A method for identifying maritime targets based on fuzzy theory and DS evidence theory

Abstract

This invention discloses a maritime target identification method based on fuzzy theory and D-S evidence theory, belonging to the field of target identification technology. The method includes: 1) Sensors detecting target parameter information at a certain moment, constructing a membership function for each parameter belonging to a target; 2) Constructing corresponding weight coefficients by comprehensively considering the characteristics of each target parameter, environmental influences, and interference; 3) Obtaining the target basic probability assignment (BPA) function at different moments based on a comprehensive evaluation method using fuzzy theory; 4) Calculating the mutual distance between evidence at each moment, obtaining the weight coefficients of the credibility of each piece of evidence, and averaging them to obtain synthetic evidence; 5) Obtaining the BPA function value of the sensor using the D-S evidence synthesis rule; 6) Calculating the mutual distance between the evidence from each sensor based on the BPA function value of each sensor, obtaining the weight coefficients of the credibility between the evidence, averaging them to obtain synthetic evidence, and finally synthesizing it using the D-S evidence synthesis rule to determine the target identity and make a decision.

Description

Technical Field

[0001] This invention relates to the field of maritime target identification, and more specifically, to a maritime target identification method based on fuzzy theory and DS evidence theory. Background Technology

[0002] As the navy moves towards the open ocean, the blue seas, and the global arena, the accurate and effective identification of maritime target types has received continuous attention. Rapid perception of maritime battlefield information and the ability to identify maritime target types can provide crucial information support for commanders' decision-making.

[0003] Sensors used to detect maritime targets include radar, infrared, optical, visual sensors, sonar, and more. Improving the precision of a single sensor can increase the accuracy of target identification, but this approach often reaches certain technological limitations. Furthermore, accurate judgment requires data from multiple layers of land, sea, and air. Therefore, maritime target identification must rely on data fusion, fully utilizing multi-source, multi-period, and multi-dimensional data fusion to improve the reliability and accuracy of maritime target identification. Summary of the Invention

[0004] The technical problem to be solved by this invention is to design a maritime target identification method based on algorithms such as fuzzy theory and DS evidence theory, and to utilize single-sensor single-cycle fusion identification, single-sensor multi-cycle fusion identification, and multi-sensor multi-source fusion identification to effectively and accurately identify maritime targets and meet the needs of situation assessment and threat assessment.

[0005] The technical solution adopted in this invention is as follows:

[0006] A method for identifying maritime targets based on fuzzy theory and DS evidence theory includes the following steps:

[0007] (1) The sensor at a certain time t i Detection target parameter information x j Where 1≤j≤N, and N is the number of objective parameters;

[0008] (2) Based on the correlation between parameter information and the target, construct the membership function μ of each parameter belonging to a certain target. ik (x j ); where μ ik (x j ) represents sensor t i Target parameter information x measured at time 1 j The membership degree of target k, 1≤k≤M, where M is the number of targets;

[0009] (3) Construct the corresponding weighting coefficients ω for each parameter of the target by comprehensively considering the characteristics of each parameter, the degree of parameter identification, environmental impact, interference, and sensor performance. k (x j ); where ω k (x j ) represents the target parameter information x j The weight values ​​belonging to target k;

[0010] (4) A comprehensive evaluation method based on fuzzy theory, which calculates the basic probability allocation (BPA) function value based on the target parameter information:

[0011]

[0012] In the formula, m ik Indicates sensor t i The basic probability assignment (BPA) function value of the target parameter information detected at any time belonging to a certain target k;

[0013] (5) Based on the target BPA function values ​​at each time point, calculate the mutual distance between the evidence at each time point to obtain the weighting coefficient ω of the credibility of each piece of evidence. k (t i After weighted average normalization, the synthetic evidence is obtained; then, the DS evidence synthesis rules are used to synthesize the synthetic evidence to obtain the BPA function value of the sensor.

[0014] (6) Based on the BPA function values ​​of each sensor, calculate the mutual distance between the evidence from each sensor, and obtain the weighting coefficient ω of the credibility of each piece of evidence. k (s n After weighted average normalization, synthetic evidence is obtained; then, the DS evidence synthesis rule is used to synthesize the synthetic evidence, ultimately determining the target's identity and making a decision; where s n This refers to a specific sensor.

[0015] The beneficial effects of this invention are as follows:

[0016] (1) This method takes into account multiple factors such as target parameters, environmental influence, interference, sensor performance and accuracy, and constructs corresponding weight coefficients, which have the characteristics of good anti-interference performance and strong robustness.

[0017] (2) Before DS evidence synthesis, this method calculates the mutual distance between each sensor evidence, obtains the weight coefficient of the credibility of each piece of evidence, and obtains the synthesized evidence after weighted average normalization, thus avoiding the problem of DS synthesis failure when the initial evidence conflicts.

[0018] (3) This method can fully utilize the various parameter data detected at each moment, and the large amount of data support makes the recognition results more accurate and reliable.

[0019] (4) Each sensor node first processes the various parameter data it detects. After processing, it obtains information such as the BPA function of each sensor. Based on the characteristics of each sensor node, the data is fused to form a global decision. The distributed fusion system structure has the advantages of low communication volume, high reliability, and low cost. Attached Figure Description

[0020] Figure 1 This is a flowchart of the single-sensor single-cycle fusion recognition process in this invention.

[0021] Figure 2 This is a flowchart of the single-sensor multi-cycle fusion recognition process in this invention.

[0022] Figure 3 This is a flowchart of the multi-sensor fusion recognition process in this invention. Detailed Implementation

[0023] The following is in conjunction with the appendix Figure 1-3 The present invention will be further described below.

[0024] A method for identifying maritime targets based on fuzzy theory and DS evidence theory includes the following steps:

[0025] like Figure 1 As shown:

[0026] (1) The sensor at a certain time t i Detection target parameter information x j (1≤j≤N), where N is the number of target parameters. For example, in a sensor radar, the parameters x such as the target's distance, velocity, and acceleration detected at time t1 are... j = {Target distance, target velocity, target acceleration};

[0027] (2) Construct parameter information x based on the correlation between parameter information and target. j =Membership function μ of {target distance, target velocity, target acceleration} belonging to a certain target ik (x j (1≤k≤M). For example, based on the characteristics of the target distance parameters, a trapezoidal membership function can be constructed. The probabilities of belonging to targets A, B, and C (e.g., setting the target recognition frame Ф={A, B, C}) based on the distance parameters are respectively (μ...). iA (x1),μ iB (x1),μ iC(x1)). Similarly, based on the characteristics of the target velocity and target acceleration parameters, a suitable membership function is constructed, and the probabilities of belonging to targets A, B, and C are respectively (μ). iA (x2),μ iA (x2),μ iC (x2)), (μ iA (x3),μ iB (x3),μ iC (x3)).

[0028] (3) Construct the corresponding weight coefficients ω for each parameter of the target by comprehensively considering factors such as the characteristics of each parameter and the degree of parameter identification. k (x j ). For example, setting parameter information x j =The weighting coefficients for {target distance, target velocity, target acceleration} are (ω) k (x1),ω k (x2),ω k (x3)).

[0029] (4) A comprehensive evaluation method based on fuzzy theory, which determines the basic probability assignment (BPA) function based on the target parameter information and after normalization:

[0030]

[0031] k = {A, B, C}, therefore M = 3, and k = 1, 2, 3 correspond to the three targets A, B, and C respectively. If three time points are set, m can be obtained according to the formula. 1k m 2k m 3k , representing the basic probability allocation (BPA) function value used to determine whether a target A, B, or C belongs to a target based on the distance, velocity, and acceleration parameters detected by the sensor at times t1, t2, and t3.

[0032] like Figure 2 As shown:

[0033] (5) Based on the obtained target BPA function values ​​at each time point, calculate the mutual distance between the evidence at each time point to obtain the weighting coefficient ω of the credibility of each piece of evidence. k (t i The data is then processed by weighted average normalization to obtain synthetic evidence. The DS evidence synthesis rules are then used to synthesize the synthetic evidence to obtain the BPA function value of the sensor.

[0034] like Figure 3 As shown:

[0035] (6) Based on the obtained BPA functions of each sensor (e.g., setting up 4 sensors and obtaining the BPA function values ​​of s1, s2, s3, and s4), calculate the mutual distance between the evidence from each sensor and obtain the weighting coefficient ω of the credibility of each piece of evidence. k (s n After weighted average normalization, synthetic evidence is obtained. Then, the DS evidence synthesis rules are used to synthesize the synthetic evidence, ultimately determining the target's identity and making a decision.

Claims

1. A maritime target identification method based on fuzzy theory and DS evidence theory, characterized in that, Includes the following steps: (1) The sensor at a certain time t i Detection target parameter information x j Where 1≤j≤N, and N is the number of objective parameters; (2) Based on the correlation between parameter information and the target, construct the membership function μ of each parameter belonging to a certain target. ik (x j ); where μ ik (x j ) represents sensor t i Target parameter information x measured at time 1 j The membership degree of target k, 1≤k≤M, where M is the number of targets; (3) Construct the corresponding weighting coefficients ω for each parameter of the target by comprehensively considering the characteristics of each parameter, the degree of parameter identification, environmental impact, interference, and sensor performance. k (x j ); where ω k (x j ) represents the target parameter information x j The weight values ​​belonging to target k; (4) A comprehensive evaluation method based on fuzzy theory, which calculates the basic probability allocation (BPA) function value based on the target parameter information: In the formula, m ik Indicates sensor t i The basic probability assignment (BPA) function value of the target parameter information detected at any time belonging to a certain target k; (5) Based on the target BPA function values ​​at each time point, calculate the mutual distance between the evidence at each time point to obtain the weighting coefficient ω of the credibility of each piece of evidence. k (t i After weighted average normalization, the synthetic evidence is obtained; then, the DS evidence synthesis rules are used to synthesize the synthetic evidence to obtain the BPA function value of the sensor. (6) Based on the BPA function values ​​of each sensor, calculate the mutual distance between the evidence from each sensor, and obtain the weighting coefficient ω of the credibility of each piece of evidence. k (s n After weighted average normalization, synthetic evidence is obtained; then, the DS evidence synthesis rule is used to synthesize the synthetic evidence, ultimately determining the target's identity and making a decision; where s n This refers to a specific sensor.

Images