A multi-sensor target fusion recognition method based on quantum negation

Abstract

The present invention relates to a multi-sensor target fusion recognition method based on quantum negation. The recognition framework and basic probability distribution function are constructed based on the sensor data of the target. Based on the mixed quantum state, the basic probability distribution function of the sensor is converted into a quantum representation. The quantum representation of the basic probability distribution function is subjected to a negation transformation. The negation of the basic probability distribution function is obtained, and the negation #imgabs0# of n pieces of evidence are fused according to the Dempster combination rule to obtain the fused basic probability distribution function. The negation #imgabs1# of the basic probability distribution function is obtained, and the basic probability distribution function #imgabs2# is converted into a probability distribution. The recognition result is judged based on the obtained probability result, and P(θ i ) is used as the target recognition result. By constructing negation generators, the present invention can effectively represent the quantum conversion process from original evidence to evidence negation. Combined with the Dempster combination rule in evidence theory, it can effectively process uncertain information.

Description

Technical Field

[0001] The invention belongs to the field of target recognition and relates to a multi-sensor target fusion recognition method based on quantum negation. Background Art

[0002] Target recognition is a crucial aspect of helicopter flight and landing. The identified target type significantly impacts helicopter flight and landing in harsh environments. Target recognition relies primarily on the results obtained by various sensors. Based on the accuracy of each sensor's recognition of the target, these results are integrated to obtain an accurate description of the target.

[0003] In recent years, the main focus of target recognition research, both domestically and internationally, has been on extracting target features and using them to distinguish different targets. To obtain a more complete description of the target and improve recognition accuracy, multi-sensor fusion recognition is often employed. Common fusion recognition methods include DS evidence theory, Bayesian criterion, neural networks, and expert systems. DS evidence theory has garnered increasing attention and application due to its ability to effectively handle unknown and uncertain issues. In DS theory, the basic probability assignment (BPA) is a function that represents uncertain information and can be interpreted as a non-additive measure, interval probability, or subjective belief. A growing number of scholars are exploring new perspectives on the uncertainty inherent in the BPA in DS evidence theory. Some have introduced quantum mechanics into DS evidence theory, using quantum superposition to address some of the difficulties DS theory faces in uncertainty modeling and information fusion. Based on this, a quantum representation of the basic probability assignment function (BPA) based on mixed quantum states has been proposed to explain the source of uncertainty in BPA. To better handle uncertain information, the negation generation of the BPA has recently attracted considerable attention. The quantized representation and negation of BPA in evidence theory play an important role in the representation of uncertain information, multi-sensor information fusion, medical diagnosis, military command, and target recognition.

[0004] The quantization representation and negation of BPA in evidence theory have many advantages in multi-source information fusion. Applying it in target recognition can better handle the fusion of multi-sensor data and achieve more accurate target recognition. Summary of the Invention

[0005] Technical problems to be solved

[0006] In order to avoid the shortcomings of the prior art, the present invention proposes a multi-sensor target fusion recognition method based on quantum negation.

[0007] To achieve target recognition, this paper proposes a multi-sensor target fusion recognition technology based on quantum negation, drawing on the quantum representation of BPA in evidence theory. This method provides a method for fusion recognition of targets. This method effectively handles the fusion analysis of multi-sensor data and accurately determines the identified targets.

[0008] Technical Solution

[0009] A multi-sensor target fusion recognition method based on quantum negation is characterized by the following steps:

[0010] Step 1: Build an identification framework and basic probability distribution function based on the target's sensor data:

[0011] Given the number of categories L of targets to be identified, the number n of sensors, and the target recognition accuracy α of each sensor, i , the sensors are numbered as 1, 2, ..., n, and the recognition accuracy of each sensor is α1, α2, ..., α i ,...,α n ;

[0012] Step 101: Construct a recognition framework for the sensor to identify the target as Θ = {θ1, θ2, ..., θ i ,...,θ L}, where θ i , i=1,2,...,L is the target type that the sensor can identify;

[0013] Step 102: Based on the DS evidence theory and the target type θ identified by each sensor j And the recognition accuracy α of each sensor i , construct the basic probability distribution function of each sensor as in is θ j The complement of j} is {θ2}, then is {θ1,θ3,...,θ i ,...,θ L}, where θ i , i≠2;

[0014] Step 2: Based on the mixed quantum state, the basic probability distribution function m of the sensor is i Converted into quantum representation:

[0015]

[0016] Step 3: The basic probability distribution function m i The quantum representation ρ iPerform a negation transformation to obtain

[0017] Step 301: Based on the recognition framework of the recognition target Θ={θ1,θ2,...,θ i ,...,θ L} Construct a negation generator, which represents the conversion relationship between the elements in the identification framework when generating the negation of the basic probability distribution function, by establishing a unitary transfer matrix. The construction method is as follows:

[0018] The negated generator is all n-order square matrices X that satisfy the following conditions i , where n is the size of the given recognition frame:

[0019] (1) All elements are either 1 or 0;

[0020] (2) The sum of each column of the matrix is ​​1;

[0021] (3) The sum of each row of the matrix is ​​1;

[0022] (4) The main diagonal elements are 0;

[0023] All n-order square matrices X that meet the above conditions i This is the negation generator constructed to prove that every square matrix X i is satisfied A unitary matrix, where I is the n-th order identity matrix;

[0024] The number of negated generators is given by:

[0025]

[0026] Step 302: Calculate ρ based on the constructed negation generator i Negation For a given identification framework, we get all the negated generators X i After that, the corresponding probability is p i ,satisfy Due to X i is a unitary matrix, so the basic probability distribution function quantum representation ρ i Negation Computed via a unitary transformation of the original density matrix:

[0027]

[0028] Step 4: Find the basic probability distribution function m i Negation

[0029] Step 401: Based on the quantum measurement principle in quantum mechanics, the probability of each event in Θ is calculated using the corresponding quantum measurement operator. For each singleton, its probability is calculated using the following formula:

[0030]

[0031] M θ =|θ><θ| is the quantum measurement operator, It's M θ The conjugate transpose matrix of , tr(·) is the operator for finding the trace of the matrix;

[0032] Step 402: Generate the associated belief function Bel(F) of each proposition: For a proposition F, since Bel(F) is the lower limit of the interval probability in DS theory, under the existing constraints, by calculating P i (F) to derive Bel(F):

[0033]

[0034] Step 403: Obtain the basic probability distribution function m of quantum representation according to the Moss-Ubi transform i Negation According to DS theory, The negation of n pieces of evidence is calculated by the following formula:

[0035]

[0036] Step 5: According to Dempster's combination rule, the negation of the n pieces of evidence obtained in step 4 is Perform fusion to obtain the basic probability distribution function after fusion

[0037] Step 6: Replace the basic probability distribution function m of the sensor in step 2 with the basic probability distribution function m after fusion in step 5 i , repeat steps 2 to 4 to find the negation of m

[0038] Step 7: According to the formula Among them, A is a subset of the identification framework Θ, |A| represents the number of elements in A, and the basic probability distribution function Convert it into probability distribution, make judgment on the recognition result according to the obtained probability result, and take P(θ i ) is taken as the result of target recognition.

[0039] The step 2 is to transform the basic probability distribution function m of the sensor into i The process of converting to quantum representation is:

[0040] Step 201: Obtain recognition frame Θ={θ1,θ2,...,θ i ,...,θ L The quantum representation of Θ is a set of complete and mutually orthogonal basis vectors on the complex field, denoted as Θ q ={|θ1>,|θ2>,...,|θ i >,...,|θ L >}, each ground state is represented by a vector |θ i >=[v1,...,v j ,...,v L ] T , for j=i, satisfy v j =1, for j≠i, satisfy v j =0;

[0041] Step 202: The basic probability distribution function m i The proposition F in is transformed into a quantum representation where z θ is a complex number z θ =α θ +iβ θ , i is an imaginary number and satisfies ||·|| is the modulus of the complex number. After that, the basic probability distribution function m i The quantum representation of

[0042] The Dempster combination rule is in

[0043] Beneficial effects

[0044] The present invention proposes a multi-sensor target fusion recognition method based on quantum negation, which builds an identification framework and a basic probability distribution function based on the sensor data of the target, and transforms the basic probability distribution function m of the sensor into a quantum state based on the mixed quantum state. i Convert it into quantum representation and transform the basic probability distribution function m i The quantum representation ρ i Perform a negation transformation to obtain Depend on Find the basic probability distribution function m i Negation According to Dempster's combination rule, the negation of n pieces of evidence Perform fusion to obtain the basic probability distribution function after fusion Find the negation of the basic probability distribution function The basic probability distribution function Convert to probability distribution. Make a judgment on the recognition result based on the obtained probability result, and take P(θ i ) is taken as the result of target recognition.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. The present invention has simple steps, reasonable design, and is easy to implement and use.

[0047] 2. The present invention can effectively represent the quantum conversion process from original evidence to evidence negation by constructing negation generators.

[0048] 3. The present invention can effectively identify the target type through quantum negation conversion of evidence under the condition of limited known information.

[0049] 4. The quantum negation representation of the basic probability distribution function proposed in this invention, combined with the Dempster combination rule of evidence theory, can effectively realize the processing of uncertain information.

[0050] In summary, the technical solution of the present invention is reasonably designed. It represents known target information based on quantum negation and uses the Dempster combination rule of evidence theory to perform target fusion identification. It not only expands the representation of uncertain information and obtains more information contained in evidence negation, but also can perform fusion under the condition of multiple known target information sources to effectively identify the target type. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 : The overall flow chart of the present invention;

[0052] Figure 2 :The evidence fusion flow chart of the present invention,

[0053] and denote the negation of the basic probability distribution function and Dempster's combination rule, respectively. DETAILED DESCRIPTION

[0054] The present invention will now be further described with reference to the embodiments and accompanying drawings:

[0055] The method of the present invention is further described below in detail with reference to the accompanying drawings and embodiments. This example illustrates how sensors identify ground obstacles when a helicopter lands in a certain area. Different sensors may identify different target types due to varying recognition accuracy. By fusing target types identified by different sensors, the proposed target recognition implementation steps are explained.

[0056] It should be noted that, in the absence of conflict, the embodiments and properties in the embodiments of the present application can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0057] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0058] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can, for example, be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0059] For ease of description, spatially relative terms such as "above", "above", "on the upper surface of", "above", etc. may be used herein to describe the spatial positional relationship of a device or feature to other devices or features as shown in the figures. It should be understood that spatially relative terms are intended to include different orientations of the device in use or operation in addition to the orientation described in the figures. For example, if the device in the drawings is inverted, the device described as "above other devices or structures" or "above other devices or structures" will be positioned as "below other devices or structures" or "below other devices or structures". Thus, the exemplary term "above" can include both "above" and "below". The device can also be positioned in other different ways (rotated 90 degrees or in other orientations), and the spatially relative descriptions used here are interpreted accordingly.

[0060] As attached Figure 1 As shown, the present invention comprises the following steps:

[0061] Step 1: Construct the recognition framework and basic probability distribution function based on the sensor data of the target.

[0062] The number of target categories identified by the input sensor is 3: buildings (a), bridges (b), and high-voltage towers (c). The sensors on the helicopter are lidar and infrared sensors, numbered 1 and 2, respectively, with recognition accuracies of α1 = 0.8 and α2 = 0.7.

[0063] Step 101: Construct a recognition framework for the sensor to identify the target as Θ = {a, b, c}.

[0064] Step 102: Based on the DS evidence theory and the target information a, b identified by the laser radar and infrared sensor, as well as the recognition accuracy α1 = 0.8, α2 = 0.7, the basic probability distribution functions of the laser radar and infrared sensor are constructed as follows: and

[0065] Step 2: Based on the mixed quantum state, convert the basic probability distribution functions m1 and m2 of the sensor into quantum representation. The conversion steps are as follows:

[0066] Step 201: Obtain the quantum representation of the identification frame Θ = {a, b, c}. The quantum representation of Θ is Θ q ={|a>,|b>,|c>}, where |a>=[1,0,0] T 、|b>=[0,1,0] T 、|c>=[0,0,1] T .

[0067] Step 202: Convert the propositions {a}, {b, c}, {b}, {a, c} in the basic probability distribution functions m1 and m2 into quantum representations: |{a}>=|a>, |{b}>=|b>、 After the proposition is converted into quantum representation, the quantum representation of the basic probability distribution functions m1 and m2 is:

[0068]

[0069]

[0070] Step 3: Perform negation transformation on the quantum representations ρ1 and ρ2 of the basic probability distribution functions m1 and m2 to obtain and The negation conversion method is as follows:

[0071] Step 301: Construct a negative generator based on the recognition framework Θ = {a, b, c} of the recognition target. The number of negated generators is 2, which are: and

[0072] Step 302: Calculate the negation of ρ1 and ρ2 based on the constructed negation generator For a given identification framework, after obtaining the negated generators X1 and X2, their corresponding probabilities are p and 1-p. We get the negation of ρ1 and ρ2:

[0073]

[0074]

[0075] Step 4: By and Find the negation of the basic probability distribution functions m1 and m2, expressed as and The calculation steps are as follows:

[0076] Step 401: According to the formula Calculate the probabilities of {a}, {b}, and {c} respectively:

[0077]

[0078]

[0079] It's M θ The conjugate transpose of , tr(·) is the operator for finding the trace of the matrix.

[0080] Step 402: Generate the associated belief function Bel(F) of each proposition, using the formula have to:

[0081]

[0082]

[0083] Step 403: According to the Moss-Ubi transformation Find the negation of m1 and m2 and for:

[0084]

[0085]

[0086]

[0087]

[0088]

[0089]

[0090] Step 5: According to Dempster's combination rule, the negation of the n pieces of evidence obtained in step 4 Perform fusion and obtain fused evidence The Dempster combination rule is in

[0091] Will After fusion, we get the fused evidence for:

[0092] m({a})=0.2066

[0093] m({b})=0.2557

[0094] m({c})=0.3344

[0095] m({a,b})=0.0393

[0096] m({a,c})=0.0656

[0097] m({b,c})=0.0983

[0098] Step 6: Take the basic probability distribution function m after fusion in step 5 and obtain the negation of m according to steps 2, 3 and 4. for:

[0099]

[0100]

[0101]

[0102]

[0103]

[0104]

[0105]

[0106] Step 7: According to the formula |A| is the number of elements in A, and the basic probability distribution function Convert it into probability distribution. According to the obtained probability, make judgments on the recognition results and take P(θ i ) is taken as the result of target recognition.

[0107] According to the formula The basic probability distribution function The probability distribution of is:

[0108] P(a)=0.3672, P(b)=0.34265, P(c)=0.29015

[0109] Since P(a) is the largest in the probability distribution P, the recognition result determines that the target is building (a).

[0110] The above description is merely an embodiment of the present invention and does not limit the present invention in any way. Any simple modification, change and equivalent structural change made to the above embodiment based on the technical essence of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A multi-sensor target fusion recognition method based on quantum negation, characterized by Here are the steps: Step 1: Build an identification framework and basic probability distribution function based on the target's sensor data: Given the number of categories L of targets to be identified, the number n of sensors, and the target recognition accuracy α of each sensor, i , the sensors are numbered as 1, 2, ..., n, and the recognition accuracy of each sensor is α1, α2, ..., α i ,...,α n ; Step 101: Construct a recognition framework for the sensor to identify the target as Θ = {θ1, θ2, ..., θ i ,...,θ L }, where θ i , i=1,2,...,L is the target type that the sensor can identify; Step 102: Based on the DS evidence theory and the target type θ identified by each sensor j And the recognition accuracy α of each sensor i , construct the basic probability distribution function of each sensor as in is θ j The complement of j } is {θ2}, then is {θ1,θ3,...,θ i ,...,θ L }, where θ i , i≠2; Step 2: Based on the mixed quantum state, the basic probability distribution function m of the sensor is i Converted into quantum representation: Step 3: The basic probability distribution function m i The quantum representation ρ i Perform a negation transformation to obtain Step 301: Based on the recognition framework of the recognition target Θ={θ1,θ2,...,θ i ,...,θ L } Construct a negation generator to represent the conversion relationship between elements in the identification framework when generating the negation of the basic probability distribution function. This is constructed by establishing a unitary transfer matrix. The construction method is as follows: The negated generator is all n-order square matrices X that satisfy the following conditions i , where n is the size of the given recognition frame: (1) All elements are either 1 or 0; (2) The sum of each column of the matrix is ​​1; (3) The sum of each row of the matrix is ​​1; (4) The main diagonal elements are 0; All n-order square matrices X that meet the above conditions i This is the negation generator constructed to prove that every square matrix X i is satisfied A unitary matrix, where I is the n-th order identity matrix; The number of negated generators is given by: Step 302: Calculate ρ based on the constructed negation generator i Negation For a given identification framework, we get all the negated generators X i After that, the corresponding probability is p i ,satisfy Due to X i is a unitary matrix, so the basic probability distribution function quantum representation ρ i Negation Computed via a unitary transformation of the original density matrix: Step 4: Find the basic probability distribution function m i Negation Step 401: Based on the quantum measurement principle in quantum mechanics, the probability of each event in Θ is calculated using the corresponding quantum measurement operator. For each singleton, its probability is calculated using the following formula: M θ =|θ><θ| is the quantum measurement operator, It's M θ The conjugate transpose matrix of , tr(·) is the operator for finding the trace of the matrix; Step 402: Generate the associated belief function Bel(F) of each proposition: For a proposition F, since Bel(F) is the lower limit of the interval probability in DS theory, under the existing constraints, by calculating P i (F) to derive Bel(F): Step 403: Obtain the basic probability distribution function m of quantum representation according to the Moss-Ubi transform i Negation According to DS theory, The negation of n pieces of evidence is calculated by the following formula: Step 5: According to Dempster's combination rule, the negation of the n pieces of evidence obtained in step 4 is Perform fusion to obtain the basic probability distribution function after fusion Step 6: Replace the basic probability distribution function m of the sensor in step 2 with the basic probability distribution function m after fusion in step 5 i , repeat steps 2 to 4 to find the negation of m Step 7: According to the formula Among them, A is a subset of the identification framework Θ, |A| represents the number of elements in A, and the basic probability distribution function Convert it into probability distribution, make judgment on the recognition result according to the obtained probability result, and take P(θ i ) is taken as the result of target recognition.

2. The multi-sensor target fusion recognition method based on quantum negation according to claim 1 is characterized in that: The step 2 is to transform the basic probability distribution function m of the sensor into i The process of converting to quantum representation is: Step 201: Obtain recognition frame Θ={θ1,θ2,...,θ i ,...,θ L The quantum representation of Θ is a set of complete and mutually orthogonal basis vectors on the complex field, denoted as Θ q ={|θ1>,|θ2>,...,|θ i >,...,|θ L >}, each ground state is represented by a vector |θ i >=[v1,...,v j ,...,v L ] T , for j=i, satisfy v j =1, for j≠i, satisfy v j =0; Step 202: The basic probability distribution function m i The proposition F in is transformed into a quantum representation where z θ is a complex number z θ =α θ +iβ θ , i is an imaginary number and satisfies ||·|| is the modulus of the complex number; we can get After that, the basic probability distribution function m i The quantum representation of 3. The multi-sensor target fusion recognition method based on quantum negation according to claim 1 is characterized in that: The Dempster combination rule is in

Images