Aircraft threat assessment method based on multi-dimensional situation fusion in complex electromagnetic environment

Abstract

The application discloses a method for aircraft threat assessment based on multi-dimensional situation fusion in a complex electromagnetic environment, divides the whole flight process of the aircraft, selects main threat factors of different flight stages according to actual conditions, selects appropriate threat factors according to different aircraft stages and performs threat calculation, and finally fuses all threat calculation results to perform threat assessment on the aircraft generating threats. The application utilizes the LSTM method to judge the landing point error, introduces a threat intention element, simultaneously considers the influence of radar elements in the terminal flight section, solves the problem that many radar elements cannot directly calculate the threat value by relying on a single radar factor based on the Dempster-Shafer theory, considers the influence of traditional threat factors such as the aircraft-target distance and aircraft threat intention, simultaneously considers the radar factors to meet the threat assessment demand of the aircraft in the complex electromagnetic environment, and makes the application more suitable for actual engineering demand.

Description

Technical Field

[0001] This invention belongs to the field of radar countermeasures, specifically relating to a method for aircraft threat assessment based on multi-dimensional situational fusion in complex electromagnetic environments. Background Technology

[0002] In current practical working environments, traditional, static selection of aircraft threat factors is no longer sufficient to meet the needs of aircraft threat assessment in complex electromagnetic environments. With the continuous development of aircraft technology and the widespread influence of electromagnetic effects in practical engineering environments, the elements involving aircraft threats constitute a complex and multidimensional dataset. Simple threat factor selection leads to increasingly larger errors in situation assessment. Therefore, it is necessary to divide the entire flight phase of the aircraft in detail to select threat factors that are more in line with the engineering environment, and are multidimensional and dynamic.

[0003] Threat factors influencing aircraft threat values ​​constitute a complex, multi-dimensional dataset. This dataset includes not only the aircraft's flight trajectory data but also a wide range of threat intent and signal data. The data exhibits high redundancy, and the degree to which different data points influence threat values ​​varies significantly; some threat factors may even be decisive. Therefore, it is essential to process this dataset to select primary threat factors, ignore irrelevant factors, and integrate as many factors as possible. Aircraft threat factors include aircraft-target distance, aircraft threat intent, and radar operating parameters. Each category contains a variety of factors, essentially covering all characteristic data from the entire flight process. Considering all this data would be extremely labor-intensive and insufficient for real-time situational assessment in realistic scenarios. However, by categorizing these aircraft threat factors according to their importance, appropriate categories can be selected for aircraft threat assessment based on different needs.

[0004] The flight stage of an aircraft also influences the selection of threat factors. The entire flight process is constantly changing; not only are the threat factor data continuously evolving, but the degree of their impact also changes dynamically. If the entire flight process is treated as a static process with unchanging threat factors, the threat calculation results will deviate significantly from the actual results, greatly impacting decision-making. Therefore, it is necessary to dynamically select threat factors based on the different stages of the aircraft's flight.

[0005] Therefore, we aim to propose a method that aligns with real-world engineering scenarios, enabling detailed segmentation of aircraft operations and the selection of appropriate threat factors for each stage, thereby facilitating more accurate threat calculations. This method considers distance, threat intent, and the operational characteristics of the aircraft's radar equipment to meet the threat assessment needs in complex electromagnetic environments, thus providing a basis for real-time decision-making. Summary of the Invention

[0006] This invention proposes a multi-dimensional situational awareness fusion-based aircraft threat assessment method for complex electromagnetic environments. The method divides the entire flight process of the aircraft and selects the main threat factors for different flight stages based on actual conditions. The LSTM method is used to predict the aircraft's landing point, incorporating the aircraft's threat intent as a threat judgment factor. Simultaneously, the influence of radar factors is considered in the terminal phase of the flight. Subsequently, different methods are used to calculate the threat for different threat factors, and the threat values ​​of each factor are fused to obtain the overall threat assessment value for the aircraft. This provides a methodological basis for aircraft threat assessment in practical engineering environments. This invention considers factors such as aircraft threat intent, aircraft-target distance, and radar operating status, avoiding the large errors in threat calculation results caused by the previous single-factor selection. This results in more accurate threat assessment results, meeting the needs of aircraft threat assessment in complex electromagnetic environments.

[0007] The technical solution for achieving this invention is as follows: a method for aircraft threat assessment based on multi-dimensional situational fusion in complex electromagnetic environments, comprising the following steps:

[0008] Step 1: When the target formation detects an unknown aircraft entering the measurement range, the LSTM method is used to estimate and predict the aircraft. The LSTM neural network is trained using the center of mass position, center of mass velocity, aircraft-target distance, attitude angles and attitude angular rates in the flight trajectory. The relevant network can be used for real-time aircraft trajectory prediction, where the attitude angles include slip angle, pitch angle and yaw angle.

[0009] Step Two: Analyze the entire flight process of the aircraft, distinguishing it based on dynamic characteristics, and divide the entire flight process into initial, middle, and terminal phases. Simultaneously analyze the terminal flight process, dividing it into far-field, middle-field, and near-field phases based on radar operating principles and layered azimuth prediction schemes. The following phase division function is obtained:

[0010]

[0011] In equation (1), since there are multiple threatened targets in a scenario, the target with the highest value is selected as the reference object. d(t) is the distance between the aircraft and target at time t. d1 is the aircraft-target distance threshold between the initial and middle phases of the aircraft. d2 is the aircraft-target distance threshold between the middle and far-end phases of the aircraft. d3 is the aircraft-target distance threshold between the far-end and middle-end phases of the aircraft. d4 is the aircraft-target distance threshold between the middle and near-end phases of the aircraft. Phase(t) represents the phase division function. If Phase(t) = 0, it means the aircraft is in the initial flight phase; if Phase(t) = 1, it means the aircraft is in the middle flight phase; if Phase(t) = 2, it means the aircraft is in the far-end phase; if Phase(t) = 3, it means the aircraft is in the middle-end phase; and if Phase(t) = 4, it means the aircraft is in the near-end phase.

[0012] Correspondingly, a flight phase selection function is constructed for further selection of phase elements. The flight phase selection function Phase_M(t) is as follows:

[0013]

[0014] The distance between the aircraft and the target can effectively determine the flight stage and operational status of the aircraft. However, the threat factors for assessing the threat level of the aircraft differ depending on the flight stage and operational status.

[0015] Step 3: Use the real-time measured distance between the aircraft and the target to determine the flight stage of the aircraft. Assume the flight stage at time t and output the Phase(t) value to select the parameter value under the corresponding stage.

[0016] Step 4: Calculate the threat value of the aircraft's threatening intent. Based on the center-of-mass position, center-of-mass velocity, attitude angle, and attitude angular velocity values ​​of the aircraft M measured at time t, input them into the trained LSTM neural network to complete the trajectory prediction of the aircraft. Then, use the trajectory prediction results to obtain the predicted landing point of the aircraft. Subsequently, based on the landing point prediction results, determine the target threatened by the aircraft. Finally, use the value of the threatened target to calculate the threat value of the aircraft's threatening intent.

[0017] Step 5: Calculate the distance threat value of the aircraft. Using the real-time measured aircraft-target distance d(t) of the target with the highest aircraft distance value, calculate the distance threat value of the aircraft.

[0018] Step Six: Calculation of the Threat Value of Aircraft Radar Factors. In the terminal phase of flight, the active radar of the aircraft begins to operate. The operating status of the radar directly affects the threat level of the aircraft. Therefore, this invention considers radar factors in the terminal aircraft threat calculation and calculates the threat value of radar factors based on the Dempster-Shafer evidence theory.

[0019] Step 7: Calculate the overall threat value of the aircraft. Based on the threat value w1(t) about the aircraft's threatening intent obtained in Step 4, the threat value w2(t) about the aircraft-target distance obtained in Step 5, and the threat value w3(t) about radar factors calculated in Step 6, and combined with the phase division function Phase(t) obtained in Step 2 based on the aircraft-target distance, the overall threat value W(t) of the aircraft is calculated.

[0020] Compared with the prior art, the significant advantages of this invention are:

[0021] (1) The new aircraft threat assessment method based on multi-dimensional situational fusion is more in line with actual combat conditions. The aircraft flight process is segmented according to the actual combat situation, and the changes of threat factors throughout the entire flight phase are comprehensively considered. Threat factors such as aircraft-target distance, aircraft threat intent, and radar factors are comprehensively considered, and the threat factors and weights of each phase are divided and allocated in detail, making the threat calculation more in line with the actual scenario.

[0022] (2) More accurate threat assessment of aircraft in complex electromagnetic environments. In threat assessment, the working principle of the main radar is combined, and radar factors are introduced as the basis for threat assessment in the terminal flight phase of the aircraft. The threat level of radar factors is calculated based on the Dempster-Shafer evidence theory.

[0023] (3) Introducing the aircraft threat intent factor into threat assessment makes it more suitable for aircraft threat assessment in multi-target scenarios. This method is based on the LSTM method to predict the aircraft trajectory, determines the aircraft landing point based on the trajectory prediction, determines the aircraft threat intent in real time based on the landing point result, and determines the threat value of the aircraft threat intent based on the value of the threatened target. Attached Figure Description

[0024] Figure 1 This is a flowchart of the aircraft trajectory prediction process of the present invention.

[0025] Figure 2 This is a schematic diagram of the threat factor corresponding to the stage of the present invention. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0027] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indication will also change accordingly.

[0028] The technical solutions of the various embodiments of the present invention can be combined with each other, but only if they can be implemented by those skilled in the art. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0029] The following section will further introduce the specific implementation method, as well as the technical difficulties and inventive points of this invention, using this design example as an example.

[0030] A method for aircraft threat assessment based on multi-dimensional situational awareness fusion in complex electromagnetic environments, comprising the following steps:

[0031] Step 1: When the target formation detects an unknown aircraft entering the measurement range, the LSTM method is used to estimate and predict the aircraft. The LSTM neural network is trained using parameters such as the center of mass position, center of mass velocity, aircraft-target distance, attitude angles (including slip angle, pitch angle, yaw angle) and attitude angular rate in the existing flight trajectory. The relevant network can be used for real-time aircraft trajectory prediction.

[0032] S1.1. By comprehensively analyzing the situational factors contained in the complex electromagnetic environment, the parameters required for aircraft trajectory prediction are extracted. A flight trajectory dataset is constructed based on existing flight trajectory test data. The parameters include the aircraft's center of mass position, center of mass velocity, aircraft-target distance, attitude angles (including roll, pitch, and yaw angles), and attitude angular rate in the ground coordinate system. The aircraft's flight trajectory dataset is then divided into a training set and a test set, with a weighting of 9:1.

[0033] S1.2 The training set is used as the raw input for training the neural network, and the test set is used to validate the trained model. An LSTM neural network is constructed using the aircraft's training set. The forget gate, input gate, and output gate of each neuron in the LSTM network are determined, and the trained model is validated using the test set. Finally, an LSTM neural network is trained, and the network can be used to predict the flight trajectory of the aircraft.

[0034] S1.3. Utilize the aircraft trajectory prediction results, continuously update the data as the flight trajectory dataset changes, correlate and match the results with the actual environment, assess the aircraft trajectory prediction and threatened targets, and determine the aircraft's true intentions. Specific steps are shown in the example. Figure 1 As shown.

[0035] Step Two: Analyze the entire flight process of the aircraft, distinguishing it based on dynamic characteristics, and divide the entire flight process into initial, middle, and terminal phases. Simultaneously analyze the terminal flight process, dividing it into far-field, middle-field, and near-field phases based on radar operating principles and layered azimuth prediction schemes. The following phase division function is obtained:

[0036]

[0037] In equation (1), since there are multiple threatened targets in a scenario, the target with the highest value is selected as the reference object. d(t) is the distance between the aircraft and target at time t. d1 is the aircraft-target distance threshold between the initial and middle phases of the aircraft. d2 is the aircraft-target distance threshold between the middle and far phases of the aircraft. d3 is the aircraft-target distance threshold between the far and middle phases of the aircraft. d4 is the aircraft-target distance threshold between the middle and near phases of the aircraft. The values ​​of d1 to d4 can be given based on experience. Phase(t) represents the phase division function. If Phase(t) = 0, it means the aircraft is in the initial flight phase. If Phase(t) = 1, it means the aircraft is in the middle flight phase. If Phase(t) = 2, it means the aircraft is in the far phase of the final flight phase. If Phase(t) = 3, it means the aircraft is in the middle phase of the final flight phase. If Phase(t) = 4, it means the aircraft is in the near phase of the final flight phase.

[0038] Correspondingly, a flight phase selection function is constructed for further selection of phase elements. The flight phase selection function Phase_M(t) is as follows:

[0039]

[0040] The distance between the aircraft and the target can effectively determine the aircraft's flight phase and operational status. However, the threat factors used to assess the aircraft's threat level differ depending on the flight phase and operational status. The main threat factors considered include... Figure 2 As shown. Due to the limited information and long distance of the aircraft during the initial phase of flight, the aircraft is considered to pose little threat to the targets, and therefore the aircraft threat assessment during the initial phase of flight is not considered.

[0041] Step 3: Use the real-time measured distance between the aircraft and the target to determine the flight stage of the aircraft. Assume the flight stage at time t and output the Phase(t) value to select the parameter value under the corresponding stage.

[0042] S3.1 Assume that aircraft M may pose a threat to various targets in the scenario. There are three different types of targets in the scenario: large, medium and small. Large targets are those with the largest size and the highest economic value; medium targets are those with medium size and medium economic value; and small targets are those with the smallest size and the lowest economic value. According to the target value ranking (i.e., large > medium > small), we obtain target 1, target 2, ..., target i, ..., target N, where target 1 is the target with the highest value in the scenario.

[0043] S3.2. d(t) is measured at time t. Substitute d(t) into equation (1) to obtain the value of Phase(t). Determine the flight stage of the aircraft based on the value of Phase(t).

[0044] Step 4: Calculate the threat value of the aircraft's threatening intent. Based on the center-of-mass position, center-of-mass velocity, attitude angle, and attitude angular velocity values ​​of the aircraft M measured at time t, input them into the trained LSTM neural network to complete the trajectory prediction of the aircraft. Then, use the trajectory prediction results to obtain the predicted landing point of the aircraft. Subsequently, based on the landing point prediction results, determine the target threatened by the aircraft. Finally, use the value of the threatened target to calculate the threat value of the aircraft's threatening intent.

[0045] S4.1. At time t, the center of mass position, center of mass velocity, attitude angle and attitude angular velocity of the aircraft M are measured in real time. The center of mass position, center of mass velocity, attitude angle and attitude angular velocity of the aircraft M measured at time t and before are input into the LSTM network trained in step one to obtain the flight trajectory prediction result of the aircraft.

[0046] S4.2 Based on the flight trajectory prediction results, obtain the predicted landing point coordinates of the aircraft at time t.

[0047] S4.3, Let the real-time coordinates of target i be (x... i (t),0,zi (t)), the real-time landing point prediction value of aircraft M is The formula for calculating the distance between the estimated landing point of aircraft M and the real-time position of target i in the scene is as follows:

[0048]

[0049] MD i (t) represents the predicted landing point error between aircraft M and target i. Based on equation (3), input the real-time target coordinates at time t: (x1(t), 0, z1(t)), (x2(t), 0, z2(t)), ..., (x...). i (t),0,z i (t)), ... (x) N (t),0,z N (t) and the predicted landing point of the aircraft obtained in S4.2 The calculated predicted impact point errors between the estimated aircraft M and targets 1 through N are MD1(t), MD2(t), ..., MD i (t), ..., MD N (t).

[0050] The predicted impact point errors are sorted to obtain the minimum values, and the target with the smallest predicted impact point error is the primary threat to aircraft M; if

[0051] MD i (t)=min(MD1(t),MD2(t),…,MD i (t),…MD N (t)) (4)

[0052] Therefore, the primary threat target of aircraft M is target i. The threatened target vector T(t) is obtained according to equation (4).

[0053] T(t)=[t1(t) … t i (t) … t N (t)] T (5)

[0054] In the formula, t represents the threat determination value of the aircraft against targets 1 to N. i (t) is as follows:

[0055]

[0056] If t i If (t) is 1, it indicates that the aircraft mainly poses a threat to target i; if the value is 0, the threat of the aircraft to target i is not considered.

[0057] S4.4. Based primarily on target value elements and combined with experience, set up a 5×N dimensional threat value matrix M1 for aircraft threat intent. The matrix is ​​as follows:

[0058]

[0059] In the formula, the subscript 1 of M1 indicates a relation to threat intent. The elements in the first row of the matrix represent the threat intent coefficients for targets 1 to N during the mid-course flight phase; the elements in the second row represent the threat intent coefficients for targets 1 to N during the terminal far-field flight phase; the elements in the third row represent the threat intent coefficients for targets 1 to N during the terminal mid-field flight phase; and the elements in the fourth row represent the threat intent coefficients for targets 1 to N during the terminal near-field flight phase. Let n 3i For example, this element represents the threat value of the aircraft M regarding its threat intent when the aircraft is in the terminal phase of flight and the target threatened by aircraft M is target i. The threat value parameters in matrix M1 satisfy...

[0060]

[0061] Wherein, value11 is the threat value of the mid-course threat intent against large targets, value12 is the threat value of the mid-course threat intent against medium-sized targets, and value13 is the threat value of the mid-course threat intent against small targets. The values ​​are determined by expert judgment, and the threat values ​​satisfy value11 > value12 > value13. Based on equation (8), the elements of the aircraft threat intent threat value matrix during the mid-course process are set as shown in equation (7), that is, the first row of the M1 matrix. Similarly:

[0062]

[0063] Wherein, value21 is the threat intention threat value against large targets in the terminal far-field region, value22 is the threat intention threat value against medium-sized targets in the terminal far-field region, and value23 is the threat intention threat value against small targets in the terminal far-field region. The values ​​are determined based on expert judgment, and the threat values ​​satisfy value21 > value22 > value23. Based on equation (9), the elements of the aircraft threat intention threat value matrix in the terminal far-field process shown in equation (7) are set as the second row elements of matrix M1. Similarly:

[0064]

[0065] Wherein, value31 is the threat value of the intent to threaten large targets in the mid-terminal zone, value32 is the threat value of the intent to threaten medium targets in the mid-terminal zone, and value33 is the threat value of the intent to threaten small targets in the mid-terminal zone. The values ​​are given based on expert judgment, and the threat values ​​satisfy value31 > value32 > value33. Based on equation (10), the elements of the threat value matrix for aircraft in the mid-terminal zone process shown in equation (7) are set, that is, the third row of the M1 matrix.

[0066]

[0067] Wherein, value41 is the threat value of the intent to threaten large targets in the terminal near-field region, value42 is the threat value of the intent to threaten medium targets in the terminal near-field region, and value43 is the threat value of the intent to threaten small targets in the terminal near-field region. The values ​​are given based on expert judgment, and the threat values ​​satisfy value41 > value42 > value43. Based on equation (11), the elements of the threat value matrix for aircraft intent during the terminal near-field process are set as shown in equation (7), that is, the fourth row of the M1 matrix.

[0068] S4.5. Based on the target threatened by the aircraft, the value of the threatened target, and the flight phase, determine the threat value of the aircraft's threatening intent. That is, based on the threat intent judgment vector T(t) shown in equation (5), the threat intent threat value matrix M1 shown in equation (7), and the flight phase selection function shown in equation (2) in step two, calculate the threat value formula of the aircraft's threatening intent.

[0069]

[0070] In the formula, w1(t) is the threat value of the aircraft's threatening intent at time t, and the subscript 1 indicates that it is related to the aircraft's threatening intent.

[0071] Step 5: Calculate the distance threat value of the aircraft. Using the real-time measured aircraft-target distance d(t) of the target with the highest distance value, calculate the distance threat value of the aircraft as follows:

[0072] Two thresholds are set for the aircraft-target distance. When the distance is greater than the first threshold, a constant weight is used, and the threat level of the aircraft-target distance is low. When the distance is less than the first threshold but greater than the second threshold, a linear function is used to ensure that the distance factor increases continuously as the distance decreases. When the distance is less than the second threshold, a larger constant weight is used to highlight the influence of the distance factor. The distance d(t) between the aircraft and target at time t is known. The following formula is used to calculate the distance threat level of the aircraft at each time point.

[0073]

[0074] In the formula, w2(t) is the distance threat value of the aircraft. The subscript 2 indicates that this value is related to the distance factor between the aircraft and the target. Since the calculation formula for the distance threat value is the same in the middle and end sections, no subscript is added to distinguish them. D1 and D2 are the number of intervals for the distance threat value. The number of intervals for the distance threat value can be determined based on experience to ensure that it is more in line with the actual engineering environment.

[0075] Step Six: Calculation of the Threat Value of Aircraft Radar Factors. In the terminal phase of flight, the active radar of the aircraft begins to operate. The operating status of the radar directly affects the threat level of the aircraft. Therefore, this invention considers radar factors in the terminal aircraft threat calculation and calculates the threat value of radar factors based on the Dempster-Shafer evidence theory.

[0076] In scenario S6.1, each target can measure the active radar signal parameters of aircraft M, thereby obtaining real-time radar signal parameters: pulse width τ(t), instantaneous bandwidth B(t), and pulse repetition frequency f(t). These three parameters are used as threat factors for radar threat assessment. For each threat factor, five levels of membership functions are established: high (H), medium-high (MH), medium (M), medium-low (ML), and low (L). Based on the trapezoidal membership function method, five levels of membership functions are established for the three threat factors: pulse width, instantaneous bandwidth, and pulse repetition frequency, forming the following table.

[0077] Table 1 Membership Functions of Threat Factors at Different Levels

[0078]

[0079] The membership functions m1(H), m1(MH), ..., m2(M), ..., m3(L) in the table are trapezoidal membership functions. Specifically, m1(H), ..., m1(L) are the membership functions of pulse width τ(t) at the five levels from H to L; m2(H), ..., m2(L) are the membership functions of instantaneous bandwidth B(t) at the five levels from H to L; and m3(H), ..., m3(L) are the membership functions of pulse repetition frequency f(t) at the five levels from H to L. The trapezoidal membership function values ​​in Table 1 are calculated using the pulse width τ(t), instantaneous bandwidth B(t), and pulse repetition frequency f(t) obtained from real-time measurements.

[0080] S6.2. Using the membership functions m1(H), m1(MH), ..., m2(M), ..., m3(L) obtained in S6.1, calculate the Dunn entropy for the three threat factors: pulse width τ(t), instantaneous bandwidth B(t), and pulse repetition frequency f(t).

[0081] The Deng entropy formula is as follows:

[0082]

[0083]

[0084]

[0085] In the formula E d (τ), E d (B), E d (f) represents the Dunn entropy with respect to the pulse width τ(t), the instantaneous bandwidth B(t), and the pulse repetition frequency f(t), respectively.

[0086] S6.3. Using the membership functions m1(H), m1(MH), ..., m2(M), ..., m3(L) obtained in S6.1, calculate the correlation coefficient between the threat factor pulse width τ(t), instantaneous bandwidth B(t), and pulse repetition frequency f(t). The formula is as follows:

[0087]

[0088]

[0089]

[0090] In the formula, c(τ,B) is the correlation coefficient between pulse width and instantaneous bandwidth; c(τ,f) is the correlation coefficient between pulse width and pulse repetition frequency; and c(B,f) is the correlation coefficient between instantaneous bandwidth and pulse repetition frequency.

[0091] S6.4 Calculate the non-normalized weighting coefficients for pulse width τ(t), instantaneous bandwidth B(t), and pulse repetition frequency f(t), using the following formula:

[0092]

[0093]

[0094]

[0095] In the formula, w(τ), w(B), and w(f) are the nonnormalized weighting coefficients for pulse width τ(t), instantaneous bandwidth B(t), and pulse repetition frequency f(t), respectively.

[0096] S6.5 Calculate the normalized weighting coefficients for pulse width τ(t), instantaneous bandwidth B(t), and pulse repetition frequency f(t), using the following formula:

[0097]

[0098]

[0099]

[0100] In the formula w normal (τ), w normal (B), w normal (f) are the normalized weighting coefficients for pulse width τ(t), instantaneous bandwidth B(t), and pulse repetition frequency f(t), respectively.

[0101] S6.6 Calculate the average membership degree corresponding to the five threat levels H, MH, M, ML, and L, using the following formula:

[0102] T(H) = w normal (τ)·m1(H)+w normal (B)·m2(H)+w normal (f)·m3(H) (26)

[0103] T(MH)=w normal (τ)·m1(MH)+w normal (B)·m2(MH)+w normal (f)·m3(MH) (27)

[0104] T(M) = w normal (τ)·m1(M)+w normal (B)·m2(M)+w normal (f)·m3(M) (28)

[0105] T(ML) = w normal (τ)·m1(ML)+w normal (B)·m2(ML)+w normal (f)·m3(ML) (29)

[0106] T(L)=w normal (τ)·m1(L)+w normal (B)·m2(L)+w normal (f)·m3(L) (30)

[0107] In the formula, T(H), T(MH), T(M), T(ML), and T(L) are the average membership values ​​for the five threat levels H, MH, M, ML, and L, respectively.

[0108] S6.7. Data fusion is performed on the membership functions of the five threat levels: H, MH, M, ML, and L.

[0109]

[0110]

[0111]

[0112]

[0113]

[0114] In the formula, m(H), m(MH), m(M), m(ML), and m(L) are the membership functions of the five threat levels H, MH, M, ML, and L after fusion.

[0115] S6.8. Threat coefficients k(H), k(MH), k(M), k(ML), and k(L) are given for the five threat levels H, MH, M, ML, and L. Threat coefficients k(H), k(MH), k(M), k(ML), and k(L) can be given by experience.

[0116] S6.9 Calculate the real-time threat value w3(t) for radar factors, using the following formula:

[0117]

[0118] Step 7: Calculate the overall threat value of the aircraft. Based on the threat value w1(t) about the aircraft's threatening intent obtained in Step 4, the threat value w2(t) about the aircraft-target distance obtained in Step 5, and the threat value w3(t) about radar factors calculated in Step 6, and combined with the phase division function Phase(t) obtained in Step 2 based on the aircraft-target distance, the overall threat value W(t) of the aircraft is calculated.

[0119] Based on the calculated threat levels w1(t), range threat level w2(t), and radar factor threat level w3(t) at time t, and combined with the phase division function Phase(t), the comprehensive threat level of aircraft M is calculated. The specific calculation formula is as follows:

[0120]

[0121] In the formula, k 11 k is the weighting coefficient for the threat value of the aircraft's threatening intent during the mid-course flight phase. 12This represents the weighting coefficient for range threat values ​​during the mid-course flight phase. Since the aircraft radar is not activated during mid-course flight, the threat impact of radar factors is not considered. 21 k is the weighting coefficient for the threat value of the aircraft's threat intent during the terminal long-range flight phase. 22 k is the weighting coefficient for the distance threat value of the aircraft during the terminal long-range flight phase. 23 k represents the weighting coefficient for the radar factor threat value of the aircraft during the terminal long-range flight phase. 31 k is the weighting coefficient for the threat value of the aircraft's threatening intent during the mid-term flight phase of the terminal phase. 32 k is the weighting coefficient for the aircraft's distance threat value during the mid-term flight phase. 33 k represents the weighting coefficient for the radar factor threat value of the aircraft during the mid-term flight phase of the terminal phase. 41 k is the weighting coefficient for the threat value of the aircraft's threat intent during the terminal phase of near-field flight. 42 k is the weighting coefficient for the aircraft's range threat value during the terminal phase of near-field flight. 43 The correlation coefficient is the weighting coefficient for the radar threat value of the aircraft during the terminal phase of near-field flight; the value of the correlation coefficient can be given empirically and can be adjusted in real time to adapt to actual engineering needs.

Claims

1. A method for aircraft threat assessment based on multi-dimensional situational fusion in complex electromagnetic environments, characterized in that, The steps are as follows: Step 1: When the target formation detects an unknown aircraft entering the measurement range, the LSTM method is used to estimate and predict the aircraft. The LSTM neural network is trained using the center of mass position, center of mass velocity, aircraft-target distance, attitude angles and attitude angular rate in the flight trajectory. The relevant network is used for real-time aircraft trajectory prediction, where the attitude angles include slip angle, pitch angle and yaw angle. Step Two: Analyze the entire flight process of the aircraft, distinguishing it based on dynamic characteristics, and divide the entire flight process into initial, middle, and terminal phases; simultaneously analyze the terminal phase flight process, dividing it into far-field, middle-field, and near-field phases based on radar working principles and layered azimuth prediction schemes; obtain the following phase division function: (1)， In equation (1), since there are multiple threatened targets in a scenario, the target with the highest value is selected as the reference object. for The distance between the spacecraft and target at any given moment. The threshold for the aircraft-target distance between the initial and mid-stages of the aircraft. The threshold for the aircraft-target distance between the mid-course and terminal phases of the aircraft. The threshold for the aircraft-target distance between the far-field and mid-field regions of the terminal phase. The phase threshold is defined as the vehicle-target distance between the mid-phase and near-phase regions of the terminal phase, where Phase(t) represents the phase division function; if This indicates that the aircraft is in the initial flight phase. This indicates that the aircraft is in the mid-flight phase. This indicates that the aircraft is in the terminal, far-field phase of flight. This indicates that the aircraft is in the mid-terminal flight phase. This indicates that the aircraft is in the terminal near-field flight phase. Correspondingly, a flight phase selection function is constructed for further selection of phase elements. The flight phase selection function Phase_M(t) is as follows: (2)， The distance between the aircraft and the target can effectively determine the flight stage and operational status of the aircraft. However, the threat factors for assessing the threat level of the aircraft differ depending on the flight stage and operational status. Step 3: Determine the flight stage of the aircraft using the real-time measured distance between the aircraft and the target. Assume... The current flight phase and output. The value is selected based on the parameter value for the corresponding stage. Step 4: Calculate the threat value regarding the aircraft's threatening intent, based on... The center of mass position, center of mass velocity, attitude angle and attitude angular velocity values ​​of the aircraft M measured at the time and before are input into the trained LSTM neural network to complete the trajectory prediction of the aircraft. The trajectory prediction results are then used to obtain the predicted landing point of the aircraft. Subsequently, based on the landing point prediction results, the target threatened by the aircraft is determined, and the threat value of the aircraft's threat intent is calculated. Step 5: Calculate the distance threat value of the aircraft, using the aircraft-target distance of the target with the highest real-time measured distance value. Calculate the distance threat value for the aircraft; Step Six: Calculation of aircraft radar factor threat value. In the terminal phase of flight, the aircraft's active radar begins to operate. The radar's operating status directly affects the threat level of the aircraft. The radar factor is considered in the terminal aircraft threat calculation, and the threat value of the radar factor is calculated based on the Dempster-Shafer evidence theory. Step 7: Calculate the overall threat value of the aircraft, based on the threat value regarding the aircraft's threatening intent obtained in Step 4. The threat value regarding the aircraft-target distance obtained in step five. And the threat value related to radar factors calculated in step six. And combined with the stage division function obtained in step two based on the aircraft-target distance. The overall threat level of the aircraft was calculated comprehensively. .

2. The aircraft threat assessment method based on multi-dimensional situational fusion in complex electromagnetic environments according to claim 1, characterized in that, Step one, as follows: S1.1 By comprehensively analyzing the situational factors contained in the complex electromagnetic environment, the parameters required for aircraft trajectory prediction are extracted; and a flight trajectory dataset is constructed based on existing flight trajectory test data. The parameters include the aircraft's center of mass position, center of mass velocity, aircraft-target distance, attitude angle, and attitude angular rate in the ground coordinate system; the aircraft's flight trajectory dataset is divided into two parts: a training set and a test set, with a ratio of 9:

1. S1.2 The training set is used as the original input for training the neural network, and the test set is used to verify the trained model. The training set of the aircraft is used to train and construct an LSTM neural network, determine the forget gate, input gate and output gate of each neuron of the LSTM network, and use the test set to verify the trained model. Finally, the LSTM neural network is trained and the relevant network is used to predict the flight trajectory of the aircraft. S1.

3. Utilize the aircraft trajectory prediction results, continuously update the data as the flight trajectory dataset changes, correlate and match the results with the actual environment, judge the aircraft trajectory prediction and threatened targets, and derive the aircraft's true intention.

3. The aircraft threat assessment method based on multi-dimensional situational fusion in complex electromagnetic environments according to claim 1, characterized in that, Step three, as follows: S3.1 Assume that aircraft M may pose a threat to various targets in the scenario. There are three different types of targets in the scenario: large, medium and small. Targets are sorted by value, i.e., large > medium > small, to obtain target one, target two, ..., target i, ..., target N, where target one is the target with the highest value in the scenario. S3.2, in Time measurement ,Will Substituting into equation (1) yields The value, based on The value determines the flight stage of the aircraft.

4. The aircraft threat assessment method based on multi-dimensional situational fusion in complex electromagnetic environments according to claim 1, characterized in that, Step four, as follows: S4.1, in The position, velocity, attitude angles, and angular velocity of the center of mass of the aircraft M are measured in real time. The center of mass position, center of mass velocity, attitude angle, and attitude angular velocity of the aircraft M measured at or before time are input into the LSTM network trained in step one to obtain the flight trajectory prediction result of the aircraft. S4.2 Based on the flight trajectory prediction results, the following is obtained: Predicted landing coordinates of the spacecraft at any given time ; S4.3, Let the real-time coordinates of target i be... The real-time landing point prediction value of aircraft M is The formula for calculating the distance between the estimated landing point of aircraft M and the real-time position of target i in the scene is: (3)， The expected landing point error between aircraft M and target i; input according to equation (3) Real-time target coordinates , ... ... And the predicted landing point of the aircraft obtained in S4.2 The calculated error of the predicted landing point between the estimated aircraft M and targets 1 through N is: , ... ... ; The predicted impact point errors are sorted to obtain the minimum values, and the target with the smallest predicted impact point error is the primary threat to aircraft M; if (4)， Therefore, the main threat target of aircraft M is target i; according to equation (4), the threatened target vector is obtained. , (5)， In the formula, the threat determination value of the aircraft to targets 1 to N is... as follows: (6)， if A value of 1 indicates that the aircraft mainly poses a threat to target i, while a value of 0 indicates that the aircraft's threat to target i is not considered. S4.

4. Focus on target value elements and combine them with experience to set up Dimensional threat value matrix regarding aircraft threat intent The matrix is (7)， In the formula Subscript 1 indicates that it is related to threat intent. The elements in the first row of the matrix represent the threat intent threat coefficient of targets 1 to N during the mid-flight phase. The elements in the second row of the matrix represent the threat intent threat coefficient of targets 1 to N during the far-field flight phase. The elements in the third row of the matrix represent the threat intent threat coefficient of targets 1 to N during the mid-field flight phase. The elements in the fourth row of the matrix represent the threat intent threat coefficient of targets 1 to N during the near-field flight phase. When the aircraft is in the mid-terminal flight phase, and the target threatened by aircraft M is target i, the threat value regarding the aircraft's threat intent is represented by the matrix. The medium threat value parameter satisfies (8)， in, The threat value represents the mid-range threat intent towards large targets. The threat value represents the threat intent towards medium-sized targets in the mid-range. This represents the threat value for medium-sized targets, determined by expert judgment. The threat value must meet certain criteria. ; Based on equation (8), the elements of the threat value matrix regarding the aircraft's threat intent during the mid-course process are set as shown in equation (7), i.e. The first row of the matrix; Similarly: (9)， in, For the far region of the last segment Threat Intent Threat Value For the centering of the far zone in the last segment Threat Intent Threat Value For the last segment of the far region to the small The threat intent and threat value are determined by expert judgment. The threat value must meet certain conditions. ; Based on equation (9), the elements of the threat value matrix regarding the aircraft's threat intent during the final far-field process are set as shown in equation (7), i.e. The second row of the matrix; Similarly: (10)， in, For the middle section of the final segment, the large Threat Intent Threat Value For the centering of the middle section of the last segment Threat Intent Threat Value For the small section in the middle of the last segment The threat intent and threat value are determined by expert judgment. The threat value must meet certain conditions. ; Based on equation (10), the elements of the threat value matrix regarding the aircraft's threat intent during the final mid-range process are set as shown in equation (7), i.e. The elements in the third row of the matrix; Similarly: (11)， in, For the last segment near region of the large Threat Intent Threat Value For the centering of the near zone of the last segment Threat Intent Threat Value For the small near-field region of the last segment The threat intent and threat value are determined by expert judgment. The threat value must meet certain conditions. Based on equation (11), the threat value matrix elements related to the aircraft's threat intent during the final near-field process are set as shown in equation (7), i.e. The fourth row element of the matrix; S4.

5. Determine the threat value of the aircraft's threatening intent based on the target threatened by the aircraft, the value of the threatened target, and the flight phase, i.e., based on the threat intent judgment vector formula shown in equation (5). The threat intent threat value matrix shown in equation (7) The threat value formula for the aircraft's threat intent is obtained by comprehensively calculating the flight phase selection function shown in equation (2) of step two: (12)， In the formula for The threat value related to the aircraft's threatening intent at any given time; the subscript 1 indicates that it is related to the aircraft's threatening intent.

5. The aircraft threat assessment method based on multi-dimensional situational fusion in complex electromagnetic environments according to claim 1, characterized in that, Step 5: Calculate the distance threat level of the aircraft at each instant using the following formula: (13)， In the formula, This is the distance threat value of the aircraft. The subscript 2 indicates that the value is related to the distance factor between the aircraft and the target. Since the calculation formula for the distance threat value is the same in the middle and the end, no subscript is added to distinguish it. and The number of intervals between distance and threat values ​​is determined based on experience to ensure that it better reflects the actual engineering environment.

6. The aircraft threat assessment method based on multi-dimensional situational fusion in complex electromagnetic environments according to claim 1, characterized in that, Step six, as follows: In scenario S6.1, each target measures the active radar signal parameters of aircraft M to obtain real-time radar signal parameters: pulse width. Instantaneous bandwidth and pulse repetition frequency These three parameters are used as threat factors for radar threat assessment. For each threat factor, a membership function with five levels (high, medium-high, medium, medium-low, and low) is established. Based on the trapezoidal membership function method, a membership function with five levels (pulse width, instantaneous bandwidth, and pulse repetition frequency) is established for the three threat factors. S6.2 Using the membership function obtained in S6.1 , ... ... Calculate pulse width Instantaneous bandwidth and pulse repetition frequency The Deng entropy of the three threat factors; the Deng entropy formula is as follows: (14)， (15)， (16)， In the formula , , Regarding pulse width Instantaneous bandwidth and pulse repetition frequency Deng's entropy; ... pulse width Membership functions at the five levels from H to L; ... Instantaneous bandwidth Membership functions at the five levels from H to L; ... pulse repetition frequency Membership functions at the five levels from H to L; pulse width obtained through real-time measurement. Instantaneous bandwidth and pulse repetition frequency The value of the trapezoidal membership function is calculated from the value of the trapezoidal membership function. S6.3 Using the membership function obtained in S6.1 , ... ... Calculate the pulse width of threat factors Instantaneous bandwidth and pulse repetition frequency The correlation coefficient between them is calculated using the following formula: (17)， (18)， (19)， In the formula, This is the correlation coefficient between pulse width and instantaneous bandwidth; This is the correlation coefficient between pulse width and pulse repetition frequency; This is the correlation coefficient between instantaneous bandwidth and pulse repetition frequency; S6.4 Calculate the pulse width Instantaneous bandwidth and pulse repetition frequency The non-normalized weighting coefficients are calculated using the following formula: (20)， (21)， (22)， In the formula , , Regarding pulse width Instantaneous bandwidth and pulse repetition frequency The non-normalized weighting coefficients; S6.5 Calculate the pulse width Instantaneous bandwidth and pulse repetition frequency The normalized weighting coefficients are calculated using the following formula: (23)， (24)， (25)， In the formula , , Regarding pulse width Instantaneous bandwidth and pulse repetition frequency Normalized weighting coefficients; S6.6 Calculate the average membership degree corresponding to the five threat levels H, MH, M, ML, and L, using the following formula: (26)， (27)， (28)， (29)， (30)， In the formula, , , , , These are the average membership values ​​for the five threat levels: H, MH, M, ML, and L. S6.

7. Data fusion is performed on the membership functions of the five threat levels: H, MH, M, ML, and L. (31)， (32)， (33)， (34)， (35)， In the formula , , , , The membership functions for the five threat levels H, MH, M, ML, and L after fusion; S6.8 provides threat level coefficients corresponding to the five threat levels: H, MH, M, ML, and L. , , , , Threat level coefficient , , , , Based on experience; S6.9 Calculate the real-time threat value for radar factors. The formula is as follows: (36)。 7. The aircraft threat assessment method based on multi-dimensional situational fusion in complex electromagnetic environments according to claim 1, characterized in that, Step seven, as follows: Based on calculations Threat level of aircraft threat intent at any given moment Distance Threat Value and radar factor threat value And combined with the stage division function Calculate the overall threat level of aircraft M using the following formula: (37)， In the formula, This refers to the weighting coefficients for the threat value of an aircraft's threatening intent during the mid-course flight phase. This is the weighting coefficient for the distance threat value during the mid-course flight phase. Since the aircraft radar is not turned on during the mid-course flight, the threat impact of radar factors is not considered. This refers to the weighting coefficients for the threat value of the aircraft's threat intent during the terminal phase of long-range flight. This refers to the weighting coefficients for the aircraft's distance threat value during the terminal phase of long-range flight. The weighting coefficients for the radar factor threat value of the aircraft during the terminal long-range flight phase; This refers to the weighting coefficients for the threat value of the aircraft's threatening intent during the mid-term flight phase. This refers to the weighting coefficients for the aircraft's distance threat value during the mid-term flight phase. The weighting coefficients for the radar factor threat value of the aircraft during the terminal mid-zone flight phase; This refers to the weighting coefficients for the threat value of an aircraft's threatening intent during the terminal phase of near-field flight. This refers to the weighting coefficients for the aircraft's distance threat value during the terminal phase of near-field flight. The correlation coefficient is the weighting coefficient for the radar factor threat value of the aircraft during the terminal phase of near-field flight; the value of the correlation coefficient is given empirically and adjusted in real time to adapt to actual engineering needs.

Images