergence speed of the population is accelerated, and the fitness of the population after stabilization is improved, thereby optimizing the algorithm structure and achieving the purpose of improving detection accuracy.

[0156] In order to solve the problem of forwarding deception interference, this paper proposes a deception signal detection method based on an improved adaptive immune algorithm. c and mutation probability P m The nonlinear adaptive adjustment is performed to optimize the performance of the traditional immune algorithm. The experimental results show that the accuracy of the traditional immune algorithm in detecting real signals and deceptive signals is 95.13% and 93.02% respectively; after the introduction of the adaptive adjustment algorithm, the detection accuracy of the real signal is increased by 4.15%, and the detection accuracy of the deceptive signal is increased by 5.78%, which effectively improves the detection performance.

[0157] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine different embodiments or examples described in this specification.

[0158] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the appended claims.

Claims

1. A fraudulent signal detection method based on adaptive immune algorithm, characterized in that: include: Using an immune algorithm to obtain real monitors and deception monitors in a real signal detection mode and a deception signal detection mode, respectively, wherein the number of the real monitors and the number of the deception monitors are both preset values; Wherein, an adaptive algorithm is used to perform nonlinear adaptive adjustment on the crossover probability and the mutation probability in the immune algorithm; Calculate the data to be detected with each of the real monitors and each of the deception monitors to obtain the affinity index of the data to be detected with the real signal of each of the real monitors and the affinity index of the deception signal of each of the deception monitors; According to the real signal affinity index and the spoof signal affinity index, the probability that the data to be tested belongs to the real signal detection mode and the probability that the spoof signal detection mode are calculated.

2. The method for detecting fraudulent signals based on the adaptive immune algorithm according to claim 1, characterized in that: The immune algorithm comprises: Setting a fault mode, and setting the number of monitors generated in the real signal detection mode and the spoof signal detection mode to a preset value; The carrier phase double difference observation values ​​of the real signal and the spoofing signal are used as antigen inputs, and selection, crossover, and mutation evolution are performed in the real signal detection mode and the spoofing signal detection mode respectively to obtain a preset value of initial real monitors and a preset value of initial spoofing monitors; The fitness values ​​of each of the initial true monitors and each of the initial deception monitors are calculated respectively, and the maximum individual of the true fitness value and its true fitness value, the maximum individual of the deception fitness value and its deception fitness value among all the initial true monitors, all the initial deception monitors are obtained respectively.

3. The method for detecting fraudulent signals based on the adaptive immune algorithm according to claim 2, characterized in that: The immune algorithm also includes: When the true fitness value is less than the required fitness threshold, the evolution is stopped and the true monitor is obtained; When the deception fitness value is less than a required fitness threshold, the evolution is stopped and the deception detector is obtained.

4. The method for detecting fraudulent signals based on the adaptive immune algorithm according to claim 3 is characterized in that: The immune algorithm also includes: When the true fitness value is greater than or equal to the fitness threshold, the individual with the largest true fitness value is used as an antigen for evolution, and a new initial true monitor is generated; When the deceptive fitness value is greater than or equal to the fitness threshold, the individual with the largest deceptive fitness value is used as an antigen for evolution, and a new initial deceptive monitor is generated.

5. The method for detecting fraudulent signals based on the adaptive immune algorithm according to claim 2, characterized in that: In the failure mode, the real signal and the spoof signal are further stimulated and clustered for learning to generate the gradually mature initial real monitor and the initial spoof monitor.

6. The method for detecting fraudulent signals based on an adaptive immune algorithm according to claim 1, characterized in that: The crossover probability and the mutation probability are adaptively adjusted to improve the operation speed and the accuracy of the detection result.

7. The method for detecting fraudulent signals based on the adaptive immune algorithm according to claim 1, characterized in that: Obtaining a preset value of real signal affinity indexes and a preset value of spoof signal affinity indexes includes: The Euclidean distances between the data to be detected and the real monitor and the deception monitor are calculated respectively, and the real signal affinity index and the deception signal affinity index are obtained.

8. The method for detecting fraudulent signals based on the adaptive immune algorithm according to claim 1, characterized in that: The calculating the probability that the data to be tested belongs to the real signal detection mode and the spoof signal detection mode includes: Comparing each of the real signal affinity indexes and each of the spoof signal affinity indexes with a similarity threshold respectively; The number of real signals with affinity index less than the similarity threshold obtained by statistics is a first number, and the number of spoofed signals with affinity index less than the similarity threshold is a second number; According to the first number and the second number, the probability that the data to be tested belongs to the real signal detection mode and the probability that the spoof signal detection mode are respectively calculated.

9. The method for detecting fraudulent signals based on the adaptive immune algorithm according to claim 1, characterized in that: The method further includes: The real signal detection mode and the spoof signal detection mode are constructed by utilizing the difference in the time domain variation of the carrier phase double difference observation values ​​of the real signal and the spoof signal.

10. The method for detecting fraudulent signals based on the adaptive immune algorithm according to claim 9, characterized in that: When the noise is ignored, the double difference observation value of the carrier phase of the real signal is affected by the satellite motion and will produce significant changes in the time domain, while the double difference observation value of the carrier phase of the spoofed signal is 0.

Images