Resource management method based on target real-time threat degree and tracking accuracy
By introducing two-dimensional optimization of target threat level and tracking accuracy into radar resource management, and using the entropy weight method and the optimal linear unbiased filter algorithm to optimize radar beam illumination and dwell time, the problem of unreasonable resource allocation in the existing technology is solved, and effective tracking and accurate allocation of high-threat targets are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHANGHAI SPACEFLIGHT ELECTRONICS & COMM EQUIP RES INST
- Filing Date
- 2023-07-24
- Publication Date
- 2026-07-31
AI Technical Summary
In existing radar resource scheduling, target threat assessment is not fully considered, resulting in unreasonable resource allocation and difficulty in effectively prioritizing the tracking of high-threat or interested targets.
The target threat level and tracking accuracy are used as two-dimensional optimization requirements. The entropy weight method is used to quantify the threat level evaluation index. Combined with the optimal linear unbiased filter algorithm, the allocation of radar beam illumination targets and dwell time is optimized.
It improves the targeting and rationality of radar resource allocation, ensures focus on high-threat targets, and enhances tracking accuracy and real-time performance.
Smart Images
Figure CN116973907B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of radar resource management technology, and in particular to a resource management method and apparatus based on real-time target threat level and tracking accuracy. Background Technology
[0002] In practical radar resource scheduling, target threat assessment is an indispensable and crucial component. Reliable target threat assessment results can effectively assist in the allocation of radar operational resources. Real-time assessment of the threat levels of multiple targets in the tracking environment allows radar to prioritize tracking of high-threat targets or targets of greater interest, which is of great significance in actual combat. Research addressing this issue primarily utilizes existing information to analyze, prioritize, evaluate, and optimize various threat factors, ultimately deriving a more realistic resource scheduling scheme. However, research that incorporates threat level into its considerations is rare.
[0003] Threat indicators come in various types. For example, "target distance indicators" are quantitative, providing a specific numerical value that can be quantified and normalized according to actual circumstances. "Whether there is a weapon threat" is a qualitative indicator, yielding results of yes, no, or uncertain. Forcing quantification of this would be too difficult, making it challenging to standardize the quantification criteria and increasing the complexity of comprehensively assessing the target's threat level. To address this, most research on target threat levels employs fuzzy processing or normalization. Before comprehensive evaluation, the original indicators of the target are normalized, transforming them into dimensionless data ranging from 0 to 1.
[0004] There are various methods for assessing target threat levels, and many mathematical theories, such as decision theory, fuzzy logic, and grey relational analysis, have been applied to this research. Because different indicators have varying degrees of influence on threat levels, it is necessary to assign appropriate weights to these indicators during the threat assessment process. Commonly used weighting methods include the entropy information method and the maximum interest rate spread method.
[0005] The addition of these algorithms undoubtedly presents new challenges to radar resources. However, in terms of radar resource scheduling, there is a lack of research that takes threat level into account. Summary of the Invention
[0006] To address the aforementioned problems, the present invention aims to provide a resource management method and apparatus based on real-time target threat level and tracking accuracy. This method and apparatus employs target threat level and target tracking accuracy as two-dimensional optimization requirements, uses the entropy weight method to quantify and integrate key indicators of threat level evaluation under different dimensions, couples the obtained quantified threat level with target tracking accuracy as the optimization objective function for resource allocation, and achieves the allocation of radar beam illumination targets and dwell time through optimization algorithms.
[0007] To achieve the above objectives, the technical solution of the present invention is as follows:
[0008] A resource management method based on real-time target threat level and tracking accuracy includes the following steps:
[0009] Step 1: Obtain target threat level assessment indicators;
[0010] Step 2: Construct threat degree membership attribute vectors and matrices for each of the N targets, and use the entropy weight method to evaluate the comprehensive threat degree weight ξ of the targets;
[0011] Step 3: At time k, perform a filtered estimation of the target based on the optimal linear unbiased filter (BLUE) to calculate the unbiased estimator of the target.
[0012] Step 4: Based on the unbiased estimator of the target obtained in Step 3 Calculate the posterior Cramer-Rao bound C of the variance of the tracking error for each target. PCRLB (x k );
[0013] Step 5: Establish a radar resource management model for multi-target scenarios. The model includes resource allocation in two aspects: beam selection and dwell time allocation.
[0014] Step 6: Based on the threat level of each target in Step 2, select the target to be illuminated by the beam at time k;
[0015] Step 7: For the targets that need beam illumination identified in Step 6, use the posterior Cramer-Rao bound C of the target tracking error variance from Step 4. PCRLB (x k As a cost function, it measures the target tracking accuracy and calculates the minimum dwell time resources required to meet the tracking accuracy.
[0016] Step 8: Update and predict the status of all targets;
[0017] Step 9: Repeat steps 1 to 8 to complete the entire tracking process.
[0018] Preferably, in step 1, the target threat level evaluation index is determined as follows:
[0019] By using the membership method, the original indicators of the target are normalized and transformed into dimensionless data ranging from 0 to 1. The attribute values of each target threat indicator are then plotted. The threat membership function for the target speed indicator is:
[0020] Where, γ v = -0.55 represents the speed threat attenuation index, αv =0.2 indicates the minimum speed threat level, where speed v is in m / s and the threat threshold v0 = 100;
[0021] The threat membership function of the target distance index is:
[0022] Among them, the distance threat attenuation index γ d = -0.16, minimum speed threat level α d =0.2, the unit of speed d is km, and the threat threshold d0 = 2;
[0023] The threat membership function for the target type indicator is:
[0024]
[0025] Preferably, in step 2, threat degree membership attribute vectors and matrices are constructed for each of the N targets, and the comprehensive weight ξ of the target threat degree is evaluated using the entropy weight method as follows:
[0026] Step 201: For N targets and P indicators, construct the threat level membership attribute vector as follows:
[0027]
[0028] The target threat index membership matrix A is:
[0029] A = [a1, a2, ..., a n ,…,a N ], n=1,…,N;
[0030] Step 202: Standardize each element p in the matrix P ij The standardization method is as follows:
[0031]
[0032] The information entropy of each indicator is calculated as follows:
[0033]
[0034] Step 203: Calculate the weights of each indicator as follows:
[0035]
[0036] The overall weight of the target threat level obtained is:
[0037] ξ j =W*A T .
[0038] Preferably, in step 3, the target is estimated using a filtered method based on the optimal linear unbiased filter (BLUE), as follows:
[0039] Step 301: Calculate the one-step prediction value of the target state. in It is the target state estimate at time k-1;
[0040] Step 302: One-step prediction of the target covariance matrix Where P k-1 It is the target estimation covariance matrix at time k-1;
[0041] Step 303: One-step prediction of the observations is in It is a prediction of the observation value at time k;
[0042] Step 304: Calculate measurement conversion error
[0043] Step 305: Calculate the filter gain factor at time k. Where S k yes The covariance matrix;
[0044] Step 306: Correct the predicted values to obtain the state estimate.
[0045] Preferably, in step 4, based on the unbiased estimator of the target calculated in step 3, the posterior Cramer-Rao bound C of the variance of the tracking error for each target is calculated. PCRLB (x k The method is as follows:
[0046] BIM calculations involve numerous matrix operations, represented as the target prior Fisher information matrix J. p (x k ) and Fisher information matrix J D (x k The sum of the two parts can be expressed as:
[0047] J(x k ) = J p (x k )+J D (x k )
[0048] Step 401: Calculate the target prior Fisher information matrix:
[0049]
[0050] Where F represents the target state transition matrix, and Q represents the variance;
[0051] Step 402: Calculate the Fisher information matrix:
[0052]
[0053] in, For the error covariance, G(x) k Let be a Jacobian matrix, satisfying
[0054]
[0055] Preferably, in step 5, the method for establishing a radar resource management model under multi-target conditions is as follows:
[0056] Step 501: Define indicators to reflect target tracking accuracy.
[0057] Step 502: To balance target threat level and tracking accuracy, define evaluation metrics.
[0058] This indicator ensures that when different targets have the same tracking performance, the radar will prioritize allocating resources to targets with a higher threat level; when the target threat levels are the same, the radar will prioritize allocating resources to targets with less satisfactory tracking performance. The corresponding beam pointing variable at the current moment is...
[0059] Step 503: η n ΔT represents the expected tracking accuracy of target n. n,k The dwell time of the target when it is illuminated should be less than the upper limit T of the total time T used by the radar for tracking tasks. track , The total dwell time consumed for all targets, when A value of 1 indicates that target n is illuminated by the beam at time k. A value of 0 indicates that target n was not illuminated, and the corresponding dwell time is 0.
[0060] By constructing the constraints and objective function in the optimized mathematical model, we can obtain the mathematical expression of the new optimization problem as follows:
[0061] Preferably, in step 6, the target to be illuminated by the beam at time k is selected based on the threat level of each target in step 2. The implementation method is as follows:
[0062] For target n, calculate When b n When >0, Add the target n to the tracking list, which contains n elements; b n When ≤0,
[0063] Preferably, in step 7, the minimum dwell time resource required to meet the tracking accuracy is calculated, and the implementation method is as follows:
[0064] Step 701: Calculate the posterior Cramer-Rao bound C of the variance of the tracking error for targets entering the tracking list. PCRLB (x k The method is as follows:
[0065] If num ≥ Q, track the first Q targets in list A; if num < Q, track the first num targets and calculate the Cramer-Rao bound for target n at time k.
[0066] Step 702: The allocation of stay time is as follows:
[0067] Return satisfied Minimum dwell time t d,n Determine the beam dwell time assigned to target n at time k+1.
[0068] Preferably, in step 8, the state of all targets is updated and predicted, which is implemented as follows:
[0069] Tracking of untracked targets (excluding those already selected for tracking) is performed using an optimal linear unbiased filter. The predicted state and covariance at time k are used as estimates for time k+1.
[0070] Based on the same concept, this invention also provides an apparatus for a multi-target tracking resource management algorithm based on real-time target threat level and tracking accuracy, comprising: an identification module for determining target threat level evaluation indicators; an evaluation module for constructing threat level membership attribute vectors and matrices for N targets respectively, and evaluating the comprehensive target threat level weight ξ using the entropy weight method; and a filtering estimation module for performing filtering estimation on the target based on the optimal linear unbiased filter (BLUE) at time k, and calculating the unbiased estimate of the target. C PCRLB (x k The module is used to calculate the posterior Cramer-Rao bound C of the target tracking error variance based on the unbiased estimator obtained from the filtering estimation module. PCRLB (x k The resource management module is used to establish a radar resource management model for multi-target scenarios, mainly involving resource allocation in two aspects: beam selection and dwell time allocation. The target selection module is used to select the target to be illuminated by the beam at time k based on the threat level of each target in the assessment module. The dwell time module is used to select the target that needs beam illumination from the target selection module, utilizing C...PCRLB (x k The PCRLB of each target in the module is used as a cost function to measure the target tracking accuracy and calculate the minimum dwell time resources required to meet the tracking accuracy; the prediction module is used to update the state and predict the target for all targets.
[0071] Because the present invention adopts the above technical solution, it has the following advantages and positive effects compared with the prior art:
[0072] 1. This invention identifies the target to be illuminated by the radar beam by assessing the real-time threat level of the target, and can focus on targets with a greater threat.
[0073] 2. This invention uses the posterior Cramer-Rao bound based on the BLUE (Best Linear Unbiased Estimation) algorithm as an evaluation index, which improves both the real-time performance of the algorithm and the rationality of the allocation of radar beam dwell time.
[0074] 3. This invention first studies the target threat level index and threat level assessment, coupling tracking accuracy and threat level to obtain a new index, ensuring target tracking accuracy. Finally, a two-step decomposition algorithm is used to obtain the resource allocation scheme that best suits the tracking situation. Attached Figure Description
[0075] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings, wherein:
[0076] Figure 1 This is a flowchart of the method proposed in this invention;
[0077] Figure 2 Three simulated target motion scenarios;
[0078] Figure 3 The threat level curves for the three simulated targets are shown in the graph.
[0079] Figure 4 This is a schematic diagram of the estimated covariance curves of the three targets and the corresponding Cramer-Rao bounds obtained by the method proposed in this invention.
[0080] Figure 5 This is a schematic diagram showing the residence time distribution of three targets obtained using the method proposed in this invention;
[0081] Figure 6 The target tracking RMSE curves for the three targets obtained using the method proposed in this invention are shown. Detailed Implementation
[0082] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The advantages and features of the present invention will become clearer from the following description and claims. It should be noted that the drawings are all in a very simplified form and use non-precise ratios, and are only used to facilitate and clarify the illustration of the embodiments of the present invention.
[0083] 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.
[0084] like Figure 1 As shown, the multi-target tracking resource management algorithm based on real-time target threat level and tracking accuracy of the present invention includes the following steps: Step 1, determining the target threat level evaluation index; Step 2, constructing threat level membership attribute vectors and matrices for N targets respectively, and evaluating the comprehensive weight of target threat level using the entropy weight method; Step 3, at time k, performing filtering estimation on the target based on the optimal linear unbiased filter (BLUE), and calculating the target unbiased estimator; Step 4, based on the target unbiased estimator obtained in Step 3... Calculate the posterior Cramer-Rao bound C of the variance of the tracking error for each target. PCRLB (x k Step 5: Establish a radar resource management model for multi-target scenarios, which includes resource allocation in two aspects: beam selection and dwell time allocation; Step 6: Based on the threat level of each target in Step 2, select the target to be illuminated by the beam at time k; Step 7: For the targets that need to be illuminated by the beam in Step 6, use the posterior Cramer-Rao bound C of the target tracking error variance in Step 4. PCRLB (x k As a cost function, the target tracking accuracy is measured, and the minimum dwell time resource required to meet the tracking accuracy is calculated; Step 8, update and predict the state of all targets; Step 9, repeat steps 1 to 8 to complete the entire tracking process.
[0085] The specific explanations for each of the above steps are as follows:
[0086] Step 1: Determine the target threat level assessment indicators;
[0087] The membership method will be used to normalize the original indicators of the targets, transforming them into dimensionless data ranging from 0 to 1. This will be used to represent the attribute values of each target threat indicator. The specific representation method is as follows:
[0088] Step 101: Assume the speed threat attenuation exponent γ v = -0.55, minimum speed threat level αv =0.2, the unit of speed v is m / s, the threat threshold v0 = 100, the threat membership function representing the target speed index is:
[0089]
[0090] Step 102: Assume the distance threat attenuation exponent γ d = -0.16, minimum speed threat level α d =0.2, the unit of speed d is km, the threat threshold d0 = 2, and the threat membership function representing the target distance index is:
[0091]
[0092] Step 103: The threat membership function representing the target type indicator is:
[0093]
[0094] By defining target threat level evaluation indicators, threat level can be transformed into dimensionless data, which makes it easier to characterize the attribute values of threat level indicators.
[0095] Step 2: Construct threat degree membership attribute vectors and matrices for each of the N targets, and use the entropy weight method to evaluate the comprehensive threat degree weight ξ of the targets;
[0096] Step 201: For N targets and P indicators, construct the threat level membership attribute vector as follows:
[0097]
[0098] The target threat index membership matrix A is:
[0099] A = [a1, a2, ..., a n ,…,a N ], n=1,…,N;
[0100] Step 202: Standardize each element p in the matrix P ij The standardization method is as follows:
[0101]
[0102] The information entropy of each indicator is calculated as follows:
[0103]
[0104] Step 203: Calculate the weights of each indicator as follows:
[0105]
[0106] The overall weight of the target threat level obtained is:
[0107] ξ j =W*A T .
[0108] By constructing a threat level membership attribute vector and matrix and using the entropy weight method, the comprehensive weight coefficient of the target threat level can be obtained, which prepares for the subsequent beam target selection.
[0109] Step 3: At time k, perform a filtered estimation of the target based on the optimal linear unbiased filter (BLUE) to calculate the unbiased estimator of the target.
[0110] Step 301: Assumption This is the target state estimate at time k-1, calculating the one-step prediction of the target state.
[0111] Step 302: Assume P k-1 It is the target estimation covariance matrix at time k-1, and the one-step prediction of the target covariance matrix.
[0112] Step 303: Assumption It is a prediction of the observation at time k, and the one-step prediction of the observation is...
[0113] Step 304: Calculate measurement conversion error
[0114] Step 305: Assume S k yes The covariance matrix, the filter gain factor at time k
[0115] Step 306: Correct the predicted values to obtain the state estimate.
[0116] By filtering and estimating the target, we can obtain the corresponding unbiased estimator, which prepares us for the subsequent calculation of tracking error.
[0117] Step 4: Based on the unbiased estimator of the target obtained in Step 3, calculate the posterior Cramer-Rao bound C of the variance of the tracking error for each target. PCRLB (x k );
[0118] BIM calculations involve numerous matrix operations, which can be represented as the target prior Fisher Information Matrix (FIM). p (x k ) and Fisher information matrix JD (x k The sum of the two parts can be expressed as:
[0119] J(x k ) = J p (x k )+J D (x k )
[0120] Step 401: Assuming the target state transition matrix F and variance Q, calculate the target prior Fisher information matrix:
[0121]
[0122] Step 402: The method for calculating the Fisher information matrix is as follows:
[0123]
[0124] in, For the error covariance, G(x) k Let be a Jacobian matrix, specifically satisfying:
[0125]
[0126] Step 403: The target posterior Cramer-Rao bound (PCRLB) can be expressed as:
[0127]
[0128] The lower bound of the target tracking accuracy error variance is extracted as follows:
[0129] C PCRLB (x k ) = C n =C PCRLB (1,1)+C PCRLB (4,4).
[0130] The lower bound of the tracking error variance can be obtained by calculating the posterior Cramer-Rao bound of the target tracking error variance.
[0131] Step 5: Establish a radar resource management model for multi-target scenarios, which mainly consists of two aspects: beam selection and dwell time allocation.
[0132] Step 501: Define indicators to reflect target tracking accuracy.
[0133] Step 502: To balance target threat level and tracking accuracy, define evaluation metrics.
[0134] This indicator ensures that when different targets have the same tracking performance, the radar will prioritize allocating resources to targets with a higher threat level; similarly, when targets have the same threat level, the radar will prioritize allocating resources to targets with less effective tracking. The corresponding beam pointing variable at the current moment is...
[0135] Step 503: Assume η n ΔT represents the expected tracking accuracy of target n. n,k The dwell time of the target when it is illuminated should be less than the upper limit T of the total time T used by the radar for tracking tasks. track . The total dwell time consumed for all objectives. When A value of 1 indicates that target n is illuminated by the beam at time k. A value of 0 indicates that target n has not been illuminated, and the corresponding dwell time is 0.
[0136] By constructing the constraints and objective function in the optimized mathematical model, we can obtain the mathematical expression of the new optimization problem as follows:
[0137]
[0138]
[0139] By establishing a multi-target radar resource management model, the allocation of resources in terms of target beam selection and dwell time can be determined, thereby ensuring the rationality of radar resource allocation and tracking accuracy.
[0140] Step 6: Based on the threat level of each target in Step 2, select the target to be illuminated by the beam at time k;
[0141] For target n, calculate When b n When >0, Add the target n to the tracking list, which contains n elements; b n When ≤0,
[0142] Targets can be easily and quickly located based on their threat level.
[0143] Step 7: For the targets that need beam illumination as determined in Step 6, use the PCRLB of each target in Step 4 as the cost function to measure the target tracking accuracy and calculate the minimum dwell time resource required to meet the tracking accuracy.
[0144] Step 701: The PCRLB calculation method for targets entering the tracking list is as follows:
[0145] If num ≥ Q, track the first Q targets in list A; if num < Q, track the first num targets. Calculate the Cramer-Rao bound for target n at time k.
[0146] Step 702: The allocation of stay time is as follows:
[0147] Return satisfied Minimum dwell time t d,n Determine the beam dwell time assigned to target n at time k+1.
[0148] By calculating the posterior Cramer-Rao bound (PCRLB) for each target, not only can the accuracy of target tracking be satisfied, but also the minimum dwell time of the beam allocated to the target can be calculated.
[0149] Step 8: Update and predict the status of all targets;
[0150] The BLUE algorithm is used to track targets. For untracked targets (all targets other than those already selected for tracking), the predicted state and covariance at time k are used as the estimated value at time k+1.
[0151] By updating and predicting the target state, the accuracy and precision of target determination can be guaranteed.
[0152] Step 9: Repeat steps 1 to 8 to complete the entire tracking process.
[0153] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments. Even if various changes are made to the present invention, if these changes fall within the scope of the claims of the present invention and their equivalents, they shall still fall within the protection scope of the present invention.
Claims
1. A resource management method based on real-time target threat level and tracking accuracy, characterized in that, Includes the following steps: Step 1: Obtain target threat level assessment indicators; Step 2: Construct threat level membership attribute vectors and matrices for each of the N targets, and use the entropy weight method to evaluate the overall threat level weight of the targets. ; Step 3: At time k, perform a filtered estimation of the target based on the optimal linear unbiased filter, and calculate the unbiased estimator of the target. ; Step 4: Based on the unbiased estimator of the target obtained in Step 3 Calculate the posterior Cramer-Rao bound of the variance of the tracking error for each target. ; Step 5: Establish a radar resource management model for multi-target scenarios. The model includes resource allocation in two aspects: beam selection and dwell time allocation. Step 6: Based on the threat level of each target in Step 2, select the target to be illuminated by the beam at time k; Step 7: For the targets that need beam illumination identified in Step 6, use the posterior Cramer-Rao bound of the target tracking error variances from Step 4. As a cost function, it measures the target tracking accuracy and calculates the minimum dwell time resources required to meet the tracking accuracy. Step 8: Update and predict the status of all targets; Step 9: Repeat steps 1 to 8 to complete the entire tracking process.
2. The resource management method based on real-time target threat level and tracking accuracy according to claim 1, characterized in that, In step 1, the target threat level assessment indicators are determined as follows: By using the membership method, the original indicators of the target are normalized and transformed into dimensionless data ranging from 0 to 1. The attribute values of each target threat indicator are then plotted. The threat membership function for the target speed indicator is: ,in, Indicates the speed threat decay index. Indicates the minimum speed threat level, speed The unit is m / s, threat threshold ; The threat membership function of the target distance index is: Among them, the distance threat attenuation index Minimum speed threat level The unit of speed d is km, and the threat threshold is... ; The threat membership function for the target type indicator is: 。 3. The resource management method based on real-time target threat level and tracking accuracy according to claim 1, characterized in that, In step 2, threat level membership attribute vectors and matrices are constructed for each of the N targets, and the entropy weight method is used to evaluate the comprehensive weight of the target threat level. The method is as follows: Step 201: For N targets and P indicators, construct the threat level membership attribute vector as follows: ； Target threat index membership matrix for: ； Step 202: Standardization Matrix Each element The standardization method is as follows: ； The information entropy of each indicator is calculated as follows: ； Step 203: Calculate the weights of each indicator as follows: ， The overall weight of the target threat level obtained is: 。 4. The resource management method based on real-time target threat level and tracking accuracy according to claim 1, characterized in that, In step 3, the target is estimated using a filtered method based on the optimal linear unbiased filter. The calculation method is as follows: Step 301: Calculate the one-step prediction value of the target state. ,in yes Target state estimation at time t; Step 302: One-step prediction of the target covariance matrix ,in yes The target estimation covariance matrix at time t; Step 303: One-step prediction of the observations is ,in It is a prediction of the observation at time k. It is the target observation at time k; Step 304: Calculate measurement conversion error ; Step 305: Calculate the filter gain factor at time k. ,in yes The covariance matrix; Step 306: Correct the predicted values to obtain the state estimate. .
5. The resource management method based on real-time target threat level and tracking accuracy according to claim 1, characterized in that, In step 4, the unbiased estimator of the target calculated in step 3 is used. Calculate the posterior Cramer-Rao bound of the variance of the tracking error for each target. The method is as follows: BIM calculations involve numerous matrix operations, represented as the target prior Fisher information matrix. and Fisher information matrix The sum of the two parts can be expressed as: Step 401: Calculate the target prior Fisher information matrix: ， Wherein represents Target state transition matrix, Indicates variance; Step 402: Calculate the Fisher information matrix: ， in, For error covariance, Let the Jacobian matrix satisfy the following conditions: 。 6. The resource management method based on real-time target threat level and tracking accuracy according to claim 1, characterized in that, In step 5, the method for establishing a radar resource management model under multi-target conditions is as follows: Step 501: To reflect the target tracking accuracy, define an index tracking accuracy index. ; Step 502: To balance target threat level and tracking accuracy, define a comprehensive evaluation index. ; This indicator ensures that when different targets have the same tracking performance, the radar will prioritize allocating resources to targets with a higher threat level; when the target threat levels are the same, the radar will prioritize allocating resources to targets with less satisfactory tracking performance. The corresponding beam pointing variable at the current moment is... ; Step 503: Indicate target The expected tracking accuracy, The dwell time for a target when it is illuminated should be less than the upper limit of the total time the radar uses for tracking tasks. , The total dwell time consumed for all targets, when A value of 1 indicates that target n is illuminated by the beam at time k. A value of 0 indicates that target n was not illuminated, and the corresponding dwell time is 0. By constructing the constraints and objective function in the optimized mathematical model, we can obtain the mathematical expression of the new optimization problem as follows: .
7. The resource management method based on real-time target threat level and tracking accuracy according to claim 1, characterized in that, In step 6, based on the threat level of each target in step 2, the target to be illuminated by the beam at time k is selected. The implementation method is as follows: For the target ,calculate Add the target n to the tracking list, and the list contains n elements; hour, .
8. The resource management method based on real-time target threat level and tracking accuracy according to claim 1, characterized in that, In step 7, the minimum dwell time resource required to meet the tracking accuracy is calculated, and the implementation method is as follows: Step 701: Calculate the posterior Cramer-Rao bound of the variance of the tracking error for targets entering the tracking list. The method is as follows: like , for list Center front Track each target; if , to the front Track one target and calculate the Cramer-Rao boundary for target n at time k. ; Step 702: The allocation of stay time is as follows: Return satisfied Minimum stay time Determine the beam dwell time assigned to target n at time k+1.
9. The resource management method based on real-time target threat level and tracking accuracy according to claim 1, characterized in that, In step 8, the state of all targets is updated and predicted. The implementation method is as follows: The system tracks targets that were not selected for tracking based on the optimal linear unbiased filter, using the predicted state and covariance at time k as the estimate at time k+1. , .
10. A device for multi-target tracking resource management algorithm based on real-time target threat level and tracking accuracy, characterized in that, include: The identification module is used to determine the target threat level assessment indicators; The evaluation module is used to construct threat level membership attribute vectors and matrices for N targets, and to evaluate the overall threat level of the targets using the entropy weight method. ; The filtering estimation module is used to perform filtering estimation of the target based on the optimal linear unbiased filter (BLUE) at time k, and calculate the unbiased estimator of the target. ; This module is used to calculate the posterior Cramer-Rao bound of the target tracking error variance based on the unbiased estimator obtained from the filtering estimation module. ; The resource management module is used to establish a radar resource management model under multi-target conditions. It mainly consists of two aspects: beam selection and dwell time allocation. The target selection module is used to select the target to be illuminated by the beam at time k based on the threat level of each target in the assessment module. The dwell time module is used to identify targets that require beam illumination from the target selection module. The PCRLB of each target in the module is used as a cost function to measure the target tracking accuracy and calculate the minimum dwell time resources required to meet the tracking accuracy. The prediction module is used to update and predict the state of all targets.