A method for evaluating the boundary performance of electronic reconnaissance equipment based on big data analytics

Abstract

This invention discloses a method for evaluating the boundary performance of electronic reconnaissance equipment based on big data analytics, comprising the following steps: S1, establishing an indicator system for the electronic reconnaissance equipment system; S2, establishing a confidence database and a rule-based confidence database inference model; S3, loading the equipment model, including but not limited to equipment parameters such as transmission frequency, transmission antenna height, and transmission power, as well as environmental parameters such as temperature, water vapor density, and visibility; S4, loading existing equipment data, mainly focusing on the equipment's existing boundary performance under known parameter conditions; S5, expanding the dataset using a small sample dataset generation algorithm to predict the equipment's boundary performance under the conditions of step S3; S6, calculating the confidence level in step S5 using the inference model from step S2. This invention utilizes equipment performance data including environmental factors and applies big data synthesis and correlation analysis inference techniques to effectively evaluate the performance of electronic reconnaissance equipment.

Description

Technical Field

[0001] This invention belongs to the field of electronic reconnaissance equipment effectiveness evaluation technology, and in particular relates to a method for evaluating the boundary performance of electronic reconnaissance equipment based on big data analysis technology. Background Technology

[0002] For a long time, various types of weapons, navigation systems, and electronic information equipment in the military or civilian fields have faced the challenge of how to conduct equipment evaluation, performance assessment, and generational comparison. In particular, the boundary performance assessment of electronic reconnaissance equipment is rarely reported in public, and there is an urgent need to strengthen relevant research and exploration.

[0003] The outcome of modern warfare is undoubtedly determined by a comprehensive comparison of factors such as the technological strength of nations worldwide, the competence of their military personnel, the level of research and application of advanced equipment, the effectiveness of scientific military training, the level of daily equipment management, and intelligence gathering and processing capabilities. Among these factors, the enhancement and improvement of intelligence gathering and processing capabilities are increasingly influenced and constrained by the boundary performance of various types of electronic reconnaissance equipment. Therefore, research on boundary performance assessment technologies for electronic reconnaissance equipment, and its provision of technical support for future warfare, such as equipment deployment and joint performance analysis, has become increasingly important.

[0004] Currently, various solutions for equipment performance evaluation have the following advantages and disadvantages, as shown in Table 1.

[0005] Table 1. Advantages and disadvantages of existing solutions for equipment performance evaluation

[0006] Method Name advantage insufficient Analytic Hierarchy Process Simple to implement and easy to do The importance ranking process is largely subjective. Bayes Network The theoretical foundation is mature and it is relatively flexible. Requires prior knowledge and cannot handle qualitative knowledge. Neural Networks Possesses learning and adaptive capabilities Due to the significant impact of sample set quality, sample collection recommendations and methods are usually not integrated. DS Theory Able to meet weaker conditions There are difficulties in reliability allocation.

[0007] Currently, technologies for evaluating the performance and effectiveness of electronic reconnaissance equipment are still in their early stages of development. This is partly because current research largely relies on the experience of experts in a few fields; and partly because it typically relies on limited, incomplete, and scattered signals, information, and data to make highly hypothetical but impractical judgments or analyses. Furthermore, the lack of effective integration between the tactical and strategic needs of frontline military personnel and the comprehensive assessments of military experts across various domains, coupled with the absence of effective support from big data and artificial intelligence technologies, has prevented the formation of a standardized, unified, and reasonable evaluation system and principles, as well as the development of effective evaluation methods for discussion and clarification. Summary of the Invention

[0008] To address the aforementioned problems, the present invention aims to provide a method for evaluating the boundary performance of electronic reconnaissance equipment based on big data analytics. This method utilizes equipment performance data that includes the influence of environmental factors and applies big data integration and correlation analysis reasoning techniques to effectively evaluate the performance of electronic reconnaissance equipment.

[0009] To achieve the above-mentioned objectives, the present invention adopts the following technical solution:

[0010] A method for evaluating the boundary performance of electronic reconnaissance equipment based on big data analytics includes the following steps:

[0011] S1. Establish an indicator system for electronic reconnaissance equipment systems;

[0012] S2. Establish a confidence base and build a rule-based confidence base inference model; including the following sub-steps:

[0013] S2.1 Confidence library initialization

[0014] S2.2, Rule Vectorization: The rule is represented in the following vector form: Rk: If x is A k Then DWith belief degree β k Where x represents the input vector; A k The reference value for the k-th rule after combination (set) is represented; D represents the result vector; β k This represents the confidence level assessment of the output result; the result assessment is expressed as: S(A) k )={(D i ,β 1,k ),i=1,2...,N};

[0015] S2.3. Use the ER algorithm to fuse the rules, and evaluate the confidence β of the i-th result. i (x) is:

[0016]

[0017] in, ω k This represents the weight of the k-th rule;

[0018] S3. Load the equipment model, including but not limited to the transmission frequency, transmission antenna height, transmission power equipment parameters, as well as temperature, water vapor density, visibility environmental parameters;

[0019] S4. Load existing equipment data, mainly targeting the equipment's existing boundary performance under known parameters including transmission frequency, transmission antenna height, transmission power, and known environmental conditions including temperature, water vapor density, and visibility.

[0020] S5. Using a small sample dataset generation algorithm, expand the dataset and predict the boundary performance of the equipment under the conditions of step S3.

[0021] S6. Using the inference model in step S2, calculate the confidence level in step S5.

[0022] Furthermore, in step S5 above, the steps of the small sample dataset generation algorithm are as follows:

[0023] Input: The data table before sample augmentation is (TK*L), the position of the attribute column is P, the number of attribute columns is n, the small sample discrimination threshold is m, and the number of samples to be augmented is H;

[0024] Output: The data table after sample augmentation is T(H+K)*L;

[0025] Step 1: Select the attribute columns according to the input requirements. Each attribute column is a feature of the sample to be generated. The data type and numerical results of the feature do not show any missing attribute data in each sample row. Use these features to construct the feature space.

[0026] Step 2: Locate all known samples in the feature space and perform sample clustering;

[0027] Step 3: Based on the sample clustering results, samples with a number less than the small sample discrimination threshold belong to the small sample dataset;

[0028] Step 4: Generate samples in the feature space using the SMOTE, Borderline SMOTE, or ADASYN algorithm;

[0029] Step 5: Project the generated samples in the feature space onto each feature dimension to obtain the corresponding feature values ​​on different attribute columns;

[0030] Step 6: Fill the feature values ​​from each attribute column in Step 5 into the data table.

[0031] Due to the adoption of the technical solution described above, the present invention has the following advantages:

[0032] This invention presents a method for evaluating the boundary performance of electronic reconnaissance equipment based on big data analytics. It utilizes an improved confidence rule base and leverages big data technology to assess the performance of electronic reconnaissance equipment. By fully utilizing data and reducing inference errors through training, it avoids the limitations of expert experience while effectively handling both quantitative and qualitative knowledge. This method is suitable for evaluating the boundary performance of equipment and can be further transferred to the performance evaluation of electronic reconnaissance equipment, ultimately revealing its boundary performance. While fully utilizing data and reducing inference errors through training, it analyzes the impact of sample distribution, particularly small samples, anchoring small sample datasets and proposing suggestions for subsequent data collection. This effectively alleviates the problems of unbalanced sample collection and lack of clear collection objectives, leading to a lack of targeted sample collection. It allows users to more consciously improve the signal acquisition process, reducing the risk of blind collection, promoting continuous and standardized collection, and improving the quality of the dataset. This, in turn, enables better optimization of the analysis model based on big data.

[0033] This invention presents a method for evaluating the boundary performance of electronic reconnaissance equipment based on big data analytics. Starting with data, it first evaluates the quality of the data, supplements the data content, refines the data results, and finally judges and optimizes the data based on the quality of the inference rules, thereby extracting better rules and proposing optimization directions for equipment performance evaluation. By utilizing small sample localization and enhancement techniques, it promptly identifies data defects and performs data cleaning, ultimately providing further suggestions for data collection.

[0034] This invention presents a method for evaluating the boundary performance of electronic reconnaissance equipment based on big data analytics. Through reasonable data cleaning and supplementation strategies, it ensures that even incomplete inference input data can still be used to ultimately evaluate equipment performance. Based on the system linkage structure and signal processing flow of the electronic reconnaissance equipment, a hierarchical confidence rule base is established. The parameters of the initially defined rule base are optimized using subsequent data, resulting in a continuously improving and updated performance evaluation process and system for electronic reconnaissance equipment, adaptable to changes in new equipment. Furthermore, it establishes solutions for imbalanced input data and the occurrence of small sample data, integrating them into the overall scheme's process, increasing the scheme's completeness and practicality, and providing effective suggestions for equipment data acquisition and testing. Attached Figure Description

[0035] Figure 1 This is a flowchart of the electronic reconnaissance equipment boundary performance evaluation method based on big data analysis technology according to the present invention;

[0036] Figure 2 This invention relates to an evaluation index system for electronic reconnaissance equipment systems based on a big data analytics-driven method for evaluating the boundary performance of electronic reconnaissance equipment.

[0037] Figure 3 This is a schematic diagram of the population migration mode in the multi-level genetic algorithm of the electronic reconnaissance equipment boundary performance evaluation method based on big data analysis technology in this invention. Detailed Implementation

[0038] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0039] like Figure 1 As shown, a method for evaluating the boundary performance of electronic reconnaissance equipment based on big data analytics includes the following steps:

[0040] S1. Establish an indicator system for electronic reconnaissance equipment systems. Considering that electronic reconnaissance equipment systems mainly include two categories, communication reconnaissance and radar reconnaissance, two different indicator systems are established to reflect their differences. During the analysis, a similar performance evaluation process is used to optimize each indicator system as necessary, so that the evaluation method has universality.

[0041] like Figure 2 As shown, the operational effectiveness of radar reconnaissance equipment needs to be analyzed and comprehensively evaluated independently from four basic aspects: detection performance, signal interception capability, signal sorting and identification capability, and parameter measurement capability. Based on the relevant knowledge of communication reconnaissance technology and equipment, as well as the main process of communication reconnaissance, the communication reconnaissance capability of electronic warfare equipment is divided into signal search and interception capability, signal detection capability, and signal feature identification and processing capability.

[0042] The signal search and interception capability consists of the ability to search for conventional signals, the ability to search for frequency-hopping signals, and the ability to search for spread spectrum signals.

[0043] Communication signal detection refers to the process of intercepting and receiving a target signal of interest after it has been detected by a reconnaissance receiving device, and measuring its various technical characteristic parameters; the signal detection capability includes the ability to measure the signal center frequency, the signal bandwidth, the signal level, and the frequency hopping rate.

[0044] The signal feature recognition capability includes the ability to recognize signal technical features and the ability to recognize subtle signal features;

[0045] S2. Establish a confidence base and build a rule-based confidence base inference model; including the following sub-steps:

[0046] S2.1 Confidence base initialization, constructed from the results of manual analysis of existing data;

[0047] S2.2, Rule Vectorization: The rule is represented in the following vector form: Rk: If x is A k Then DWith belief degree β k Where x represents the input vector; A k The reference value for the k-th rule after combination (set) is represented; D represents the result vector; β k This represents the confidence level assessment of the output result; further, the result assessment is expressed as: S(A) k )={(D i ,β 1,k ),i=1,2...,N};

[0048] S2.3. Use the ER algorithm to fuse the rules, and evaluate the confidence β of the i-th result. i (x) is:

[0049]

[0050] in, ω k This represents the weight of the k-th rule;

[0051] S3. Load the equipment model, which mainly includes equipment parameters such as transmission frequency, transmission antenna height, and transmission power, as well as environmental parameters such as temperature, water vapor density, and visibility.

[0052] S4. Load existing equipment data, mainly targeting the equipment's existing boundary performance under known parameters including transmission frequency, transmission antenna height, transmission power, and known environmental conditions including temperature, water vapor density, visibility, etc.

[0053] S5. Expand the dataset using the small sample dataset generation algorithm to predict the boundary performance under the conditions of step S3; the specific steps of the above small sample dataset generation algorithm are as follows:

[0054] Input: The data table before sample augmentation is (TK*L), the position of the attribute column is P, the number of attribute columns is n, the small sample discrimination threshold is m, and the number of samples to be augmented is H;

[0055] Output: The data table after sample augmentation is T(H+K)*L;

[0056] Step 1: Select the attribute columns according to the input requirements. Each attribute column is a feature of the sample to be generated. The data type and numerical results of the feature do not show any missing attribute data in each sample row. Use these features to construct the feature space.

[0057] Step 2: Locate all known samples in the feature space and perform sample clustering;

[0058] Step 3: Based on the sample clustering results, samples with a number less than the small sample discrimination threshold belong to the small sample dataset;

[0059] Step 4: Generate samples in the feature space using the SMOTE, Borderline SMOTE, or ADASYN algorithm;

[0060] Step 5: Project the generated samples in the feature space onto each feature dimension to obtain the corresponding feature values ​​on different attribute columns;

[0061] Step 6: Fill the feature values ​​from each attribute column in Step 5 into the data table;

[0062] S6. Using the inference model in step S2, calculate the confidence level in step S5.

[0063] The above description is only a preferred embodiment of the present invention and not a limitation thereof. Any equivalent changes and modifications made in accordance with the scope of the present invention without departing from the spirit and scope of the present invention shall be within the scope of patent protection of the present invention.

Claims

1. A method for evaluating the boundary performance of electronic reconnaissance equipment based on big data analytics, characterized by: It includes the following steps: S1. Establish an indicator system for electronic reconnaissance equipment systems; S2. Establish a confidence base and build a rule-based confidence base inference model; S3. Load the equipment model, including transmission frequency, transmission antenna height, transmission power equipment parameters, as well as temperature, water vapor density, visibility environmental parameters; S4. Load existing equipment data to obtain the existing boundary performance of the equipment under known parameters, including transmission frequency, transmission antenna height, transmission power, and known environmental conditions, including temperature, water vapor density, and visibility. S5. Expand the dataset using a small sample dataset generation algorithm to predict the boundary performance of the equipment model from step S3; the steps of the small sample dataset generation algorithm are as follows: Input: The data table before sample augmentation is (TK*L), the position of the attribute column is P, the number of attribute columns is n, the small sample discrimination threshold is m, and the number of samples to be augmented is H; Output: The data table after sample augmentation is T(H+K)*L; Step 1: Select the attribute columns according to the input requirements. Each attribute column is a feature of the sample to be generated. The data type and numerical results of the feature do not show any missing attribute data in each sample row. Use these features to construct the feature space. Step 2: Locate all known samples in the feature space and perform sample clustering; Step 3: Based on the sample clustering results, samples with a number less than the small sample discrimination threshold belong to the small sample dataset; Step 4: Generate samples in the feature space using the SMOTE, Borderline SMOTE, or ADASYN algorithm; Step 5: Project the generated samples in the feature space onto each feature dimension to obtain the corresponding feature values ​​on different attribute columns; Step 6: Fill the feature values ​​from each attribute column in Step 5 into the data table; S6. Using the inference model in step S2, calculate the confidence level in step S5.

2. The method for evaluating the boundary performance of electronic reconnaissance equipment based on big data analysis technology according to claim 1, characterized in that: Its step S2, Includes the following sub-steps: S2.1 Confidence library initialization S2.2, Rule Vectorization: The rule is represented in the following vector form: Rk: If x is A k Then D With belief degree β k Where x represents the input vector; A k The reference value for the k-th rule after combination is represented; D represents the result vector; β k This represents the confidence level assessment of the output result; the result assessment is expressed as: S(A) k )={(D i ,β1,k),i=1,2...,N}; S2.

3. Use the ER algorithm to fuse the rules, and evaluate the confidence β of the i-th result. i (x) is: in, ω k This represents the weight of the k-th rule.

Images