Global ellipsoid density fusion method for multi-sensor information fusion

Through the global ellipsoid density fusion method, the problems of insufficient local ellipsoid domain fusion accuracy and high computational complexity in multi-sensor information fusion are solved, and the global state estimation with higher accuracy and robustness are achieved, adapting to complex environments and high dimensions Data fusion.

CN118916832APending Publication Date: 2024-11-08SHANGHAI MARITIME UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410949082.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-16
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The existing multi-sensor information fusion technology has problems such as insufficient accuracy, high computational complexity, serious conservativeness and inability to adapt to complex target system state estimation scenarios in local ellipsoid domain fusion.

Method used

The global ellipsoid density fusion method is used to calculate the sampling points of the ellipsoid by calculating the local state estimating ellipsoid, set constraints and density conditions, filter out the sampling points that meet the conditions, construct an external rectangle and incisate the ellipsoid to obtain the global state estimation result.

Benefits of technology

It improves estimation accuracy and robustness, can describe global state areas more accurately, reduces calculation complexity, and has stronger noise and outlier resistance, and has stronger adaptability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118916832A_ABST
    Figure CN118916832A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of multi-sensor information fusion, in particular to a global ellipsoid density fusion method for multi-sensor information fusion. Obtaining a local state estimation ellipsoid according to a distributed set membership estimation algorithm; calculating sampling points in all the local estimation ellipsoids according to the local state estimation ellipsoids; after the sampling points are obtained, density parameters are set, and all the sampling points meeting the density condition are calculated; according to all the sampling points meeting the density condition, a sampling point set meeting the density condition is obtained, an internally tangent ellipsoid is constructed in a surrounding mode through a circumscribed rectangle, and a global state estimation result is obtained. Compared with the prior art, the global ellipsoid density fusion method has higher resistance to noise and abnormal values, so that the global ellipsoid density fusion method has the advantages of being capable of keeping higher estimation precision and better robustness when facing complex environments and sensor errors, and the like.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of multi-sensor information fusion, and in particular to a global ellipsoid density fusion method for multi-sensor information fusion. Background Art

[0002] Distributed set-membership state estimation is an information fusion scheme that combines multiple sensors to work together and uses set-membership estimation theory to accurately estimate the estimated system. It has shown great potential in target tracking, environmental monitoring and other fields. Its core idea is that multiple sensor nodes can obtain a more accurate target state estimation set than a single sensor through information interaction and fusion. In the distributed set-membership state estimation process, each sensor node independently collects the state information of the estimated system and can exchange data with surrounding nodes. Then, based on the observed data and the system model, an ellipsoid domain containing the true state is calculated, which is the local state estimation ellipsoid. Then, the local state estimation ellipsoids of each node are fused to finally obtain a global state estimation ellipsoid that contains the true state and is more accurate than the local state estimation ellipsoid.

[0003] Existing fusion schemes for local ellipsoid domains have many limitations. Although the union scheme ensures the integrity of the global estimation, it sacrifices the estimation accuracy. The scheme of obtaining the global minimum volume ellipsoid of the intersection of local state estimation ellipsoids pursues the improvement of accuracy. It not only has the problems of conservative and high computational complexity of different optimization schemes, but also only focuses on the common part of all ellipsoids, which will discard the state information of the non-public part and may lose valuable data. The covariance fusion scheme obtains the global fusion result by fusing the covariance matrix of the ellipsoid. The covariance matrix can only reflect the statistical characteristics of the data and cannot accurately describe the geometric shape of the ellipsoid. The fusion result is usually too conservative and the estimation range is too large. However, when facing the problem of multi-sensor fusion, it is necessary to repeat the operation on multiple local information, and the algorithm complexity will increase accordingly. The circumscribed rectangle method fusion scheme assumes that the local state estimation result is an ellipsoid parallel to the coordinate axis, which may not be consistent with the actual fusion situation. The accuracy improvement of this scheme is very limited compared with the intersection scheme, and there is also a conservative problem, which is difficult to adapt to complex target system state estimation scenarios. It is worth noting that for practical problems with low fault tolerance, higher precision estimation of regions is required, and the existing fusion schemes are highly conservative and cannot meet the requirements. Summary of the invention

[0004] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and to provide a global ellipsoid density fusion method for multi-sensor information fusion.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] The present invention provides a global ellipsoid density fusion method for multi-sensor information fusion, comprising the following steps:

[0007] S1, obtain the local state estimation ellipsoid according to the distributed set membership estimation algorithm;

[0008] S2, calculating the sampling points of all local estimated ellipsoids according to the local state estimated ellipsoid;

[0009] S3, after obtaining the sampling points, setting constraint conditions, screening out sampling points that meet the constraint conditions, and combining them into a global sampling point set;

[0010] S4, setting a density condition, calculating the global sampling points in the global sampling point set that meet the density condition, and obtaining a global sampling point set that meets the density condition;

[0011] S5. Construct a circumscribed rectangle according to a set of global sampling points that meet the density condition, and use the circumscribed rectangle to construct an inscribed ellipsoid to obtain a global state estimation result.

[0012] The local state estimation ellipsoid is composed of a state estimation value and a parameter matrix.

[0013] The constraint condition is expressed by the following formula:

[0014]

[0015] in, is the local ellipsoid sampling point vector, is the estimated value of the state, is the parameter matrix.

[0016] The density condition includes two density parameters: the distance between sampling points and the number of sampling points, which are expressed by the following formula:

[0017]

[0018] in, represents the set of sampling points within the neighborhood radius γ, For sampling point and The Euclidean distance between h and Represents the index in the collection, ranging from 1 to Represents a set of sampling points that meet density requirements.

[0019] The process of constructing the bounding rectangle includes: firstly determining the maximum value and the minimum value of the sampling point set that meets the conditions in each dimension, and taking the center of the maximum value and the minimum value as the center point of the bounding rectangle.

[0020] The long side and the short side of the circumscribed rectangle are used as the long axis and the short axis of the inscribed ellipse respectively; and the center point of the circumscribed rectangle is used as the center point of the inscribed ellipse.

[0021] In a second aspect, the present invention provides a computer program product, comprising a computer program / instruction, which implements any of the above-mentioned methods when executed by a processor.

[0022] In a third aspect, the present invention provides an electronic device, comprising a memory, a processor, and a program stored in the memory, wherein the processor implements any of the above-described methods when executing the program.

[0023] In a fourth aspect, the present invention provides a storage medium having a program stored thereon, wherein the program implements any of the above-described methods when executed.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] 1. The global ellipsoid density fusion method adopted in the present invention has stronger resistance to noise and outliers, so that the global ellipsoid density fusion method can maintain higher estimation accuracy and better robustness when facing complex environments and sensor errors.

[0026] 2. The global ellipsoid density fusion method adopted in the present invention is not limited to the geometric boundary of the ellipsoid, but deeply analyzes the density distribution of sampling points inside each ellipsoid, which can more accurately describe the area of ​​global state estimation and effectively improve the estimation accuracy.

[0027] 3. For high-precision requirements, the present invention can be achieved by modifying density parameters and the number of sampling points, and the accuracy of the fusion result can be controlled more flexibly.

[0028] 4. For high-dimensional data fusion, the present invention is faster and has lower computational complexity than existing optimization algorithms. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 Schematic diagram of the density information of the ellipsoid sampling points for estimating multiple local states on a two-dimensional plane;

[0030] Figure 2 It is a schematic diagram of the steps of the present invention;

[0031] Figure 3 Schematic diagram of the global ellipsoid density fusion method. DETAILED DESCRIPTION

[0032] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0033] Taking into account the problems and shortcomings of the traditional state estimation ellipsoid fusion scheme, the present invention designs a global ellipsoid density fusion method to fuse the local state estimation ellipsoids and obtain the global state estimation result. Figure 1 It is a schematic diagram of the density information of the ellipsoid sampling points for multi-local state estimation on a two-dimensional plane, as shown in Figure 1 As shown in the figure, after the results of the local state ellipsoid are fused, the density of sampling points in different areas will be different. The red sampling points are located in areas with higher density, and the blue sampling points are located in areas with lower density. It is more likely that the true state is in areas with higher density.

[0034] Based on the above content, the present invention designs a method of global ellipsoid density fusion, which is applied to multi-sensor distributed set membership estimation. The method can reflect different density areas to obtain areas with different conservative degrees. First, each local state estimation ellipsoid is sampled to obtain local state sampling points, and then, according to the number of neighborhood sampling points and the Euclidean distance between the sampling points, the high-density area sampling points that meet the requirements are obtained. Finally, a rectangle is used to contain all the sampling points that meet the conditions, and the inscribed ellipsoid is calculated, which is the final global state estimation result.

[0035] The specific process is as follows Figure 2 As shown, the following steps are included:

[0036] Consider a basic class of nonlinear discrete-time systems:

[0037]

[0038] in, represents the system state matrix at time k+1; represents the system state matrix at time k; represents the system input matrix at time k; A represents the process noise matrix at time k; P ,B P ,D P is a known parameter matrix.

[0039] Step S1: Obtain a local state estimation ellipsoid according to an existing distributed set membership estimation algorithm.

[0040] Specifically, consider a distributed sensor network consisting of N sensor nodes, each of which estimates the same system. Assume that at time k, each sensor node can obtain a local state estimation ellipsoid containing the true state value according to the set membership estimation algorithm. in, It represents the state estimation value obtained by the i-th sensor node at time k+1, which is also the center point of the local state estimation ellipsoid; represents the parameter matrix of the local state estimation ellipsoid, and are the solved ellipsoid parameters.

[0041] Step S2: Calculate the sampling points of all local estimated ellipsoids according to the local state estimated ellipsoid.

[0042] Specifically, Figure 3 As shown, after obtaining all local state estimation ellipsoids After that, it is sampled. First, the horizontal axis is sampled, and the sampling points of the horizontal axis are The vertical axis is sampled, and the vertical axis sampling points are Then the local ellipsoid sampling point vector is expressed as and is the starting value of the horizontal axis and the starting value of the vertical axis, and is the end value of the horizontal axis and the end value of the vertical axis, [P i,k+1 ] 11 ,[P i,k+1 ] 22 YesP i,k+1 The diagonal elements of , λ is the number of generated points, and the index l starts from 0 to λ-1.

[0043] for In the case of multiple dimensions, use n x Indicates status The dimension of the local ellipsoid sampling point vector can be expressed as:

[0044]

[0045] Among them, the starting vector End Vector Where P i,k+1 The diagonal elements of The weights are

[0046]

[0047] Step S3: In order to obtain accurate sampling points in the ellipsoid, constraint conditions are set, and all sampling points in the ellipsoid that meet the constraint conditions are screened out and combined into a global sampling point set.

[0048] Among them, each local ellipsoid sampling point vector The following constraints are met:

[0049]

[0050] All sampling points of the local estimated ellipsoid that meet the constraints are screened out, and the sampling point vector that meets the above conditions is recorded as Then the global sampling point set is expressed as

[0051] Step S4: setting density conditions, and calculating all sampling points that meet the density conditions based on the global sampling point set that meets the constraint conditions, to obtain a global sampling point set that meets the density conditions.

[0052] Specifically, Figure 3 As shown in the figure, the density condition mainly includes two density parameters: the distance between sampling points and the number of sampling points. The distance between sampling points is the radius of the sampling point set with this distance as the sampling point set. The number of sampling points is the number of sampling points within the circle with the distance between sampling points as the radius.

[0053] They are given by the following formulas:

[0054]

[0055] in, represents the set of sampling points within the neighborhood radius γ, For sampling point and The Euclidean distance between h and The range of the index in the collection is 1 to represents the final set of sampling points that meets the density requirement, that is, a set of sampling points that contains at least m sampling points and a neighborhood radius of γ.

[0056] Step S5: enclose the global sampling point set that meets the density condition with a circumscribed rectangle and calculate its inscribed ellipsoid to obtain a global state estimation result.

[0057] Specifically, Figure 3 As shown, the inscribed ellipsoid is obtained through the circumscribed rectangle, and the two-dimensional inscribed ellipse is first constructed. The process of constructing the inscribed ellipse includes: the sampling point set The center point of the minimum enclosing rectangle is Among them, assuming that the set n is a set The number of elements in Represents the sampling point set The coordinate minimum and maximum vectors. Use the center point and axis length of the minimum circumscribed rectangle to construct the inscribed ellipse, and the axis length vector is Γ k+1 =X max,k+1 -X min,k+1 , the center of the inscribed ellipse is the center point of the minimum circumscribed rectangle, and the length of the axis coincides with the length of the minimum outer boundary rectangle, so the inscribed ellipse can be expressed as:

[0058]

[0059] in, is the global fusion ellipsoid estimate, is the parameter matrix of the global fusion ellipsoid. is the axis length vector Γ k+1 The amount.

[0060] For dimension n x The ellipsoid of other dimensions can be obtained by the same method as that of constructing the inscribed ellipse. First, find the sampling point set The minimum enclosing rectangle of . The center point of the minimum enclosing rectangle can be expressed as in, Represents a set of sampling points The coordinate minimum vector of Represents a set of sampling points The maximum coordinate vector of the ellipsoid. The axis length vector of the ellipsoid is Γ k+1 =X max,k+1 -X min,k+1 .

[0061] The inscribed ellipsoid is formed using the center point of the minimum circumscribed rectangle and the axis length of each dimension. The center point of the inscribed ellipsoid is the same as the center point of the minimum circumscribed rectangle, and the length of each axis coincides with the length of the minimum circumscribed rectangle.

[0062] Therefore, the inscribed ellipsoid can be described as follows:

[0063]

[0064] in, is the global fusion ellipsoid estimate, is the parameter matrix of the global fusion ellipsoid. is the axis length vector Γ k+1 The components in each dimension.

[0065] This embodiment provides a computer program product, including a computer program / instruction, which implements the method mentioned in the above embodiment when executed by a processor.

[0066] This embodiment provides an electronic device, including a memory, a processor, and a program stored in the memory, and the processor implements the method mentioned in the above embodiment when executing the program.

[0067] This embodiment provides a storage medium on which a program is stored, and when the program is executed, the method mentioned in the above embodiment is implemented. The storage medium includes: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0068] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A global ellipsoid density fusion method for multi-sensor information fusion, It is characterized in that The following steps are involved: S1, obtain the local state estimation ellipsoid according to the distributed set membership estimation algorithm; S2, calculating the sampling points of all local estimated ellipsoids according to the local state estimated ellipsoid; S3, after obtaining the sampling points, setting constraint conditions, screening out sampling points that meet the constraint conditions, and combining them into a global sampling point set; S4, setting a density condition, calculating the global sampling points in the global sampling point set that meet the density condition, and obtaining a global sampling point set that meets the density condition; S5. Construct a circumscribed rectangle according to the global sampling point set that meets the density condition, and use the circumscribed rectangle to construct an inscribed ellipsoid to obtain a global state estimation result.

2. A global ellipsoid density fusion method for multi-sensor information fusion according to claim 1, It is characterized in that The local state estimation ellipsoid is composed of a state estimation value and a parameter matrix.

3. The global ellipsoid density fusion method for multi-sensor information fusion according to claim 1, It is characterized in that The constraint condition is expressed by the following formula: in, is the local ellipsoid sampling point vector, is the estimated value of the state, is the parameter matrix.

4. The global ellipsoid density fusion method for multi-sensor information fusion according to claim 1, It is characterized in that The density condition includes two density parameters: the distance between sampling points and the number of sampling points, which are expressed by the following formula: in, represents the set of sampling points within the neighborhood radius γ, For sampling point and The Euclidean distance between h and Represents the index in the collection, ranging from 1 to Represents a set of sampling points that meet density requirements.

5. The global ellipsoid density fusion method for multi-sensor information fusion according to claim 1, It is characterized in that The process of constructing the bounding rectangle includes: firstly determining the maximum value and the minimum value of the sampling point set that meets the conditions in each dimension, and taking the center of the maximum value and the minimum value as the center point of the bounding rectangle.

6. A global ellipsoid density fusion method for multi-sensor information fusion according to claim 5, It is characterized in that The axis length of the circumscribed rectangle is used as the axis length of the inscribed ellipsoid, and the center point of the circumscribed rectangle is used as the center point of the inscribed ellipsoid.

7. The global ellipsoid density fusion method for multi-sensor information fusion according to claim 6, It is characterized in that The inscribed ellipsoid includes a two-dimensional inscribed ellipse and a three-dimensional or higher inscribed ellipsoid.

8. A computer program product comprising a computer program / instructions, It is characterized in that When the computer program / instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

9. An electronic device comprising a memory, a processor, and a program stored in the memory, It is characterized in that When the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

10. A storage medium having a program stored thereon, It is characterized in that When the program is executed, the method according to any one of claims 1 to 7 is implemented.