Electromagnetic Compatibility Prediction Method and System for Navigation Receiver and Aircraft Equipment

Abstract

The present disclosure provides a method and system for electromagnetic compatibility prediction of a navigation receiver and an airborne device. The method includes: constructing a training data set, where the training data set includes characteristic parameters and path loss obtained through experiments, training a path loss prediction model using the training data set, and optimizing the weight threshold of the path loss prediction model through an intelligent algorithm during the training process to obtain a trained path loss prediction model; collecting the transmission signals of the airborne device in real time, and extracting target characteristic parameters based on the transmission signals; inputting the target characteristic parameters into the trained path loss prediction model to obtain a target path loss; and inputting the target path loss into an electromagnetic compatibility prediction model to determine whether the navigation receiver and the airborne device are electromagnetic compatible. The method according to the present disclosure solves the problem in the prior art that the prediction effect of path loss is not good, resulting in inaccurate prediction of electromagnetic compatibility.

Description

Technical Field

[0001] The present disclosure relates to the field of electromagnetic compatibility performance prediction, and particularly to a method and system for predicting the electromagnetic compatibility between a navigation receiver and airborne equipment. Background Art

[0002] Electromagnetic compatibility refers to the ability of electronic devices and components to operate properly in the presence of other devices that emit electromagnetic waves. This means that each device that emits electromagnetic waves must limit its emission power to a certain level, and each individual device must have a certain immunity to electromagnetic interference in the environment in which it is to operate.

[0003] The navigation system is one of the most important components in an aircraft. It provides accurate position, speed, and attitude information for the aircraft, which is of great significance for aviation safety. Navigation satellites are located more than twenty thousand kilometers above the ground. By the time the satellite signals reach the Earth, they have become very weak. In the complex electromagnetic environment of an aircraft, the space signals of navigation satellites are greatly affected by the signals of various types of airborne equipment. There are many types of airborne equipment, diverse signal systems, high signal power, and complex spatio-temporal distributions. Currently, the electromagnetic compatibility analysis methods for GNSS navigation receivers used in aircraft mostly rely on prior knowledge, and judge compatibility by whether the performance of the GNSS navigation receiver under interference signals drops to the allowable range. Domestic researchers directly use the interference power received by the receiver when there is interference when simulating the performance of GNSS navigation receivers. If the performance of the receiver is actually evaluated, it is necessary to establish a system model to measure information such as the interference power received by the GNSS navigation receiver, which is costly, time-consuming, and inflexible. Whenever the system changes, it is necessary to re-establish the system model and measure. This method is difficult to apply under the requirements of new signal systems and unmanned aerial vehicle platforms.

[0004] In addition, Zhang Wenwen et al. used the Monte Carlo model to predict the electromagnetic compatibility performance between systems (Zhang Wenwen, Yang Shiming, Li Weiming, Gao Benqing, Prediction and Analysis Method between Mobile Communication Systems, The 15th National Conference on Electromagnetic Compatibility, pp. 38-39 (March 2005)). However, the detection probability therein was obtained by inputting parameters such as transmission power, receiving bandwidth, and path loss into the SEAMCAT simulation software. The path loss considered was point-to-point. In actual situations, the interference signals received by the navigation receiver are not only directly emitted by the interference transmitter, but the propagation paths of the interference signals are complex and diverse, including reflection and scattering on the walls of the system model, etc. Under the requirements of new signal systems and unmanned aerial vehicle platforms, this prediction method is not comprehensive enough and it is difficult to achieve the required accuracy.

[0005] In a communication system, the transmission power of a transmitter and the interference limit of a receiver are known. What is difficult to predict is the signal propagation path loss, which is defined as the attenuation degree of the radiation field of the signal emitted by an airborne device in the direction of a navigation receiver to the navigation receiver antenna. Measuring the true propagation path loss requires building a large shielded chamber, measuring the receiving / transmitting cable loss, the correction factor, load factor, and background noise of the shielded chamber, etc., so that the path loss of signal propagation can be accurately measured. However, the measurement cost is high and the time consumption is long. Once the device position or platform changes, re-measurement is required. Summary of the Invention

[0006] The present disclosure aims to solve at least one of the technical problems in the related art to some extent.

[0007] To this end, the first object of the present disclosure is to propose a method for predicting the electromagnetic compatibility between a navigation receiver and an airborne device, so as to solve the problem that the prediction of electromagnetic compatibility is inaccurate due to the poor prediction effect of path loss in the prior art.

[0008] The second object of the present disclosure is to propose a system for predicting the electromagnetic compatibility between a navigation receiver and an airborne device.

[0009] To achieve the above object, an embodiment of the first aspect of the present disclosure proposes a method for predicting the electromagnetic compatibility between a navigation receiver and an airborne device, including:

[0010] Construct a training data set, where the training data set includes characteristic parameters and path loss obtained through experiments;

[0011] Use the training data set to train a path loss prediction model, and optimize the weight threshold of the path loss prediction model through an intelligent algorithm during the training process, so as to obtain a trained path loss prediction model;

[0012] Collect the transmission signals of the airborne device in real time, and extract target characteristic parameters based on the transmission signals;

[0013] Input the target characteristic parameters into the trained path loss prediction model to obtain a target path loss;

[0014] Input the target path loss into an electromagnetic compatibility prediction model for calculation to determine whether the navigation receiver and the airborne device are electromagnetic compatible.

[0015] The method of the embodiments of the present disclosure constructs a training data set, which includes feature parameters and path loss obtained through experiments. The path loss prediction model is trained using the training data set, and during the training process, the weight threshold of the path loss prediction model is optimized through an intelligent algorithm, so as to obtain a trained path loss prediction model; the transmission signal of the airborne device is collected in real time, and the target feature parameters are extracted based on the transmission signal; the target feature parameters are input into the trained path loss prediction model to obtain the target path loss; the target path loss is input into the electromagnetic compatibility prediction model to determine whether the navigation receiver and the airborne device are electromagnetic compatible. In this case, since the intelligent algorithm is used to optimize the weight threshold of the path loss prediction model during the training process, it prevents the model from falling into local optimization, thus achieving a better prediction effect of the path loss, and further more accurately predicting the electromagnetic compatibility between the navigation receiver and the airborne device, solving the problem in the prior art that the prediction effect of the path loss is not good, resulting in inaccurate prediction of the electromagnetic compatibility.

[0016] In a method for predicting the electromagnetic compatibility between a navigation receiver and an airborne device according to an embodiment of the first aspect of the present disclosure, the construction of the training data set includes: conducting a field experiment, placing a reference antenna at the transmitting antenna of the airborne device, and placing a spectrum analyzer and a tracking source between the reference antenna and the navigation receiver; transmitting a reference signal with a preset power through the reference antenna, measuring the received power at the navigation receiver through the spectrum analyzer and the tracking source, and obtaining the path loss using the preset power and the received power; extracting the feature parameters of the reference signal, and tagging the path loss measured using the reference signal, and obtaining the training data set based on the feature parameters and the tagged path loss.

[0017] In a method for predicting the electromagnetic compatibility between a navigation receiver and an airborne device according to an embodiment of the first aspect of the present disclosure, before training the path loss prediction model using the training data set, the training data set also needs to be normalized.

[0018] In a method for predicting the electromagnetic compatibility between a navigation receiver and an airborne device according to an embodiment of the first aspect of the present disclosure, the path loss prediction model is selected from one of a BP neural network model, an attention mechanism-convolutional neural network-long short-term memory network fusion model, and a convolutional neural network-bidirectional gated recurrent unit neural network-convolutional pooling layer fusion network model.

[0019] In a method for predicting the electromagnetic compatibility between a navigation receiver and an airborne device according to an embodiment of the first aspect of the present disclosure, the intelligent algorithm is selected from one of a simulated annealing algorithm, a genetic algorithm, and an ant colony algorithm.

[0020] In a method for predicting the electromagnetic compatibility between a navigation receiver and an airborne device according to an embodiment of the first aspect of the present disclosure, the electromagnetic compatibility prediction model is selected from one of a Monte Carlo model, an interference prediction equation, and a combined interference frequency test method.

[0021] In a method for predicting the electromagnetic compatibility between a navigation receiver and an airborne device according to an embodiment of the first aspect of the present disclosure, the characteristic parameters include at least one of signal form, signal frequency, signal transmission power, fuselage length, number of portholes, number of cabin doors, antenna polarization mode, and relative position between the porthole and the receiving antenna.

[0022] In a method for predicting the electromagnetic compatibility between a navigation receiver and an airborne device according to an embodiment of the first aspect of the present disclosure, the activation function used in the training process is selected from one of a ReLU function, a Sigmoid function, and a tanh function.

[0023] To achieve the above object, an embodiment of the second aspect of the present disclosure proposes a system for predicting the electromagnetic compatibility between a navigation receiver and an airborne device, including:

[0024] A training data set construction module for constructing a training data set, where the training data set includes characteristic parameters and path loss obtained by experiments.

[0025] A path loss modeling module for training a path loss prediction model using the training data set and optimizing the weight threshold of the path loss prediction model through an intelligent algorithm during the training process, so as to obtain a trained path loss prediction model.

[0026] An acquisition module for real-time collecting the transmission signals of the airborne device and extracting target characteristic parameters based on the transmission signals.

[0027] A calculation module for inputting the target characteristic parameters into the trained path loss prediction model to obtain a target path loss.

[0028] A prediction module for inputting the target path loss into the electromagnetic compatibility prediction model for calculation to determine whether the navigation receiver and the airborne device are electromagnetically compatible.

[0029] In the system according to the embodiments of the present disclosure, a training dataset construction module constructs a training dataset, which includes feature parameters and path loss obtained through experiments. A path loss modeling module trains a path loss prediction model using the training dataset, and optimizes the weight threshold of the path loss prediction model through an intelligent algorithm during the training process, so as to obtain a trained path loss prediction model; an acquisition module collects the transmission signals of the airborne device in real time and extracts target feature parameters based on the transmission signals; a calculation module inputs the target feature parameters into the trained path loss prediction model to obtain a target path loss; a prediction module inputs the target path loss into an electromagnetic compatibility prediction model to determine whether the navigation receiver and the airborne device are electromagnetic compatible. In this case, since the intelligent algorithm is used to optimize the weight threshold of the path loss prediction model during the training process, it prevents the model from falling into local optimization, thus achieving a better prediction effect of the path loss, and further more accurately predicting the electromagnetic compatibility between the navigation receiver and the airborne device, solving the problem in the prior art that the prediction effect of the path loss is not good, resulting in inaccurate prediction of the electromagnetic compatibility.

[0030] In an electromagnetic compatibility prediction system for a navigation receiver and an airborne device according to the second aspect of the present disclosure, the path loss modeling module is further configured to: the path loss prediction model is selected from one of a BP neural network model, an attention mechanism-convolutional neural network-long short-term memory network fusion model, and a convolutional neural network-bidirectional gated recurrent unit neural network-convolutional pooling layer fusion network model; the prediction module is further configured to: the electromagnetic compatibility prediction model is selected from one of a Monte Carlo model, an interference prediction equation, and a combined interference frequency test method.

[0031] Additional aspects and advantages of the present disclosure will be given in part in the following description, become apparent in part from the following description, or be learned through the practice of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The above and / or additional aspects and advantages of the present disclosure will become apparent and be readily understood from the following description of the embodiments in conjunction with the drawings, where:

[0033] Figure 1 is a schematic flowchart of an electromagnetic compatibility prediction method for a navigation receiver and an airborne device provided by an embodiment of the present disclosure;

[0034] Figure 2 is a schematic diagram of the antenna positions of a navigation receiver and an airborne device provided by an embodiment of the present disclosure;

[0035] FIG. 3(a) is a first signal propagation path diagram provided by an embodiment of the present disclosure;

[0036] FIG. 3(b) is a second signal propagation path diagram provided by an embodiment of the present disclosure;

[0037] Figure 4(a) is a schematic position diagram of the first loss measurement device provided by the embodiment of the present disclosure;

[0038] Figure 4(b) is a schematic position diagram of the second loss measurement device provided by the embodiment of the present disclosure;

[0039] Figure 4(c) is a schematic position diagram of the third loss measurement device provided by the embodiment of the present disclosure;

[0040] Figure 4(d) is a schematic position diagram of the fourth loss measurement device provided by the embodiment of the present disclosure;

[0041] Figure 4(e) is a schematic position diagram of the fifth loss measurement device provided by the embodiment of the present disclosure;

[0042] Figure 5 is a schematic flow diagram of another method for predicting the electromagnetic compatibility between a navigation receiver and an airborne device provided by the embodiment of the present disclosure;

[0043] Figure 6 is a schematic diagram of a model based on the simulated annealing algorithm and the DCNN - BiGRUs - CP neural network provided by the embodiment of the present disclosure;

[0044] Figure 7 is a schematic structural diagram of a system for predicting the electromagnetic compatibility between a navigation receiver and an airborne device provided by the embodiment of the present disclosure. Detailed implementation manners

[0045] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the embodiments of the present disclosure. On the contrary, they are merely examples of devices and methods consistent with some aspects of the embodiments of the present disclosure as detailed in the appended claims.

[0046] In the description of this specification, the description referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials, or characteristics described in connection 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 can be combined in a suitable manner in any one or more embodiments or examples. In addition, without conflict, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0047] In addition, the terms "first" and "second" are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In the description of the present disclosure, "a plurality of" means at least two, such as two, three, etc., unless otherwise specifically defined. It should also be understood that the term "and / or" used in the present disclosure refers to and includes any and all possible combinations of one or more of the associated listed items.

[0048] Embodiments of the present disclosure will be described in detail below. Examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present disclosure and should not be construed as limiting the present disclosure.

[0049] The present disclosure will be described in detail below in conjunction with specific embodiments.

[0050] Figure 1 It is a schematic flow chart of a method for predicting the electromagnetic compatibility between a navigation receiver and an airborne device provided by an embodiment of the present disclosure. The embodiment of the present disclosure provides a method for predicting the electromagnetic compatibility between a navigation receiver and an airborne device to solve the problem that the prediction of path loss in the prior art is not good, resulting in inaccurate prediction of electromagnetic compatibility. The method for predicting the electromagnetic compatibility between a navigation receiver and an airborne device provided by the embodiment of the present disclosure is applied in a satellite navigation system, such as Figure 1 As shown, the method for predicting the electromagnetic compatibility between a navigation receiver and an airborne device includes the following steps:

[0051] Step S101: Construct a training data set, which includes characteristic parameters and path loss obtained through experiments.

[0052] In step S101, the characteristic parameters and path loss obtained through experiments are mainly obtained by using a reference antenna and a loss measurement device with reference to the transmitting antenna of the airborne device. The loss measurement device includes a spectrum analyzer and a tracking source.

[0053] In this embodiment, the airborne equipment includes but is not limited to types such as Distance Measuring Equipment (DME), Localizers (LOC), Glide Slope (GS), VHF omnidirectional radio range (VOR), Automatic Dependent Surveillance Broadcast (ADS-B), and Very High Frequency (VHF) communication.

[0054] First, taking an aircraft using the Beidou satellite navigation system as an example, the antenna positions of the navigation receiver and the airborne equipment in the aircraft, as well as the propagation path of the transmission signal of the airborne equipment, are briefly introduced.

[0055] Figure 2 It is a schematic diagram of the antenna positions of the navigation receiver and the airborne equipment provided by the embodiment of the present disclosure. As Figure 2 shown, the antenna positions of DME, VOR, LOC, GS, VHF, ADS-B, and the Beidou receiver are distributed at the nose, fuselage, and tail of the aircraft. In addition, there are other types of airborne equipment in the aircraft, and their antenna positions are not marked in the figure. The transmission signal of the airborne equipment is emitted from the transmitting antenna of the airborne equipment and received through the receiving antenna of the Beidou receiver.

[0056] Figure 3(a) is the first signal propagation path diagram provided by the embodiment of the present disclosure; Figure 3(b) is the second signal propagation path diagram provided by the embodiment of the present disclosure. The aircraft is modeled, and the local diagrams of different positions of the modeled aircraft are shown in Figure 3(a) and Figure 3(b). As shown in Figure 3(a) and Figure 3(b), the fuselage is modeled as a cylinder, and the fuselage has several portholes and a hatch. Multiple airborne device transmitting antennas (including the first transmitting antenna A1 and the first transmitting antenna A2) and the receiving antenna B of one Beidou receiver are arranged on the fuselage shown in Figure 3(a) and Figure 3(b). Among them, there are two relative positions between the Beidou receiver antenna (i.e., the receiving antenna B) and the airborne device transmitting antenna. One is respectively above and below the aircraft (see Figure 3(a)), and the other is above the aircraft together (see Figure 3(b)). The partial propagation paths in the two relative positions are shown in the figure, where yellow represents the direct path, blue represents the diffraction path, red represents the reflection path, and green represents the transmission path. In addition to the various propagation modes shown in the figure, the absorption effects of the porthole glass and the aircraft shell on the signal also need to be considered. It should be noted that the propagation paths in the figure only serve as representatives, indicating the possible modes of signal directivity, scattering, transmission, and diffraction, and are not all propagation paths. Compared with the diffracted signal and the signal that has only undergone one reflection, transmission, and diffraction, the signal after multiple reflections, transmissions, and scatterings has become very weak and can be ignored.

[0057] After understanding the antenna positions of the navigation receiver and the airborne devices in the aircraft, as well as the propagation path of the transmission signal of the airborne devices to the receiving antenna, referring to the position of the airborne device transmitting antenna, a training data set is constructed through on-site experiments using a reference antenna.

[0058] Specifically, a training data set is constructed, including: conducting a field experiment, placing a reference antenna at the transmitting antenna of the airborne device, and placing a spectrum analyzer and a tracking source between the reference antenna and the navigation receiver; transmitting a reference signal with a preset power through the reference antenna, measuring the received power at the navigation receiver through the spectrum analyzer and the tracking source, and obtaining the path loss by using the preset power and the received power; extracting the characteristic parameters of the reference signal, labeling the path loss measured by using the reference signal, and obtaining the training data set based on the characteristic parameters and the labeled path loss. Among them, placing a reference antenna at the transmitting antenna of the airborne device specifically means using the transmitting wire of the airborne device to be measured as the test location, and placing a reference antenna at each test location, so the positions of the reference antenna and the transmitting antenna of the airborne device are the same. Placing a spectrum analyzer and a tracking source between the reference antenna and the navigation receiver means placing the spectrum analyzer and the tracking source in the middle of the path to be measured between the transmitting antenna and the Beidou receiving antenna. A set of spectrum analyzer and tracking source is set on each path, and the reference antenna and the receiving antenna are respectively connected to the tracking source and the spectrum analyzer through cables. Since the power of the reference signal transmitted by the reference antenna is known, the power at the receiving antenna (i.e., the received power) is measured by using the spectrum analyzer and the tracking source, and the ratio is the path loss. This ratio specifically refers to the ratio of the received power of the receiving antenna to the transmitting power of the reference antenna during the experiment.

[0059] In this embodiment, since there are multiple transmitting antennas of the airborne devices on the aircraft and multiple receiving antennas of the Beidou receivers, there may be propagation paths between each transmitting antenna of the airborne device and the receiving antenna of the Beidou receiver, so there are various positions of the loss measurement device. Fig. 4(a) is a schematic diagram of the position of the first loss measurement device provided by the embodiment of the present disclosure; Fig. 4(b) is a schematic diagram of the position of the second loss measurement device provided by the embodiment of the present disclosure; Fig. 4(c) is a schematic diagram of the position of the third loss measurement device provided by the embodiment of the present disclosure; Fig. 4(d) is a schematic diagram of the position of the fourth loss measurement device provided by the embodiment of the present disclosure; Fig. 4(e) is a schematic diagram of the position of the fifth loss measurement device provided by the embodiment of the present disclosure; Figures 4(a) to 4(e) The positions of the loss measurement device in five cases are exemplified. Among them, the reference antenna is set at the transmitting antenna, the path between the reference antenna and the receiving antenna is the path between the transmitting antenna and the receiving antenna, the line (solid line or dotted line) between the reference antenna and the receiving antenna represents the path, and the square set on the line represents the loss measurement device (such as a spectrum analyzer and a tracking source).

[0060] In this embodiment, the characteristic parameters may include at least one of signal form, signal frequency, signal transmission power, fuselage length, number of portholes, number of cabin doors, antenna polarization mode (transmitting antenna polarization mode or reference antenna polarization mode), and relative position between the porthole and the receiving antenna.

[0061] In step S101, when conducting on-site experiments, a tracking source and a spectrum analyzer are used to measure the path loss from the porthole to the receiving antenna. The signal form, signal transmission frequency, antenna polarization mode, porthole position, and relative position between the porthole and the receiving antenna can be changed, so as to measure several groups of data (or information). A training data set is obtained by using the characteristic parameters of the reference signal and the corresponding measured path loss in all groups of data. Among them, it is also necessary to label the characteristic parameters of the reference signal in each group and the corresponding measured path loss, and obtain the training data set based on the characteristic parameters and the labeled path loss.

[0062] In step S101, before training the path loss prediction model using the training data set (i.e., entering step S102), it is also necessary to preprocess the training data set. The preprocessing can be, for example, normalization processing.

[0063] Specifically, the sample parameters in the training data set include characteristic parameters such as the number of portholes, the number of hatches, the transmitting power of the reference antenna (divided into multiple levels), the transmitting signal frequency, the antenna polarization mode, the porthole position, and the relative position between the receiving antenna and the transmitting antenna (on the same side or on the opposite side, the distance is far or near), as well as the measured path loss (i.e., the true value of the path loss).

[0064] Perform normalization processing on the sample parameters: The fuselage length (the interval distance between portholes), the number of portholes, the number of hatches, the antenna polarization mode, and the true value of the path loss are constant values. The transmitting power of the reference antenna is one-dimensional data, the porthole position, and the relative position between the Beidou receiver and the transmitting antenna are two-dimensional data. Taking the front vertex of the aircraft as the origin of the coordinate axis, the direction from the nose to the tail is the positive x-axis direction, the wing is parallel to the y-axis, the tail is parallel to the z-axis, and the installation position of the airborne antenna is the xyz coordinates of the antenna in this three-dimensional coordinate system. Since the operating frequency band and polarization mode of the airborne antenna have a great influence on the path loss of electromagnetic wave propagation, which is related to the penetration and coupling ability of different electromagnetic waves, the path loss in the low frequency band is relatively large, and the path loss in the high frequency band is relatively small. The antenna polarization mode is represented by 0 and 1, where 0 represents horizontal polarization and 1 represents vertical polarization.

[0065] Step S102: Use the training data set to train the path loss prediction model, and optimize the weight threshold of the path loss prediction model through an intelligent algorithm during the training process, so as to obtain a trained path loss prediction model.

[0066] In step S102, when training, 70% of the training data set can be randomly selected as the training data for training, and the remaining 30% can be used as the test data for testing. Among them, the characteristic parameters in the training data set are the input data of the path loss prediction model, and the path loss in the training data set is the label.

[0067] In step S102, the activation function adopted during the training process is selected from one of the ReLU function, the Sigmoid function, and the tanh function. The activation function can introduce non-linear factors, accelerate the convergence of the network, reduce the interdependence of parameters, avoid the overfitting problem of the model, and improve the generalization ability of the model.

[0068] In step S102, the path loss prediction model can include three parts, namely, a feature extraction part, a feature learning part, and an output part. Among them, the feature extraction part is used to automatically extract the features of the input data of the path loss prediction model, perform deeper and more abstract processing on the input data, so as to realize the feature extraction of the input data. The feature learning part is used to better learn the dependency relationship between data features and improve the prediction accuracy of the model. The output part is used to output the path loss prediction value predicted by the path loss prediction model.

[0069] In step S102, the path loss prediction model can be selected from one of a BP neural network model, an attention mechanism-convolutional neural network-long short-term memory network fusion model (Attention-CNN-LSTM), and a convolutional neural network-bidirectional gated recurrent unit neural network-convolutional pooling layer fusion network model (1DCNN-BiGRUs-CP).

[0070] In step S102, since the path loss prediction model is composed of a neural network, considering that the gradient training algorithm of the neural network lacks global search ability and cannot converge to the global minimum point of the error surface for functions with a large search space, multiple peaks, and non-differentiability, the intelligent algorithm reveals certain natural processes through simulation, has high-efficiency global parallel optimization performance, robustness, generality, and does not require gradient information of the problem. Therefore, during the training process, the weight threshold of the path loss prediction model is optimized by the intelligent algorithm. Thus, it is possible to jump out of the local optimal solution and reach the global optimal solution.

[0071] In this embodiment, the intelligent algorithm can be selected from one of the simulated annealing algorithm, the genetic algorithm, and the ant colony algorithm.

[0072] Step S103, collect the transmission signals of the airborne equipment in real time, and extract the target feature parameters based on the transmission signals.

[0073] In step S103, the target feature parameters are the feature parameters of the transmission signals of the airborne equipment obtained in real time, and the target feature parameters can also be referred to as the feature parameters of the path loss to be measured. Among them, the feature parameters include but are not limited to signal form, signal frequency, signal transmission power, fuselage length, number of portholes, number of cabin doors, polarization mode of the transmitting antenna, and relative position between the porthole and the receiving antenna.

[0074] In step S103, the types of the target feature parameters extracted are the same as those of the feature parameters in the training dataset.

[0075] Step S104: Input the target feature parameters into the trained path loss prediction model to obtain the target path loss.

[0076] In step S104, the target path loss is the path loss prediction value output by the trained path loss prediction model.

[0077] Step S105: Input the target path loss into the electromagnetic compatibility prediction model for calculation to determine whether the navigation receiver and the airborne equipment are electromagnetic compatible.

[0078] In this embodiment, the electromagnetic compatibility prediction model is selected from one of the Monte Carlo model, the interference prediction equation, and the combined interference frequency test method.

[0079] Please refer to Figure 5 , Figure 5 which is a schematic flowchart of another electromagnetic compatibility prediction method for a navigation receiver and an airborne equipment provided by an embodiment of the present disclosure. Figure 6 which is a schematic diagram of a model based on the simulated annealing algorithm and the DCNN - BiGRUs - CP neural network. In the electromagnetic compatibility prediction method for a navigation receiver and an airborne equipment shown in Figure 5 and Figure 6 taking the A320 - 200 as the actual measurement aircraft model, the Beidou satellite navigation system as the navigation system, the path loss prediction model as the convolutional neural network - bidirectional gated recurrent unit neural network - convolutional pooling layer fusion network model (1DCNN - BiGRUs - CP), the intelligent algorithm as the simulated annealing algorithm, and the electromagnetic compatibility prediction model as the Monte Carlo model as an example, the electromagnetic compatibility prediction between the airborne equipment and the Beidou receiving system is carried out. Figure 5 The electromagnetic compatibility prediction method for a navigation receiver and an airborne equipment shown in

[0080] Step S201: Conduct a field experiment to measure the path loss of the signal propagation from each reference antenna to the Beidou receiving antenna, and then obtain the training dataset.

[0081] The specific process of obtaining the training dataset is as described in step S101 above and will not be elaborated here.

[0082] In step S201, a field experiment is carried out. The reference antenna is placed at the porthole position, and a spectrum analyzer and a tracking source are used to measure the power ratio between the reference antenna and the receiving antenna, that is, the path loss of the reference signal propagating between the two antennas. The fuselage of an A320-200 aircraft is 37.57 meters long in total, and the cabin length is 27.51 meters. It has 2×40 portholes, 4 emergency exits (the 16th and 17th portholes), and 4 cabin doors. By changing the signal form, signal transmission power, antenna polarization mode, porthole position, and relative position between the porthole and the receiving antenna, multiple groups of path losses are measured. The signal forms include six signals: DME, VOR, LOC, GS, VHF, and ADS-B. The reference antenna polarization modes of the six signals are vertical polarization, horizontal polarization, horizontal polarization, horizontal polarization + vertical polarization, and vertical polarization respectively. The transmission frequency ranges of the six signals are 962 - 1213 MHz, 108 - 117.95 MHz, 108.10 - 111.95 MHz, 328.6 - 335.4 MHz, 118 - 137 MHz, and 1030 MHz respectively. For each signal, 40 frequency points are equally spaced within its transmission frequency range as the transmission frequencies of the experiment. Among them, the ADS-B frequency is 1030 MHz and remains fixed. A total of 84×(5×40 + 1) groups of data are obtained, of which 70% (11819 groups of data) are used for neural network training, and 30% (5065 groups of data) are used for testing the neural network. The obtained multiple groups of data are normalized. Among them, the fuselage length (the interval distance between portholes), the number of portholes, the number of cabin doors, the antenna polarization mode, and the true value of the path loss are constant values. The reference antenna transmission power is one-dimensional data, and the porthole position and the relative position between the Beidou receiver and the transmitting antenna are two-dimensional data. Taking the front vertex of the aircraft as the origin of the coordinate axis, the direction from the nose to the tail is the positive x-axis direction, the wing is parallel to the y-axis, the tail is parallel to the z-axis, and the installation position of the airborne antenna is the xyz coordinates of the antenna in this three-dimensional coordinate system. The antenna polarization mode is represented by 0 and 1, where 0 represents horizontal polarization and 1 represents vertical polarization.

[0083] Step S202, construct a path loss prediction model, and use the training data set to train the path loss prediction model.

[0084] The specific process of constructing and training the path loss prediction model is as in step S102 above and will not be elaborated here.

[0085] In step S202, the path loss prediction model is a convolutional neural network - bidirectional gated recurrent unit neural network - convolutional pooling layer fusion network model (1DCNN - BiGRUs - CP). Therefore, constructing the path loss prediction model includes constructing a one-dimensional convolutional neural network, constructing a BiGRUs layer (bidirectional gated recurrent unit neural network), and constructing a CP layer (convolutional pooling layer).

[0086] Among them, a one-dimensional convolutional neural network (1DCNN) is constructed, and the one-dimensional convolutional neural network is used for feature extraction of input data. The one-dimensional convolutional neural network can realize automatic extraction of input data features and perform deeper and more abstract processing on the input data. However, directly using 1DCNN is likely to cause loss of some important information and affect the prediction accuracy. Specifically, as Figure 6 shown, the 1DCNN consists of an input layer (8*1), 3 convolutional layers, 3 pooling layers (a 5*1 pooling layer and 2 4*1 pooling layers), a fully connected layer, and an output layer. The number of convolutional kernels in the 3 convolutional layers is 128, 64, and 32 respectively. The first convolutional layer uses the largest number of neurons to capture a large number of data features, and then gradually decreases to extract the main features. A random dropout mechanism is added after each convolutional layer to reduce the mutual dependence between neurons, and the dropout (random inactivation regularization) index is 0.4. The sizes of the 3 pooling layers are 4, 5, and 5 respectively, and the maximum pooling method is adopted. The ReLU function is used as the activation function for the convolutional layer and the fully connected layer. During the training process, the training data set is input into the neural network. First, it passes through the one-dimensional 1DCNN layer to extract features of the sample data. The output data of the output layer of the 1DCNN layer is input into the BiGRUs layer.

[0087] Construct a BiGRUs layer. The BiGRUs layer can better learn the dependence relationship between data features and improve the prediction accuracy of the model. Replace the fully connected layer in the BiGRUs layer with a convolutional layer and a pooling layer, which reduces the model training parameters while improving the prediction performance of the hybrid model. The input of the BiGRUs layer is the output of the 1DCNN.

[0088] The bidirectional gated recurrent unit neural network (BiGRUs) model is composed of forward and backward propagating GRUs (Gated Recurrent Unit), which can prevent gradient explosion and improve the model operation efficiency. The GRU includes a reset gate and an update gate. The update gate determines the degree to which the state information of the previous moment is retained in the current state. The larger its value, the more state information of the previous moment is retained. The reset gate determines the degree to which historical information is forgotten. The smaller its value, the more information is forgotten. The BiGRUs model connects the input data to further learn deep features. The BiGRUs layer in this embodiment can be obtained by removing the fully connected layer from the BiGRUs model. Specifically, as Figure 6 shown, the BiGRUs layer in this embodiment consists of two BiGRU networks, a backward propagating GRU and a forward propagating GRU, and the number of their neurons is 32 and 16 respectively. The data output by the BiGRUs layer is input into the CP layer.

[0089] Construct the CP layer (convolutional pooling layer) to replace the fully connected layer in BiGRUs, reducing the model training parameters. The output of the BiGRUs layer is the input of the CP layer, and the output of the last pooling layer of the CP layer is the prediction result. Specifically, as Figure 6 shown, the CP layer in this embodiment includes 2 convolutional layers and 2 pooling layers. The number of convolutional kernels of the 2 convolutional layers are 16 and 8 respectively. A random dropout mechanism is added after the convolutional layer, and the dropout coefficient is 0.4. The sizes of the pooling layers are 5 and 4 respectively.

[0090] In step S202, use the training data set to train the convolutional neural network - bidirectional gated recurrent unit neural network - convolutional pooling layer fusion network model (1DCNN - BiGRUs - CP).

[0091] In step S202, when training 1DCNN - BiGRUs - CP, train until the error of the loss function meets the requirements to complete the training. Among them, the Adam optimizer is used, and the loss function uses HM - MSE - Score (the combined loss function of the harmonic mean of MSE and Score). The harmonic coefficient λ of the loss function is 0.6. In addition, after training the 1DCNN - BiGRUs - CP neural network model (which can be simply referred to as the neural network) using the training data in the training data set, it is also necessary to input the test data into the neural network for testing. After meeting the standards, the trained neural network is obtained.

[0092] Step S203, use the simulated annealing algorithm to optimize the weight threshold of the neural network.

[0093] In step S203, using the simulated annealing algorithm to optimize the parameters (weight threshold) can prevent the neural network from falling into local optimization. The simulated annealing algorithm introduces the transfer probability, a random factor, in the search process. It accepts a solution that is worse than the current solution with a certain probability, enabling the neural network to jump out of the local optimal solution and reach the global optimal solution.

[0094] Take the weight threshold of the neural network with the determined structure as the input of the simulated annealing algorithm. The weight threshold is set to a relatively high random value; the annealing rate uses the exponential decay method T(n)=αT(n), where α ranges from 0.8 to 0.99. There are enough transfer attempts for each temperature, and the convergence speed is slow to ensure that the global minimum point can be reached; the transfer probability is:

[0095]

[0096] Wherein, P is the transition probability, C(·) is the cost function, n is the number of iterations, T is the temperature. After the temperature reaches the threshold, the result of the simulated annealing algorithm is used as the weight threshold of the neural network. Using the simulated annealing algorithm to calculate the weight threshold of the neural network can enable the neural network to jump out of the local optimum and search for the global optimum.

[0097] Step S204, until the error of the loss function meets the requirements, complete the training, and then obtain the path loss prediction value.

[0098] In step S204, after completing the training, obtain the trained 1DCNN-BiGRUs-CP neural network model, and input the characteristic parameters of the path loss to be measured into the trained 1DCNN-BiGRUs-CP neural network model to obtain the path loss prediction value.

[0099] The specific process is as in the above steps S103 and S104, and will not be elaborated here.

[0100] Step S205, establish a Monte Carlo model, substitute various parameters into the Monte Carlo model, and calculate the electromagnetic compatibility between the Beidou receiver and the airborne equipment.

[0101] The specific process is as in the above step S105, and will not be elaborated here.

[0102] In step S205, construct a Monte Carlo prediction model, and establish an electromagnetic compatibility prediction model between the airborne equipment and the Beidou receiving system (i.e., the Beidou receiver). The input parameter of the Monte Carlo model is the path loss prediction value output by the 1DCNN-BiGRUs-CP neural network model. The Monte Carlo method adopted by the Monte Carlo prediction model establishes a model from the emission source (such as the airborne equipment), the receiver (such as the Beidou receiver), and the propagation path, and can calculate the interference probability according to the transmission power, reception power, and path loss of the antenna, so as to accurately predict the electromagnetic compatibility between the Beidou space signal and the airborne equipment signal.

[0103] In step S205, the useful signal power dRSS received by the Beidou receiver is:

[0104] dRSS = p wt + g wt→vr - pl wt→vr (fvr) + g vr→wt

[0105] Wherein, p wt is the power of the Beidou transmitter; f vr is the frequency of the Beidou receiver (which can be a constant or a specific distribution); pl wt→vr is the path loss between the Beidou transmitter and the receiver; g wt→vr is the gain of the transmitting antenna in the direction of the receiver; gvr→wt is the gain of the receiving antenna in the direction of the transmitter. Among them, the power of the Beidou transmitter, the frequency of the Beidou receiver, the path loss between the Beidou transmitter and the receiver, the gain of the transmitting antenna in the direction of the receiver, and the gain of the receiving antenna in the direction of the transmitter can all be obtained from the Beidou satellite navigation system.

[0106] The power of the signal (interference signal) of a certain airborne device is:

[0107]

[0108] In the formula, p it is the power of the transmitter of a certain (one) airborne device (the power of the transmitted signal); f it is the operating frequency of the airborne device; is the power control gain of a certain (one) airborne device; g it→vr is the gain of the transmitting antenna of a certain (one) airborne device in the direction of the Beidou receiver; g vr→it is the gain of the Beidou receiving antenna in the direction of a certain (one) airborne device; a vr is the attenuation of the Beidou receiver; pl it→vr is the path loss between a certain (one) airborne device and the Beidou receiver. Among them, the power of the airborne device transmitter, the operating frequency of the airborne device, the power control gain of the airborne device, the gain of the airborne device transmitting antenna in the direction of the Beidou receiver, the gain of the Beidou receiving antenna in the direction of each airborne device, and the attenuation of the Beidou receiver can all be obtained from the Beidou satellite navigation system. The path loss between each airborne device and the Beidou receiver can be obtained through the trained 1DCNN - BiGRUs - CP neural network model.

[0109] The total interference signal power iRSS received by the Beidou receiver is the sum of the weighted powers of the signals of each airborne device received by the Beidou receiver. That is:

[0110]

[0111] In the formula, k is the number of airborne devices, ω i is the weight of the signal of the i-th airborne device.

[0112] The judgment criterion for determining whether the Beidou receiver is interfered is: compare the ratio of the useful signal power dRSS received by the Beidou receiver to the total interference signal power iRSS received by the Beidou receiver with the judgment criterion value. If is greater than the judgment criterion value, it is considered that there is no interference, that is, the Beidou receiving system and the airborne device system are electromagnetic compatible.

[0113] In the method of the embodiments of the present disclosure, a training data set is constructed. The training data set includes characteristic parameters and path loss obtained through experiments. The path loss prediction model is trained using the training data set, and during the training process, the weight threshold of the path loss prediction model is optimized through an intelligent algorithm, so as to obtain a trained path loss prediction model; the transmission signals of the airborne device are collected in real time, and target characteristic parameters are extracted based on the transmission signals; the target characteristic parameters are input into the trained path loss prediction model to obtain the target path loss; the target path loss is input into the electromagnetic compatibility prediction model to determine whether the navigation receiver and the airborne device are electromagnetic compatible. In this case, since the intelligent algorithm is used to optimize the weight threshold of the path loss prediction model during the training process, the model is prevented from falling into local optimization, thereby achieving a better prediction effect of the path loss, and further more accurately predicting the electromagnetic compatibility between the navigation receiver and the airborne device, solving the problem in the prior art that the prediction effect of the path loss is not good, resulting in inaccurate prediction of the electromagnetic compatibility.

[0114] The following is an embodiment of the system of the present disclosure, which can be used to execute the embodiment of the method of the present disclosure. For the details not disclosed in the embodiment of the system of the present disclosure, please refer to the embodiment of the method of the present disclosure.

[0115] Please refer to Figure 7 , Figure 7 which is a schematic structural diagram of the electromagnetic compatibility prediction system for the navigation receiver and the airborne device provided by the embodiment of the present disclosure. The electromagnetic compatibility prediction system for the navigation receiver and the airborne device can be implemented as all or part of a device through software, hardware, or a combination of both. The electromagnetic compatibility prediction system for the navigation receiver and the airborne device is applied in a satellite navigation system. The electromagnetic compatibility prediction system 10 for the navigation receiver and the airborne device includes a training data set construction module 11, a path loss modeling module 12, an acquisition module 13, a calculation module 14, and a prediction module 15, where:

[0116] The training data set construction module 11 is used to construct a training data set, and the training data set includes characteristic parameters and path loss obtained through experiments;

[0117] The path loss modeling module 12 is used to train the path loss prediction model using the training data set, and during the training process, the weight threshold of the path loss prediction model is optimized through an intelligent algorithm, so as to obtain a trained path loss prediction model;

[0118] The acquisition module 13 is used to collect the transmission signals of the airborne device in real time and extract target characteristic parameters based on the transmission signals;

[0119] The calculation module 14 is used to input the target characteristic parameters into the trained path loss prediction model to obtain the target path loss;

[0120] The prediction module 15 is configured to input the target path loss into the electromagnetic compatibility prediction model for calculation to determine whether the navigation receiver and the airborne device are electromagnetic compatible.

[0121] Optionally, the path loss modeling module 12 is further configured to: the path loss prediction model is selected from one of a BP neural network model, an attention mechanism-convolutional neural network-long short-term memory network fusion model, and a convolutional neural network-bidirectional gated recurrent unit neural network-convolutional pooling layer fusion network model; the prediction module is further configured to: the electromagnetic compatibility prediction model is selected from one of a Monte Carlo model, an interference prediction equation, and a combined interference frequency test method.

[0122] It should be noted that when the electromagnetic compatibility prediction system of the navigation receiver and the airborne device provided in the above embodiment executes the electromagnetic compatibility prediction method of the navigation receiver and the airborne device, only the division of the above functional modules is used for illustration. In practical applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the electronic device is divided into different functional modules to complete all or part of the functions described above. In addition, the electromagnetic compatibility prediction system of the navigation receiver and the airborne device provided in the above embodiment and the embodiment of the electromagnetic compatibility prediction method of the navigation receiver and the airborne device belong to the same concept. The implementation process is detailed in the method embodiment and will not be repeated here.

[0123] In the system of the embodiments of the present disclosure, the training dataset construction module constructs a training dataset, the training dataset includes the characteristic parameters and path loss obtained by experiments, the path loss modeling module uses the training dataset to train the path loss prediction model, and optimizes the weight threshold of the path loss prediction model through an intelligent algorithm during the training process, so as to obtain a trained path loss prediction model; the acquisition module collects the transmission signals of the airborne device in real time and extracts the target characteristic parameters based on the transmission signals; the calculation module inputs the target characteristic parameters into the trained path loss prediction model to obtain the target path loss; the prediction module inputs the target path loss into the electromagnetic compatibility prediction model to determine whether the navigation receiver and the airborne device are electromagnetic compatible. In this case, since an intelligent algorithm is used during the training process, the intelligent optimization algorithm globally searches for the weight threshold of the neural network to prevent the model from falling into local optimization, so as to achieve a better prediction effect of the path loss, and further more accurately predict the electromagnetic compatibility between the navigation receiver and the airborne device, solving the problem that the prediction of the electromagnetic compatibility is inaccurate due to the poor prediction effect of the path loss in the prior art. The prediction system of the present disclosure provides a fast, flexible and accurate predictor, which can effectively predict the electromagnetic compatibility between the navigation receiver and the airborne device, and provide a basis for ensuring the functions of satellite navigation devices and the installation requirements of airborne avionics devices.

[0124] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution disclosed in this disclosure can be achieved. This disclosure places no restrictions herein.

[0125] The above specific embodiments do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure shall be included within the scope of protection of this disclosure.

Claims

1. A method for predicting the electromagnetic compatibility between a navigation receiver and airborne equipment, characterized in that, it includes: Construct a training data set, which includes characteristic parameters and path loss obtained through experiments. Among them, the characteristic parameters include at least one of signal form, signal frequency, signal transmission power, fuselage length, number of portholes, number of hatches, antenna polarization mode, and relative position between the porthole and the receiving antenna; Use the training data set to train the path loss prediction model, and optimize the weight threshold of the path loss prediction model through an intelligent algorithm during the training process, so as to obtain a trained path loss prediction model. Among them, the intelligent algorithm is selected from one of the simulated annealing algorithm, genetic algorithm, and ant colony algorithm; Collect the transmission signals of the airborne equipment in real time, and extract the target characteristic parameters based on the transmission signals; Input the target characteristic parameters into the trained path loss prediction model to obtain the target path loss; Input the target path loss into the electromagnetic compatibility prediction model for calculation to determine whether the navigation receiver and the airborne equipment are electromagnetically compatible.

2. The method for predicting the electromagnetic compatibility between a navigation receiver and airborne equipment according to claim 1, characterized in that, The construction of the training data set includes: Conduct on-site experiments, place a reference antenna at the transmitting antenna of the airborne equipment, and place a spectrum analyzer and a tracking source between the reference antenna and the navigation receiver; Transmit a reference signal with a preset power through the reference antenna, measure the received power at the navigation receiver through the spectrum analyzer and the tracking source, and obtain the path loss by using the preset power and the received power; Extract the characteristic parameters of the reference signal, label the path loss measured by using the reference signal, and obtain a training data set based on the characteristic parameters and the labeled path loss.

3. The method for predicting the electromagnetic compatibility between a navigation receiver and airborne equipment according to claim 1 or 2, characterized in that, It further includes: Before using the training data set to train the path loss prediction model, it is also necessary to perform normalization processing on the training data set.

4. The method for predicting the electromagnetic compatibility between a navigation receiver and airborne equipment according to claim 1, characterized in that, The path loss prediction model is selected from one of the BP neural network model, the attention mechanism-convolutional neural network-long short-term memory network fusion model, and the convolutional neural network-bi-directional gated recurrent unit neural network-convolutional pooling layer fusion network model.

5. The method for predicting the electromagnetic compatibility between a navigation receiver and airborne equipment according to claim 1, characterized in that, The electromagnetic compatibility prediction model is selected from one of the Monte Carlo model, the interference prediction equation, and the combined interference frequency test method.

6. The method for predicting the electromagnetic compatibility between a navigation receiver and airborne equipment according to claim 1, characterized in that, It further includes: The activation function used during the training process is selected from one of the ReLU function, Sigmoid function, and tanh function.

7. A system for predicting the electromagnetic compatibility between a navigation receiver and airborne equipment, characterized in that, it includes: A training data set construction module for constructing a training data set, where the training data set includes feature parameters and path loss obtained through experiments. Among them, the feature parameters include at least one of signal form, signal frequency, signal transmission power, fuselage length, number of portholes, number of hatches, antenna polarization mode, and relative position between the porthole and the receiving antenna; A path loss modeling module for training a path loss prediction model using the training data set and optimizing the weight threshold of the path loss prediction model through an intelligent algorithm during the training process to obtain a trained path loss prediction model. Among them, the intelligent algorithm is selected from one of simulated annealing algorithm, genetic algorithm, and ant colony algorithm; An acquisition module for real-time collecting transmission signals of airborne equipment and extracting target feature parameters based on the transmission signals; A calculation module for inputting the target feature parameters into the trained path loss prediction model to obtain a target path loss; A prediction module for inputting the target path loss into an electromagnetic compatibility prediction model for calculation to determine whether the navigation receiver and the airborne equipment are electromagnetic compatible.

8. The electromagnetic compatibility prediction system for a navigation receiver and an airborne equipment according to claim 7, characterized in that the path loss modeling module is further used for: the path loss prediction model is selected from one of a BP neural network model, an attention mechanism-convolutional neural network-long short-term memory network fusion model, and a convolutional neural network-bidirectional gated recurrent unit neural network-convolutional pooling layer fusion network model; the prediction module is further used for: the electromagnetic compatibility prediction model is selected from one of a Monte Carlo model, an interference prediction equation, and a combined interference frequency test method.

Images