A Distributed Consistency Target Passive Location Method Based on Fast Covariance Interaction

Abstract

This application provides a distributed consensus passive target localization method based on fast covariance interaction, comprising: determining the target's current-moment information matrix and information vector based on the first sensor using angle measurement data of the target collected by a first sensor and a second sensor respectively, and the position information of the first sensor and the second sensor at the current moment, and combining a state model, a measurement model, and a square root volume information filtering algorithm, to construct a local information group of the target based on the first sensor; iteratively updating the local information group of the target based on each sensor through communication between adjacent sensors in a distributed layout, and combining a Monte Carlo consensus weight matrix, until a global consensus information group is obtained; and determining the target's current-moment position based on the global consensus information group and a fast covariance interaction fusion rule. This application can improve positioning accuracy and the consistency of distributed positioning.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and more specifically, to a distributed consensus target passive location method based on fast covariance interaction. Background Technology

[0002] For the distributed consensus passive target localization method based on fast covariance interaction of multiple sensors, in target localization applications, compared with active sensing methods, passive sensor networks have strong concealment characteristics, which can significantly reduce the risk of being detected and interfered with by external factors, effectively enhance the survivability of the system, and at the same time ensure the robustness of target position estimation. It is one of the main technologies for achieving concealed target localization in various localization fields.

[0003] However, in the actual operation of a multi-sensor distributed information fusion system, due to objective factors such as differences in sensor deployment locations and inconsistent measurement accuracy, the local estimation results of the target state by each sensor often have deviations, resulting in a decrease in the overall positioning accuracy of the system. Summary of the Invention

[0004] This application provides a distributed consensus target passive localization method based on fast covariance interaction, which can improve localization accuracy and the consistency of distributed localization.

[0005] In a first aspect, embodiments of this application provide a method for passively locating a distributed consensus target based on fast covariance interaction, including:

[0006] Based on the target's angle measurement data at the current moment collected by the first sensor, the position information of the first sensor at the current moment, the target's angle measurement data at the current moment collected by the second sensor, and the position information of the second sensor at the current moment, and combined with the state model, measurement model, and square root volume information filtering algorithm, the target's information matrix and information vector based on the first sensor at the current moment are determined.

[0007] Wherein, the first sensor is any one of a plurality of sensors, the second sensor is the sensor adjacent to the first sensor among the plurality of sensors, and the plurality of sensors are arranged in a distributed manner;

[0008] Based on the target's information matrix and information vector at the current moment using the first sensor, a local information group of the target based on the first sensor is constructed.

[0009] The target is iteratively updated based on the local information group of each sensor through communication between adjacent sensors in a distributed layout, combined with the Monte Carlo consensus weight matrix, until a global consensus information group is obtained.

[0010] Based on the global consensus information group and combined with the fast covariance interaction fusion rule, the position of the target at the current moment is determined.

[0011] In one possible implementation, determining the target's information matrix and information vector based on the first sensor at the current moment, using the target's angle measurement data acquired by the first sensor at the current moment, the position information of the first sensor at the current moment, the target's angle measurement data acquired by the second sensor at the current moment, and the position information of the second sensor at the current moment, combined with a state model, a measurement model, and a square root volume information filtering algorithm, includes:

[0012] Based on the target's state vector at the previous time step, the square root factor of the target's covariance matrix at the previous time step, the state model, the noise covariance matrix of the state model at the previous time step, and in conjunction with the square root volume information filtering algorithm, the target's state vector at the intermediate time step, the square root factor of the target's covariance matrix at the intermediate time step, the information matrix of the target at the intermediate time step, and the information vector are determined.

[0013] Wherein, the intermediate time is the midpoint between the current time and the previous time;

[0014] Based on the target's state vector at the intermediate time, the square root factor of the target's covariance matrix at the intermediate time, the measurement model, the position information of the first sensor at the current time, the measurement covariance matrix of the measurement model at the current time, the target's information matrix at the intermediate time, the angle measurement data of the target collected by the first sensor at the current time, and combined with the square root volume information filtering algorithm, the target's information contribution matrix and information contribution vector based on the first sensor are determined.

[0015] Based on the target's state vector at the intermediate time, the square root factor of the target's covariance matrix at the intermediate time, the measurement model, the position information of the second sensor at the current time, the measurement covariance matrix of the measurement model at the current time, the target's information matrix at the intermediate time, the angle measurement data of the target collected by the second sensor at the current time, and combined with the square root volume information filtering algorithm, the target's information contribution matrix and information contribution vector based on the second sensor are determined.

[0016] Based on the information matrix of the target at an intermediate time, the information contribution matrix of the target based on the first sensor, and the information contribution matrix of the target based on the second sensor, the information matrix of the target based on the first sensor at the current time is determined.

[0017] Based on the information vector of the target at an intermediate time, the information contribution vector of the target based on the first sensor, and the information contribution vector of the target based on the second sensor, the information vector of the target at the current time based on the first sensor is determined.

[0018] In one possible implementation, determining the target's state vector at the intermediate time step, the square root factor of the target's covariance matrix at the previous time step, the state model, the noise covariance matrix of the state model at the previous time step, and in conjunction with the square root volume information filtering algorithm, includes:

[0019] Based on the target's state vector at the previous time step, the square root factor of the target's covariance matrix at the previous time step, and the preset volume points, determine the state sampling vector of the target at each volume step at the previous time step.

[0020] Based on each volume state sampling vector of the target at the previous time step and the state model, determine each first volume state sampling vector of the target at the intermediate time step;

[0021] The state vector of the target at each intermediate time step is determined based on the first volume state sampling vector of the target at each intermediate time step.

[0022] Based on the target's state vector at the intermediate time, each first volume state sampling vector of the target at the intermediate time, and the noise covariance matrix of the state model at the previous time, determine the square root factor of the covariance matrix of the target at the intermediate time.

[0023] Based on the square root factor of the covariance matrix of the target at the intermediate time and the state vector of the target at the intermediate time, the information matrix and information vector of the target at the intermediate time are determined.

[0024] In one possible implementation, determining the target's information contribution matrix and information contribution vector based on the target's state vector at an intermediate time, the square root factor of the target's covariance matrix at an intermediate time, the measurement model, the position information of the first sensor at the current time, the measurement covariance matrix of the measurement model at the current time, the target's information matrix at an intermediate time, the angle measurement data of the target collected by the first sensor at the current time, and in conjunction with the square root volume information filtering algorithm, includes:

[0025] Based on the target's state vector at the intermediate time, the square root factor of the target's covariance matrix at the intermediate time, and the preset volume point, determine each second volume state sampling vector of the target at the intermediate time.

[0026] Based on each second volume state sampling vector of the target at an intermediate time, the position information of the first sensor at the current time, and the measurement model, determine each volume state sampling vector of the target based on the first sensor at the current time.

[0027] Based on the target's current volume state sampling vector at the first sensor, determine the target's predicted angle measurement data at the first sensor at the current time;

[0028] Based on the target's state vector at an intermediate time, each second volume state sampling vector of the target at an intermediate time, each volume state sampling vector of the target based on the first sensor at the current time, and the target's predicted angle measurement data based on the first sensor at the current time, the cross-covariance matrix of the state model and the measurement model is determined.

[0029] Based on the cross-covariance matrix, the measurement covariance matrix of the measurement model at the current time, and the information matrix of the target at an intermediate time, the information contribution matrix of the target based on the first sensor is determined.

[0030] Based on the cross-covariance matrix, the measurement covariance matrix of the measurement model at the current time, the information matrix of the target at the intermediate time, the state vector of the target at the intermediate time, the predicted angle measurement data of the target based on the first sensor at the current time, and the angle measurement data of the target collected by the first sensor at the current time, the information contribution vector of the target based on the first sensor is determined.

[0031] In one possible implementation, the state model is:

[0032]

[0033] in, For the target at the intermediate time, the first The first volume state sampling vector, For the target at the previous time step A volume state sampling vector, Here is the state transition matrix, and T is the sampling time interval.

[0034]

[0035] The state noise transition matrix is... , This is state noise.

[0036] In one possible implementation, the measurement model is:

[0037]

[0038] in, For the target based on the current time of the first sensor, the first... The view angle in each volume state sampling vector For the target based on the current time of the first sensor, the first... The line-of-sight angle in each volume state sampling vector, ( ) represents the position information of the first sensor at the current moment. , , ) is the target at the intermediate time. The location information indicated by the second volume state sampling vector. The noise in the measurement of the line-of-sight tilt angle. This is the measurement noise for the line-of-sight deflection angle.

[0039] Secondly, embodiments of this application provide an electronic device, including a memory and a processor;

[0040] The memory is used to store computer programs;

[0041] The processor is configured to, when executing the computer program, implement the distributed consensus target passive location method based on fast covariance interaction as described in any one of the first aspects.

[0042] Thirdly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the distributed consensus target passive location method based on fast covariance interaction as described in any one of the first aspects.

[0043] The beneficial effects of the distributed consensus target passive localization method based on fast covariance interaction in this application embodiment are:

[0044] Based on the target's angle measurement data and position information at the current moment acquired by the first sensor, and the target's angle measurement data and position information at the current moment acquired by the second sensor, and combined with a state model, a measurement model, and a square root volume information filtering algorithm, the target's information matrix and information vector at the current moment based on the first sensor are determined. Here, the first sensor is any one of multiple sensors, and the second sensor is the one adjacent to the first sensor. The multiple sensors are arranged in a distributed layout. In other words, this process utilizes the positional correlation and complementary angle measurement data of adjacent sensors, combined with the square root volume information filtering algorithm, which possesses high nonlinear estimation accuracy and numerical stability. This avoids the original measurement deviation caused by differences in deployment location and inconsistent measurement accuracy of a single sensor, alleviates the problem of large errors when processing data based on a single sensor, reduces individual bias in local estimation by a single sensor, and provides accurate data for subsequent construction of local information groups.

[0045] Based on the target's information matrix and information vector at the current moment using the first sensor, a local information group of the target based on the first sensor is constructed. This means that the data that has undergone the aforementioned partial processing is integrated into a local information group, giving it a clear data structure.

[0046] By leveraging communication between adjacent sensors in a distributed array of sensors, and combining this with the Monte Carlo consensus weight matrix, the target's local information groups based on each sensor are iteratively updated until a global consensus information group is obtained. Specifically, local information exchange is achieved through communication between adjacent sensors. Utilizing the Monte Carlo consensus weight matrix's ability to adaptively allocate fixed weights based on the sensor communication topology, and through iterative processes, deviations in local information groups caused by deployment differences and inconsistent measurement accuracy between different sensors are gradually eliminated. This process unifies the local information groups, resulting in a globally consensus information group free from inter-sensor bias.

[0047] Based on the global consensus information set and combined with the fast covariance interaction fusion rule, the target's position at the current moment is determined. That is, by integrating the global consensus information set, which eliminates differences between sensors, through the fast covariance interaction fusion rule, the secondary accumulation of biases during the fusion process can be avoided, thus improving the positioning accuracy of multi-sensor distributed information fusion. Furthermore, since each sensor uses the same global consensus information set, the target position determined by each sensor is the same, thereby improving the consistency of distributed positioning. Attached Figure Description

[0048] Figure 1 A flowchart illustrating a distributed consensus target passive localization method based on fast covariance interaction provided in an embodiment of this application;

[0049] Figure 2 A flowchart illustrating a method for determining a target based on the information matrix and information vector of a first sensor at the current moment, as provided in an embodiment of this application.

[0050] Figure 3 A flowchart illustrating the process of determining the target's state vector at an intermediate time step, the square root factor of the target's covariance matrix at an intermediate time step, the target's information matrix at an intermediate time step, and the information vector provided in this application embodiment;

[0051] Figure 4 A flowchart illustrating a method for determining a target based on the information contribution matrix and information contribution vector of a first sensor, as provided in an embodiment of this application;

[0052] Figure 5 This is a schematic diagram of the communication topology of multiple sensors provided in an embodiment of this application;

[0053] Figure 6 The trajectory diagram of the sensor and target motion provided in the embodiments of this application;

[0054] Figure 7 A schematic diagram illustrating the target location estimation accuracy provided in an embodiment of this application;

[0055] Figure 8 A schematic diagram illustrating the target velocity estimation accuracy provided in an embodiment of this application;

[0056] Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0057] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Although some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present invention. It should be understood that the accompanying drawings and embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0058] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0059] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to"; the term "based on" means "at least partially based on"; the term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; and the term "optionally" means "optional embodiments". Definitions of other terms will be given in the following description. It should be noted that the concepts of "first," "second," etc., mentioned in this invention are used only to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0060] It should be noted that the terms "a" and "a plurality of" used in this invention are illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0061] The names of the messages or information exchanged between the multiple devices in the embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of these messages or information.

[0062] Figure 1 A flowchart illustrating a distributed consensus target passive localization method based on fast covariance interaction provided in this application embodiment is shown below. Figure 1 As shown, the method may include the following steps:

[0063] 110. Based on the target's angle measurement data at the current moment acquired by the first sensor, the position information of the first sensor at the current moment, the target's angle measurement data at the current moment acquired by the second sensor, and the position information of the second sensor at the current moment, and combined with the state model, measurement model, and square root volume information filtering algorithm, determine the target's information matrix and information vector based on the first sensor at the current moment.

[0064] The first sensor is any one of the multiple sensors, the second sensor is the sensor adjacent to the first sensor among the multiple sensors, and the multiple sensors are arranged in a distributed manner.

[0065] For example, a target can be located using multiple aircraft. Specifically, multiple aircraft correspond to multiple sensors, with each sensor mounted on its respective aircraft. Each sensor is used to collect angle measurement data of the target at preset time intervals. The angle measurement data may include the target's line-of-sight tilt angle and line-of-sight deflection angle.

[0066] It should be noted that this method can be executed by the aircraft corresponding to any one of the multiple sensors, and this application does not impose any special limitations on this. Below, throughout the text, the method provided in this application will be exemplarily described using the aircraft corresponding to the first sensor as an example.

[0067] For example, such as Figure 2 As shown, the implementation process of 110 can be described as follows:

[0068] 210. Based on the target's state vector at the previous time step, the square root factor of the target's covariance matrix at the previous time step, the state model, the noise covariance matrix of the state model at the previous time step, and combined with the square root volume information filtering algorithm, determine the target's state vector at the intermediate time step, the square root factor of the target's covariance matrix at the intermediate time step, the target's information matrix at the intermediate time step, and the information vector.

[0069] The intermediate time is the point in time between the current time and the previous time.

[0070] The target's state vector at the previous moment may include at least one of the target's position, velocity, and acceleration at the previous moment, and this application does not impose any special limitations on this.

[0071] For example, , Let be the target's state vector at the previous time step, ( , , ) represents the position of the target at the previous moment. , , ) represents the velocity of the target at the previous moment.

[0072] The square root factor of the target's covariance matrix at the previous time step can at least indicate the error range of the target's state estimate at the previous time step.

[0073] State models are used to indicate the pattern of state changes of a target over time.

[0074] The noise covariance matrix of the state model at the previous time step can be used to indicate the characteristics of noise during the state transition process in the time dimension.

[0075] The information matrix of the target at intermediate time can be used to indicate the degree of certainty of each component in the target's state vector at intermediate time.

[0076] The information vector of the target at intermediate time can indicate the optimal estimation trend of each component in the target's state vector at intermediate time.

[0077] For example, such as Figure 3As shown, the specific implementation of 210 can be as follows:

[0078] 310. Based on the target's state vector at the previous time step, the square root factor of the target's covariance matrix at the previous time step, and the preset volume points, determine the state sampling vector of each volume at the previous time step.

[0079] Specifically, it can be determined using the following formula:

[0080] (1)

[0081] in, For the target at the previous moment (i.e. ) A volume state sampling vector, Let be the state vector of the target at the previous time step. The square root factor of the covariance matrix of the target at the previous time step. , For the state dimension, , For volume point, Indicator set of For ,have .

[0082] 320. Based on the target's state sampling vector and state model at each volume in the previous time step, determine the target's first volume sampling vector at each intermediate time step.

[0083] Specifically, the state model can be:

[0084] (2)

[0085] in, For the target at the intermediate time (i.e. ) The first volume state sampling vector, For the target at the previous moment A volume state sampling vector, Here is the state transition matrix, and T is the sampling time interval.

[0086]

[0087] The state noise transition matrix is... , This is state noise.

[0088] 330. Determine the target's state vector at each intermediate time step based on the target's first volume state sampling vector at each intermediate time step.

[0089] Specifically, this can be achieved using the following formula:

[0090] (3)

[0091] in, , Let be the state vector of the target at the intermediate time.

[0092] 340. Based on the target's state vector at the intermediate time, the sampled state vector of each first volume at the intermediate time, and the noise covariance matrix of the state model at the previous time, determine the square root factor of the target's covariance matrix at the intermediate time.

[0093] Specifically, this can be achieved using the following formula:

[0094] (4)

[0095] In the formula, Let the square root factor of the covariance matrix of the target at the intermediate time be . Let be the noise covariance matrix of the state model at the previous time step. , To calculate the corresponding square root using the QR decomposition method, For a weighted centrality matrix, .

[0096] 350. Based on the square root factor of the covariance matrix of the target at the intermediate time and the state vector of the target at the intermediate time, determine the information matrix and information vector of the target at the intermediate time.

[0097] Specifically, this can be achieved using the following formula:

[0098] (5)

[0099] (6)

[0100] in, The information matrix of the target at the intermediate time. Let be the information vector of the target at the intermediate time.

[0101] 220. Based on the target's state vector at the intermediate time, the square root factor of the target's covariance matrix at the intermediate time, the measurement model, the position information of the first sensor at the current time, the measurement covariance matrix of the measurement model at the current time, the target's information matrix at the intermediate time, the angle measurement data of the target collected by the first sensor at the current time, and combined with the square root volume information filtering algorithm, the target's information contribution matrix and information contribution vector based on the first sensor are determined.

[0102] Measurement models can be used to indicate the relationship between a target's state and observational data. Observational data can be used to indicate the target's line-of-sight tilt angle and line-of-sight deflection angle, and the target state can include at least the target's position.

[0103] The measurement covariance matrix of the measurement model at the current moment can be used to indicate the noise of the sensor during the measurement process.

[0104] The target's angle measurement data collected by the first sensor at the current moment includes the target's line-of-sight tilt angle and line-of-sight deflection angle at the current moment.

[0105] The target's contribution matrix based on the first sensor can be used to indicate the incremental structure of state determinism;

[0106] The target-based information contribution vector from the first sensor can be used to indicate the incremental value of the weighted observations.

[0107] For example, such as Figure 4 As shown, the specific implementation of 220 can be as follows:

[0108] 410. Based on the target's state vector at the intermediate time, the square root factor of the target's covariance matrix at the intermediate time, and the preset volume point, determine the second volume state sampling vector of the target at the intermediate time.

[0109] Specifically, this can be achieved using the following formula:

[0110] (7)

[0111] Among them, here For the target at the intermediate time... The second volume state sampling vector.

[0112] 420. Based on each second volume state sampling vector of the target at the intermediate time, the position information of the first sensor at the current time, and the measurement model, determine each volume state sampling vector of the target based on the first sensor at the current time.

[0113] Specifically, the measurement model can be:

[0114] (8)

[0115] in, The target is based on the current moment of the first sensor. The The view angle in each volume state sampling vector For the target based on the current time of the first sensor, the first... The line-of-sight angle in each volume state sampling vector, ( ) represents the position information of the first sensor at the current moment. , , ) is the target at the intermediate time. The location information indicated by the second volume state sampling vector. The noise in the measurement of the line-of-sight tilt angle. This is the measurement noise for the line-of-sight deflection angle.

[0116] 430. Based on the target's current volume state sampling vector at each moment based on the first sensor, determine the target's predicted angle measurement data based on the first sensor at the current moment.

[0117] Specifically, it can be obtained through the following formula:

[0118] (9)

[0119] in, For the target based on the current time of the first sensor, the first... A volume state sampling vector, include and , The target is predicted based on the current angle measurement data from the first sensor.

[0120] 440. Based on the target's state vector at the intermediate time, the target's state sampling vector for each second volume at the intermediate time, the target's state sampling vector for each volume at the current time based on the first sensor, and the target's predicted angle measurement data for the current time based on the first sensor, determine the cross-covariance matrix between the state model and the measurement model.

[0121] Specifically, this can be achieved using the following formula:

[0122] (10)

[0123] in, The cross-covariance matrix of the state model and the measurement model. Here For the target at the intermediate time... The second volume state sampling vector, .

[0124] 450. Based on the cross-covariance matrix, the measurement covariance matrix of the measurement model at the current time, and the information matrix of the target at intermediate time, determine the information contribution matrix of the target based on the first sensor.

[0125] Specifically, this can be achieved using the following formula:

[0126] (11)

[0127] in, This is the measurement covariance matrix of the measurement model at the current time. The contribution matrix of information based on the first sensor is used to define the target.

[0128] 460. Based on the cross-covariance matrix, the measurement covariance matrix of the measurement model at the current time, the information matrix of the target at the intermediate time, the state vector of the target at the intermediate time, the predicted angle measurement data of the target based on the first sensor at the current time, and the angle measurement data of the target collected by the first sensor at the current time, determine the information contribution vector of the target based on the first sensor.

[0129] Specifically, this can be achieved using the following formula:

[0130] (12)

[0131] in, The contribution vector of the target based on the information from the first sensor. , The angle measurement data of the target at the current moment is collected by the first sensor.

[0132] 230. Based on the target's state vector at the intermediate time, the square root factor of the target's covariance matrix at the intermediate time, the measurement model, the position information of the second sensor at the current time, the measurement covariance matrix of the measurement model at the current time, the target's information matrix at the intermediate time, the angle measurement data of the target collected by the second sensor at the current time, and combined with the square root volume information filtering algorithm, the target's information contribution matrix and information contribution vector based on the second sensor are determined.

[0133] The concepts involved in this step can be found in the explanations of the corresponding concepts above, and will not be repeated here. Furthermore, the implementation method of this step is similar to that of 220; please refer to the implementation steps of 220 for details, which will not be repeated here.

[0134] 240. Based on the target's information matrix at the intermediate time, the target's information contribution matrix based on the first sensor, and the target's information contribution matrix based on the second sensor, determine the target's information matrix based on the first sensor at the current time.

[0135] Specifically, it can be obtained through the following formula:

[0136] (13)

[0137] in, The target is based on the information matrix of the first sensor at the current moment. The information contribution matrix of the target based on the j-th sensor. , It includes a first sensor and a second sensor.

[0138] The target information matrix based on the current moment of the first sensor can be used to indicate the amount of information or degree of certainty contained in the state estimation.

[0139] 250. Based on the target's information vector at the intermediate moment, the target's information contribution vector based on the first sensor, and the target's information contribution vector based on the second sensor, determine the target's information vector at the current moment based on the first sensor.

[0140] Specifically, it can be obtained through the following formula:

[0141] (14)

[0142] in, The target is based on the information vector from the first sensor at the current moment. The information contribution vector of the target based on the j-th sensor.

[0143] The target information vector based on the current moment of the first sensor can be used to indicate the optimal estimate of the state variables under the current information level or degree of determinism.

[0144] 120. Based on the target's information matrix and information vector at the current moment using the first sensor, construct a local information group of the target using the first sensor.

[0145] For example, the target based on the local information group of the first sensor is ( , , ).

[0146] 130. By communicating with adjacent sensors in a distributed array of sensors and combining the Monte Carlo consensus weight matrix, the target is iteratively updated based on the local information group of each sensor until a global consensus information group is obtained.

[0147] Specifically, each sensor in the distributed layout communicates with its neighboring sensors, and during the communication process, a Metropolis consistency weight matrix is ​​introduced. Through multiple iterations This allows the local information groups of each sensor to gradually spread and merge between adjacent nodes, continuously correcting the local information groups of each sensor.

[0148] in: , Let be the degree of the node.

[0149] As the number of iterations increases, the local information groups of all sensors will gradually converge, eventually reaching a unified global consensus information group, which can be represented as ( , , ),in, , , , Indicates the sensor number.

[0150] 140. Based on the global consensus information group and combined with the fast covariance interaction fusion rule, determine the target's position at the current moment.

[0151] Specifically, this can be achieved using the following formula:

[0152] (15)

[0153] (16)

[0154] (17)

[0155] in, Let be the target's state vector at the current moment, which includes the target's position at the current moment. The information matrix is ​​obtained after fast covariance interaction fusion. This is the information vector after fast covariance interaction fusion.

[0156] It should be noted that here, The sensor number indicates that the data was calculated based on the data from the i-th sensor.

[0157] The derivation process of formulas (15) and (16) will be explained below.

[0158] Given a sensor among multiple sensors Information vector and information matrix The fast covariance interaction criterion is:

[0159] (18)

[0160] (19)

[0161] in,

[0162] Formulas (15) and (16) can be obtained by transforming the two formulas above in the following way.

[0163] (20)

[0164] (twenty one)

[0165] In summary, based on the target's angle measurement data and position information at the current moment acquired by the first sensor, and the target's angle measurement data and position information at the current moment acquired by the second sensor, and combined with the state model, measurement model, and square root volume information filtering algorithm, the target's information matrix and information vector at the current moment based on the first sensor are determined. Here, the first sensor is any one of multiple sensors, and the second sensor is the sensor adjacent to the first sensor, with the multiple sensors arranged in a distributed layout. In other words, this process utilizes the positional correlation and complementary angle measurement data of adjacent sensors, combined with the square root volume information filtering algorithm which possesses high nonlinear estimation accuracy and numerical stability. This avoids the original measurement deviation caused by differences in deployment location and inconsistent measurement accuracy of a single sensor, alleviates the problem of large errors when processing data based on a single sensor, reduces individual bias in local estimation by a single sensor, and provides accurate data for subsequent construction of local information groups.

[0166] Based on the target's information matrix and information vector at the current moment using the first sensor, a local information group of the target based on the first sensor is constructed. This means that the data that has undergone the aforementioned partial processing is integrated into a local information group, giving it a clear data structure.

[0167] By leveraging communication between adjacent sensors in a distributed array of sensors, and combining this with the Monte Carlo consensus weight matrix, the target's local information groups based on each sensor are iteratively updated until a global consensus information group is obtained. Specifically, local information exchange is achieved through communication between adjacent sensors. Utilizing the Monte Carlo consensus weight matrix's ability to adaptively allocate fixed weights based on the sensor communication topology, and through iterative processes, deviations in local information groups caused by deployment differences and inconsistent measurement accuracy between different sensors are gradually eliminated. This process unifies the local information groups, resulting in a globally consensus information group free from inter-sensor bias.

[0168] Based on the global consensus information set and combined with the fast covariance interaction fusion rule, the target's position at the current moment is determined. That is, by integrating the global consensus information set, which eliminates differences between sensors, through the fast covariance interaction fusion rule, the secondary accumulation of biases during the fusion process can be avoided, thus improving the positioning accuracy of multi-sensor distributed information fusion. Furthermore, since each sensor uses the same global consensus information set, the target position determined by each sensor is the same, thereby improving the consistency of distributed positioning.

[0169] Below, we take the distributed consensus target passive localization based on fast covariance interaction using eight sensor platforms as an example. Each sensor can measure the target's pitch and azimuth angles. The initial positions of the eight UAVs are shown in Table 1, and the communication topology is as follows: Figure 5 As shown.

[0170]

[0171] Table 1

[0172] The motion model for the multi-sensor system is a uniform acceleration model. The velocity and acceleration of the sensors are initialized as follows: and The standard deviation of the process noise in the state model was set to... Assume the target's true initial motion state is as follows: The noise covariance matrix of the state model at the initial time. for: The covariance matrix of the target's state vector at the initial moment. The covariance of the target's state vector at the initial time is used to determine the square root factor of the target's covariance matrix at the initial time. The measurement covariance matrix of the measurement model at the initial time... for , The fusion cycle is... The simulation time is 300 Monte Carlo simulations were performed.

[0173] Simulation results Figures 6-8 As shown, Figure 6 This is a trajectory diagram of the sensor and target motion. Figure 7 A schematic diagram illustrating the accuracy of target location estimation. Figure 8 A schematic diagram illustrating the accuracy of target velocity estimation. Through analysis... Figures 6-8As can be seen, the method proposed in this application can track the target trajectory very well, with an average positioning accuracy (RMSE) of 746m, which is better than the local estimate of 1183m, and eventually stabilizes at less than 500m. Similarly, the average velocity estimation accuracy (RMSE) is 28m / s, which is better than the local estimate of 34m / s.

[0174] like Figure 9 As shown, an electronic device 900 provided in this embodiment of the invention may include a processor 910 and a memory 920; the memory 920 is used to store a computer program; the processor 910 is used to implement, when executing the computer program, a distributed consensus target passive localization method based on fast covariance interaction as described above.

[0175] This invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements a distributed consensus target passive location method based on fast covariance interaction as described above.

[0176] The present invention will now describe an electronic device 900 that can serve as a server or client of the present invention, which is an example of a hardware device that can be applied to various aspects of the present invention. Electronic device 900 is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0177] Electronic device 900 includes a computing unit that can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) or a computer program loaded from a storage unit into random access memory (RAM). The RAM may also store various programs and data required for device operation. The computing unit, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0178] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. In this application, the units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of the present invention according to actual needs. Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units can be implemented in hardware or as software functional units.

[0179] While the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A distributed consensus target passive localization method based on fast covariance interaction, characterized in that, include: Based on the target's angle measurement data at the current moment collected by the first sensor, the position information of the first sensor at the current moment, the target's angle measurement data at the current moment collected by the second sensor, and the position information of the second sensor at the current moment, and combined with the state model, measurement model, and square root volume information filtering algorithm, the target's information matrix and information vector based on the first sensor at the current moment are determined. Wherein, the first sensor is any one of a plurality of sensors, the second sensor is the sensor adjacent to the first sensor among the plurality of sensors, and the plurality of sensors are arranged in a distributed manner; Based on the target's information matrix and information vector at the current moment using the first sensor, a local information group of the target based on the first sensor is constructed. The target is iteratively updated based on the local information group of each sensor through communication between adjacent sensors in a distributed layout, combined with the Monte Carlo consensus weight matrix, until a global consensus information group is obtained. Based on the global consensus information group and combined with the fast covariance interaction fusion rule, the position of the target at the current moment is determined; The step of determining the target's information matrix and information vector based on the first sensor at the current moment, using the target's angle measurement data acquired by the first sensor at the current moment, the target's position information acquired by the second sensor at the current moment, and combining the state model, measurement model, and square root volume information filtering algorithm, includes: Based on the target's state vector at the previous time step, the square root factor of the target's covariance matrix at the previous time step, the state model, the noise covariance matrix of the state model at the previous time step, and in conjunction with the square root volume information filtering algorithm, the target's state vector at the intermediate time step, the square root factor of the target's covariance matrix at the intermediate time step, the information matrix of the target at the intermediate time step, and the information vector are determined. Wherein, the intermediate time is the midpoint between the current time and the previous time; Based on the target's state vector at the intermediate time, the square root factor of the target's covariance matrix at the intermediate time, the measurement model, the position information of the first sensor at the current time, the measurement covariance matrix of the measurement model at the current time, the target's information matrix at the intermediate time, the angle measurement data of the target collected by the first sensor at the current time, and combined with the square root volume information filtering algorithm, the target's information contribution matrix and information contribution vector based on the first sensor are determined. Based on the target's state vector at the intermediate time, the square root factor of the target's covariance matrix at the intermediate time, the measurement model, the position information of the second sensor at the current time, the measurement covariance matrix of the measurement model at the current time, the target's information matrix at the intermediate time, the angle measurement data of the target collected by the second sensor at the current time, and combined with the square root volume information filtering algorithm, the target's information contribution matrix and information contribution vector based on the second sensor are determined. Based on the information matrix of the target at an intermediate time, the information contribution matrix of the target based on the first sensor, and the information contribution matrix of the target based on the second sensor, the information matrix of the target based on the first sensor at the current time is determined. Based on the information vector of the target at an intermediate time, the information contribution vector of the target based on the first sensor, and the information contribution vector of the target based on the second sensor, the information vector of the target at the current time based on the first sensor is determined.

2. The method according to claim 1, characterized in that, The step of determining the target's state vector at the intermediate time step, the square root factor of the target's covariance matrix at the previous time step, the state model, the noise covariance matrix of the state model at the previous time step, and in conjunction with the square root volume information filtering algorithm, includes: Based on the target's state vector at the previous time step, the square root factor of the target's covariance matrix at the previous time step, and the preset volume points, determine the state sampling vector of the target at each volume step at the previous time step. Based on each volume state sampling vector of the target at the previous time step and the state model, determine each first volume state sampling vector of the target at the intermediate time step; The state vector of the target at each intermediate time step is determined based on the first volume state sampling vector of the target at each intermediate time step. Based on the target's state vector at the intermediate time, each first volume state sampling vector of the target at the intermediate time, and the noise covariance matrix of the state model at the previous time, determine the square root factor of the covariance matrix of the target at the intermediate time. Based on the square root factor of the covariance matrix of the target at the intermediate time and the state vector of the target at the intermediate time, the information matrix and information vector of the target at the intermediate time are determined.

3. The method according to claim 1, characterized in that, The step of determining the target's information contribution matrix and information contribution vector based on the target's state vector at the intermediate time, the square root factor of the target's covariance matrix at the intermediate time, the measurement model, the position information of the first sensor at the current time, the measurement covariance matrix of the measurement model at the current time, the target's information matrix at the intermediate time, and the angle measurement data of the target collected by the first sensor at the current time, combined with the square root volume information filtering algorithm, includes: Based on the target's state vector at the intermediate time, the square root factor of the target's covariance matrix at the intermediate time, and the preset volume point, determine each second volume state sampling vector of the target at the intermediate time. Based on each second volume state sampling vector of the target at an intermediate time, the position information of the first sensor at the current time, and the measurement model, determine each volume state sampling vector of the target based on the first sensor at the current time. Based on the target's current volume state sampling vector at the first sensor, determine the target's predicted angle measurement data at the first sensor at the current time; Based on the target's state vector at an intermediate time, each second volume state sampling vector of the target at an intermediate time, each volume state sampling vector of the target based on the first sensor at the current time, and the target's predicted angle measurement data based on the first sensor at the current time, the cross-covariance matrix of the state model and the measurement model is determined. Based on the cross-covariance matrix, the measurement covariance matrix of the measurement model at the current time, and the information matrix of the target at an intermediate time, the information contribution matrix of the target based on the first sensor is determined. Based on the cross-covariance matrix, the measurement covariance matrix of the measurement model at the current time, the information matrix of the target at the intermediate time, the state vector of the target at the intermediate time, the predicted angle measurement data of the target based on the first sensor at the current time, and the angle measurement data of the target collected by the first sensor at the current time, the information contribution vector of the target based on the first sensor is determined.

4. The method according to claim 2, characterized in that, The state model is as follows: in, For the target at the intermediate time, the first The first volume state sampling vector, For the target at the previous time step A volume state sampling vector, Here is the state transition matrix, and T is the sampling time interval. The state noise transition matrix is... , This is state noise.

5. The method according to claim 3, characterized in that, The measurement model is as follows: in, For the target based on the current time of the first sensor, the first... The view angle in each volume state sampling vector For the target based on the current time of the first sensor, the first... The line-of-sight angle in each volume state sampling vector, ( ) represents the position information of the first sensor at the current moment. , , ) is the target at the intermediate time. The location information indicated by the second volume state sampling vector. The noise in the measurement of the line-of-sight tilt angle. This is the measurement noise for the line-of-sight deflection angle.

6. An electronic device, characterized in that, Including memory and processor; The memory is used to store computer programs; The processor is configured to, when executing the computer program, implement a distributed consensus target passive location method based on fast covariance interaction as described in any one of claims 1 to 5.

7. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements a distributed consensus target passive location method based on fast covariance interaction as described in any one of claims 1 to 5.

Images