A method and system for estimating time-varying ocean currents based on an interactive Kalman model

Abstract

The present invention provides a method and system for estimating time-varying ocean currents based on an interactive Kalman model, the method comprising: establishing a preset number of ocean current sub-models according to the acquired ocean current information of the target area; constructing corresponding linear Kalman sub-filters according to each ocean current sub-model; obtaining the sub-filter mixing probability according to the navigation parameters and covariance matrix output by each linear Kalman sub-filter, and obtaining the fused ocean current information of each linear Kalman sub-filter according to the sub-filter mixing probability; obtaining the fused filter output according to the fused ocean current information of each linear Kalman sub-filter, and obtaining the time-varying ocean current estimation according to the fused filter output. Based on the ocean current multi-model theory, the present invention mixes and superimposes a plurality of sub-filter outputs according to the interactive multi-model, effectively improving the estimation accuracy of complex ocean currents while providing reliable protection for high-precision combined navigation systems.

Description

Technical Field

[0001] The present invention relates to the field of information measurement technology, and in particular to a method and system for estimating time-varying ocean currents based on an interactive Kalman model. Background Art

[0002] Autonomous underwater vehicles (AUVs) play an important role in many fields, and navigation and positioning technology is the core technology that determines the success or failure of underwater missions. The existing methods of combined navigation using medium- and high-precision strapdown inertial navigation systems (SINS), auxiliary sensors, and global navigation satellite systems (GNSS) can improve navigation accuracy and suppress positioning errors to a certain extent, but there is still a fatal drawback: the error of the damped horizontal channel requires the ground velocity of the carrier to be used as the observation of the linear Kalman filter (LFK). When the distance from the bottom of the water exceeds the maximum bottom-penetrating distance, the Doppler velocimeter (DVL) and the acoustic correlation velocimeter (ACL) can only measure the water velocity, but not the ground velocity. When there is no ocean current, the water velocity is the ground velocity. However, for underwater environments with complex time-varying ocean currents, if the influence of ocean currents is not considered, combined navigation using the water velocity will cause large positioning errors.

[0003] Although the commonly used ocean current estimation algorithms improve positioning accuracy by adding ocean current estimation to integrated navigation, they all have application defects and cannot be applied to ocean current estimation in unfamiliar waters such as deep seas. For example, the method of estimating the size of ocean currents using DVL water velocity information is: assuming that the ocean current is steady or slowly changing, a Kalman model containing ocean current information is established. When the ocean current is a complex time-varying ocean current, this method will fail. Inaccurate ocean current estimation will cause huge positioning errors, and sometimes the error even exceeds the conventional model that does not consider ocean current factors; 2) When using the Kalman model based on time-varying ocean currents for estimation, it is necessary to frequently receive GNSS signals to estimate the ocean currents. The parameters of the ocean current model can only describe simple time-varying ocean currents and the parameters need to be obtained by looking up tables. It requires a high level of knowledge of weather and environmental information, and the application of the method has great limitations. Summary of the invention

[0004] The purpose of the present invention is to provide a time-varying ocean current estimation method based on an interactive Kalman model. Based on the ocean current multi-model theory, several sub-filter outputs are mixed and superimposed according to the interactive multi-model to obtain a time-varying ocean current estimate, thereby solving the application defects of the ocean current estimation method in the existing combined navigation. While effectively improving the estimation accuracy of complex ocean currents, it also provides reliable protection for high-precision combined navigation systems.

[0005] In order to achieve the above objectives, it is necessary to provide a time-varying ocean current estimation method and system based on an interactive Kalman model in response to the above technical problems.

[0006] In a first aspect, an embodiment of the present invention provides a method for estimating time-varying ocean currents based on an interactive Kalman model, the method comprising the following steps:

[0007] Establishing a preset number of ocean current sub-models based on the acquired ocean current information of the target area; the ocean current information of the target area includes ocean current observation data of different change periods;

[0008] According to each ocean current sub-model, a corresponding linear Kalman sub-filter is constructed; the measurement matrix of the linear Kalman sub-filter includes a pure inertial navigation measurement sub-matrix and an underwater speed measurement sub-matrix;

[0009] According to the navigation parameters and covariance matrix output by each linear Kalman sub-filter, a sub-filter mixing probability is obtained, and according to the sub-filter mixing probability, fused ocean current information of each linear Kalman sub-filter is obtained; the fused ocean current information includes fused navigation parameters, fused variance matrix, fused ocean current model time parameters and fused ocean current model variance parameters;

[0010] According to the fused ocean current information of each linear Kalman sub-filter, a fused filter output is obtained, and according to the fused filter output, a time-varying ocean current estimation is obtained.

[0011] Furthermore, the ocean current sub-model is a first-order Markov ocean current model; the ocean current sub-model is expressed as:

[0012]

[0013] in, and represent the ocean current values ​​corresponding to the k+1th and kth sampling moments respectively; T C The time constant representing the relevant interval of ocean current velocity; σ C represents the root mean square of the ocean current velocity; Δt represents the time interval between data generation; w C Represents white noise that conforms to a standard normal distribution.

[0014] Furthermore, the step of establishing a preset number of ocean current sub-models according to the ocean current information of the target area includes:

[0015] According to the ocean current information of the target area, a preset number of sub-model parameters are determined with the minimum fitting error as the optimization goal; the sub-model parameters include time parameters and variance parameters;

[0016] According to the sub-model parameters, a corresponding ocean current sub-model is constructed.

[0017] Furthermore, the step of constructing a corresponding linear Kalman sub-filter according to each ocean current sub-model includes:

[0018] The north-eastern coordinate system is used as the navigation system, and the corresponding ocean current estimation error is obtained according to each ocean current sub-model; the ocean current estimation error is expressed as:

[0019]

[0020] in, and represent the ocean current estimation errors at the k-1th and kth moments, respectively; and represent the ocean current velocity and ideal ocean current velocity based on the linear Kalman subfilter and ocean current submodel, respectively; T i C represents the time parameter of the ith ocean current submodel; I represents the total number of ocean current submodels;

[0021] The ocean current estimation error is used as a state variable, and the corresponding navigation parameter error is determined; the navigation parameter error is expressed as:

[0022]

[0023] Among them, δL and δl represent the latitude error and longitude error respectively; δv N and δv E is the velocity component in the navigation system n; and Respectively represent the roll angle error, pitch angle error and heading angle error; and represents the zero bias of the three accelerometers in the b system; ε x , ε y and ε z Indicates the three gyro biases in the b system; and They represent the northward ocean current error and the eastward ocean current error respectively;

[0024] According to the real-time pure inertial navigation measurement, underwater velocity measurement and ocean current vector, a measurement equation is constructed; the measurement equation is expressed as:

[0025]

[0026] in, and They represent the calculated n-to-b attitude conversion matrix, accurate velocity vector, accurate ocean current vector and underwater velocity measurement output respectively;

[0027] A corresponding linear Kalman sub-filter is obtained according to the navigation parameter error and the measurement equation.

[0028] Furthermore, the step of obtaining the sub-filter mixing probability according to the navigation parameters and covariance matrix output by each linear Kalman sub-filter includes:

[0029] According to the navigation parameters and covariance matrix output by each linear Kalman sub-filter, the corresponding model credibility probability is obtained; the model credibility probability is expressed as:

[0030]

[0031] In the formula,

[0032]

[0033]

[0034] Among them, μ i,k represents the model credibility probability of the i-th linear Kalman subfilter at the k-th time; n i,k-1 represents the prediction probability of the i-th linear Kalman subfilter at the k-1th time; M represents the Markov chain transmission matrix; r i , C i and λ i denote the residual, the corresponding covariance matrix and the maximum likelihood ratio output by the i-th linear Kalman subfilter, respectively; I denotes the total number of linear Kalman subfilters;

[0035] According to the credible probabilities of each model, the sub-filter mixing probability is obtained; the sub-filter mixing probability is expressed as:

[0036]

[0037] Among them, g 11,k ,...,g 1I,k represents the superposition weight required to calculate the output of the first linear Kalman subfilter at the kth time; g I1,k, ... II,krepresents the superposition weight required to calculate the output of the I-th linear Kalman subfilter at the k-th time; m ij represents the probability of switching from the i-th linear Kalman subfilter to the j-th linear Kalman subfilter.

[0038] Furthermore, the steps of obtaining a fusion filter output according to the fusion ocean current information of each linear Kalman sub-filter, and obtaining a time-varying ocean current estimation according to the fusion filter output include:

[0039] According to the model credibility probability and fusion navigation parameters of each linear Kalman sub-filter, the fusion filter output is obtained; the fusion filter output is expressed as:

[0040]

[0041] in, and μ i,k They represent the fusion navigation parameters and model credibility probability of the i-th linear Kalman subfilter at the k-th moment respectively;

[0042] Obtaining an ideal ocean current value according to the fused ocean current model time parameter and the fused ocean current model variance parameter;

[0043] The ocean current estimation error in the fusion filter output is compensated to the ideal ocean current value to obtain the time-varying ocean current estimate.

[0044] Furthermore, after the step of obtaining the time-varying ocean current estimation according to the fusion filter output, the method further includes:

[0045] The fusion filter output is compensated to the pure inertial solution navigation parameters acquired in real time to obtain the estimated value of the combined navigation parameters.

[0046] In a second aspect, an embodiment of the present invention provides a time-varying ocean current estimation system based on an interactive Kalman model, the system comprising:

[0047] A model building module, used to establish a preset number of ocean current sub-models based on the acquired ocean current information of the target area; the ocean current information of the target area includes ocean current observation data of different change periods;

[0048] A filter construction module is used to construct a corresponding linear Kalman sub-filter according to each ocean current sub-model; the measurement matrix of the linear Kalman sub-filter includes a pure inertial navigation measurement sub-matrix and an underwater speed measurement sub-matrix;

[0049] A probability fusion module is used to obtain a sub-filter mixing probability according to the navigation parameters and covariance matrix output by each linear Kalman sub-filter, and obtain fused ocean current information of each linear Kalman sub-filter according to the sub-filter mixing probability; the fused ocean current information includes fused navigation parameters, fused variance matrix, fused ocean current model time parameters and fused ocean current model variance parameters;

[0050] The ocean current estimation module is used to obtain a fusion filter output according to the fusion ocean current information of each linear Kalman sub-filter, and obtain a time-varying ocean current estimation according to the fusion filter output.

[0051] In a third aspect, an embodiment of the present invention further provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

[0052] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above method when executed by a processor.

[0053] The present application provides a method and system for estimating time-varying ocean currents based on an interactive Kalman model. Through the method, a preset number of ocean current sub-models are established based on the ocean current information of the target area based on the acquired ocean current observation data including different change periods, and the corresponding linear Kalman sub-filters are constructed. Then, according to the navigation parameters and covariance matrix output by each linear Kalman sub-filter, the sub-filter mixing probability is obtained, and according to the sub-filter mixing probability, the fused ocean current information of each linear Kalman sub-filter including the fused navigation parameter, the fused variance matrix, the fused ocean current model time parameter and the fused ocean current model variance parameter is obtained. Then, according to the fused ocean current information of each linear Kalman sub-filter, the fused filter output is obtained, and according to the fused filter output, the time-varying ocean current estimation is obtained, and the technical scheme of calibrating the combined navigation parameters is provided. Compared with the prior art, the time-varying ocean current estimation method based on the interactive Kalman model, according to the ocean current multi-model theory, the outputs of several sub-filters are mixed and superimposed according to the interactive multi-model, which effectively solves the problem of ocean current estimation in unfamiliar waters such as deep seas, improves the estimation accuracy of complex ocean currents, and provides reliable protection for high-precision combined navigation systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 Schematic diagram of an application framework for time-varying ocean current estimation based on an interactive Kalman model in an embodiment of the present invention;

[0055] Figure 2 is a flow chart of a method for estimating time-varying ocean currents based on an interactive Kalman model in an embodiment of the present invention;

[0056] Figure 3 is another schematic flow chart of a method for estimating time-varying ocean currents based on an interactive Kalman model in an embodiment of the present invention;

[0057] Figure 4 is a schematic structural diagram of a time-varying ocean current estimation system based on an interactive Kalman model in an embodiment of the present invention;

[0058] Figure 5 is another structural schematic diagram of a time-varying ocean current estimation system based on an interactive Kalman model in an embodiment of the present invention;

[0059] Figure 6 It is a diagram of the internal structure of a computer device in an embodiment of the present invention. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical scheme and beneficial effects of the present application clearer, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. Obviously, the embodiments described below are part of the embodiments of the present invention and are only used to illustrate the present invention, but are not used to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] The time-varying ocean current estimation method based on the interactive Kalman model provided by the present invention is based on the technical defect that the existing ocean current estimation method cannot be applied to the ocean current estimation in unfamiliar sea areas such as the deep sea. Figure 1 The time-varying ocean current estimation framework shown in the figure implements a method for calculating the time-varying ocean current based on the ocean current multi-model theory, which mixes and superimposes the outputs of several sub-filters (LKF filters) according to the interactive multi-model (IMM) to obtain accurate and reliable time-varying ocean current estimates and ensures a high-precision time-varying ocean current calculation method for the integrated navigation system. Figure 1 v W 、h DG , and They represent the water velocity output by UVL, the depth output by depth, the specific force measured by the accelerometer, and the angular velocity measured by the gyroscope; The subscript ib in the figure represents the angular velocity of the b system relative to the inertial system i system, that is, the measurement result of the gyroscope; represents the complete navigation parameters (sub-filter output). It is the navigation parameter with ocean current model that needs to be fed back to the sub-filter after IMM fusion calculation; after the depth meter information enters the model, it is processed using the vertical channel damping technology, that is, using the output h of the depth meter DG Instead of the height / depth h calculated by pure inertial solution INSAs the output of the IMM, the accuracy of the height / depth information output by the IMM is guaranteed (for simplicity, Figure 1 The height channel related details are not drawn in the figure); the height / depth information of each sub-filter and the final IMM output is taken as h DG , so h DG There is no need to perform the fusion calculation of IMM, which greatly simplifies the amount of calculation. Since the horizontal channel and the vertical channel in the inertial navigation model are decoupled from each other, this processing method not only provides more accurate depth / altitude information, but also does not increase the amount of calculation of IMM, and does not affect the original accuracy of all navigation parameters of the horizontal channel. The following embodiments will explain in detail the time-varying ocean current estimation method based on the interactive Kalman model of the present invention.

[0062] In one embodiment, Figure 2 As shown, a time-varying ocean current estimation method based on an interactive Kalman model is provided, comprising the following steps:

[0063] S11. Establishing a preset number of ocean current sub-models according to the acquired ocean current information of the target area; the ocean current information of the target area includes ocean current observation data of different change periods; wherein the ocean current observation data can be understood as relevant data obtained from the statistics of the previous ocean surveying and mapping, such as wave height, wave direction and change period, and can also include an empirical reference model of the target area;

[0064] Since ocean currents or sea currents are the flow of seawater caused by the gravitational force of the moon and the sun, their speed and direction are related to time and place, and are affected by many factors such as wind, interaction between water and wind, temperature, salinity, water depth, tidal force, etc., it is difficult to accurately control ocean currents. In order to improve the accuracy of time-varying ocean current estimation, this embodiment preferably uses a first-order Markov ocean current model as the ocean current sub-model used; that is, the ocean current sub-model is a first-order Markov ocean current model, which is specifically expressed as:

[0065]

[0066] in, and represent the ocean current values ​​corresponding to the k+1th and kth sampling moments respectively; T C The time constant representing the ocean current velocity-related interval is generally related to the IMU update period; σ C represents the root mean square of the ocean current velocity; Δt represents the time interval between data generation; w C Represents white noise that conforms to the standard normal distribution; the typical value of the parameter is: T C =0.1-30h,σ C =0.005-5m / s.

[0067] In order to achieve accurate modeling and analysis of ocean currents, this embodiment preferably selects I (generally, I = 2-5) groups of parameters to establish I ocean current sub-models as shown in formula (1), and then weights and superimposes each ocean current sub-model to obtain a higher ocean current fitting accuracy. That is, by statistically analyzing the previous ocean current mapping data of the sea area, combining spectral analysis and other methods with prior knowledge, an ocean current sub-model describing the ocean current in the area is obtained, and each ocean current sub-model contains a set of T C and σ C Parameters, and recursively according to the rule shown in formula (1). Specifically, the step of establishing a preset number of ocean current sub-models according to the ocean current information of the target area includes:

[0068] According to the ocean current information of the target area, a preset number of sub-model parameters are determined with the minimum fitting error as the optimization goal; the sub-model parameters include time parameters and variance parameters;

[0069] According to the sub-model parameters, construct a corresponding ocean current sub-model;

[0070] The selection of the preset number should be determined according to the complexity of the ocean current in the target area of ​​the actual application. Generally, the more complex the ocean current is, the more ocean current sub-models are needed. For example, if the ocean current in the target area is represented by a T C =1h and σ C =0.08m / s ocean current model fitting can achieve very high accuracy (fitting error is small enough), then select the preset number I=1, T C =1h and σ C =0.08m / s; if the ocean current in the target area requires T 1 C = 0.5h and and and =0.2m / s, so that the two sets of ocean current sub-model fitting (linear superposition) can achieve higher accuracy, then the preset number I=2 is selected, T 1 C = 0.5h and and and By analogy, the more complex the ocean current, the larger the I required. That is, the changing patterns of complex ocean currents are complex, and if not enough ocean current sub-models are used, the fitting error will be relatively large.

[0071] In this embodiment, the size of the preset number I and the parameters of each ocean current sub-model need to be determined based on a large amount of ocean current prior information, statistical analysis, nonlinear function fitting calculation, etc. For example, the ocean current data obtained by mapping is recorded as V r , assuming V 1 ,V 2 and V 3There are three known ocean current submodels, and the unknown weights are k 1 , k 2 and k 3 , then V = k 1 ×V 1 +k 2 ×V 2 +k 3 ×V 3 ; Then, use nonlinear function fitting methods (such as simulated annealing, steepest gradient descent, etc.) to optimize the following objective function to obtain the three unknown weights:

[0072]

[0073] In general, V r The changing pattern of is very complicated. If the ocean current sub-model is not suitable, no matter what constant the weight is, it is difficult for the above objective function to get a small value. At this time, it is necessary to adjust the parameters of the ocean current sub-model appropriately. Only when the parameters of the three ocean current sub-models are appropriate can a set of weights be found to make the objective function small enough (smaller than a certain threshold can be considered small enough), and the corresponding ocean current sub-model is suitable.

[0074] It should be noted that if there is no prior information about the target area, we can first assume that I = 1 and a set of parameters are given arbitrarily, and use the minimum fitting error as the objective function to continuously adjust I and the parameters of each ocean current sub-model to obtain more suitable ocean current sub-models;

[0075] S12. According to each ocean current sub-model, a corresponding linear Kalman sub-filter is constructed; wherein the linear Kalman sub-filter can be understood as an LKF filter capable of estimating ocean currents, ocean current errors and navigation parameters, and the measurement matrix of the linear Kalman sub-filter includes a pure inertial navigation measurement sub-matrix (SINS sub-matrix) and an underwater velocity measurement sub-matrix (UVL sub-matrix); it should be noted that UVL is one of DVL, ACL, EML and PT, and the UVL sub-matrix is ​​established based on formula (1); at the same time, based on the measurement information of the depth meter used in the aforementioned height channel for damping, the navigation parameters contained in the SINS sub-matrix of this embodiment can remove the height (or depth) and the celestial velocity, as shown in formula (2):

[0076]

[0077] Where L and l represent latitude and longitude respectively; vector Represents the attitude angle, which includes three components: roll, pitch and heading angle; speed v N and v EIt represents the velocity component in the navigation system n (north-east-earth coordinate system), and the vector composed of it is recorded as V n =[v N v E ] T Similarly, the vectors composed of the northward and eastward currents are recorded as

[0078] Specifically, the step of constructing a corresponding linear Kalman sub-filter according to each ocean current sub-model includes:

[0079] The north-eastern coordinate system is used as the navigation system, and the corresponding ocean current estimation error is obtained according to each ocean current sub-model; the ocean current estimation error is expressed as:

[0080]

[0081] in, and represent the ocean current estimation errors at the k-1th and kth moments, respectively; and They represent the ocean current velocity and ideal ocean current velocity obtained based on the linear Kalman subfilter and ocean current submodel respectively;

[0082] The first equation in (3a) describes the ocean current estimation error, which is given by and Subtract the result. and Both contain the noise term shown in equation (1). After subtraction, the term cancels out, so the first equation in equation (3a) does not contain the noise term. Δt is the time step used in computer calculation, which is generally selected as 1, 2, 5 or 10ms. The "^" above the speed variable in equation (3a) represents the Kalman filter estimate. Assuming there are I sub-filters, then: i∈{1,2,...,I}, the subscript i represents the i-th sub-filter in IMM. From equation (3a), it is easy to get equation (3b):

[0083]

[0084] From the first equation in (3b), we can get the UVL submatrix in the i-th LKF as shown in equation (4), which is a constant matrix:

[0085]

[0086] The following discusses the selection of state variables corresponding to the linear Kalman subfilter: Since the accelerometer bias and gyro bias have a great influence on the navigation parameters, they should be included in the state variables of the SINS part of the linear Kalman subfilter; at the same time, in order to estimate the ocean current, the ocean current or ocean current error must be included in the state variables of the UVL part of the linear Kalman subfilter. In addition, in order to consider the simplicity and convenience of the Kalman filter model, this embodiment preferably adopts the indirect Kalman filter mode (estimates the error of the navigation parameters, and then calculates the navigation parameters), that is, selects the ocean current error (defined as the difference between the more accurate ocean current given by the filter and the estimated value given by formula (1)) as the state variable according to the following steps, see the last two components of the column vector on the right side of formula (5), which is consistent with the first formula of formula (3b).

[0087] The ocean current estimation error is used as a state variable, and the corresponding navigation parameter error is determined; wherein the navigation parameter error can be understood as a navigation parameter with ocean current information, and the navigation parameter error is expressed as:

[0088]

[0089] Among them, δL and δl represent the latitude error and longitude error respectively; δv N and δv E is the velocity component in the navigation system n; and Respectively represent the roll angle error, pitch angle error and heading angle error; and represents the zero bias of the three accelerometers in the b system; ε x , ε y and ε z Indicates the three gyro biases in the b system; and They represent the northward ocean current error and the eastward ocean current error respectively;

[0090] Assuming that the state transfer matrix in the time continuous LKF is F(t), the state transfer matrix in the discrete form LKF is Φ k+1,k for:

[0091]

[0092] Where ΔT is the discretized sampling period; F(t) is expressed as:

[0093]

[0094] In the formula,

[0095]

[0096]

[0097]

[0098]

[0099]

[0100]

[0101]

[0102]

[0103] Among them, M xy represents the transformation submatrix from variable x to y in FINS, where x / y is one of the attitude angle a, velocity v, and position p. Since p and v in this model only take two-dimensional quantities in the horizontal plane, M pp 、M vp 、M vv 、M pv Only the 2×2 submatrix in front of the corresponding matrix in the above formula can be taken, and M pa 、M va Take the first two rows of the corresponding matrix in the above formula, M ap 、M av Take the first two columns of the corresponding matrix in the above formula, that is is a 13×13 matrix;

[0104] Based on the state variables selected above, the corresponding state equation can be obtained, which will not be described here. The measurement equation is discussed below;

[0105] According to the real-time pure inertial navigation measurement, underwater velocity measurement and ocean current vector, a measurement equation is constructed; the measurement equation is expressed as:

[0106]

[0107] in, and They represent the calculated n-to-b attitude conversion matrix, accurate velocity vector, accurate ocean current vector and underwater velocity measurement output respectively;

[0108] From formula (8), we can know that:

[0109]

[0110] In the formula, w UVL They represent the attitude conversion matrix from n to b, the vector of misalignment angles of the three attitude angles, and the measurement noise of UVL respectively; V n and VC They represent the velocity vector solved by SINS and the ocean current vector obtained by the ocean current sub-model respectively. express The submatrix consisting of the i-th column and the j-k-th rows in the matrix. It should be noted that here v C is in the n system, so, The expression form is After considering the measurement noise of EML, it can be seen that That is: V C and is the projection vector in the n system, v C and is the projection vector in system b.

[0111] The left side of equation (9) is organized into a column vector, and the right side is organized into a matrix multiplied by a column vector and then superimposed with a noise term, and the measurement matrix shown in equation (10) is obtained:

[0112]

[0113] The last two lines in equation (10) are the UVL observation matrix, which is used for ocean current estimation. When setting the corresponding filter parameters, the coefficients in the noise term of equation (1) can be used to derive the value of the process noise variance matrix Q as shown in equation (11):

[0114]

[0115] It should be noted that, for the sake of convenience, the subscript i representing the subfilter number is omitted in equations (8) and (11).

[0116] According to the navigation parameter error and the measurement equation, a corresponding linear Kalman subfilter is obtained;

[0117] At this point, for any ocean current sub-model as shown in formula (1), the model can be used to establish a corresponding linear Kalman sub-filter, which is used to adopt the following steps and perform superposition fusion through IMM theory to fit the ocean current velocity as accurately as possible and improve the accuracy of complex ocean current estimation.

[0118] S13, according to the navigation parameters and covariance matrix output by each linear Kalman sub-filter, obtain the sub-filter mixing probability, and according to the sub-filter mixing probability, obtain the fused ocean current information of each linear Kalman sub-filter; the fused ocean current information includes fused navigation parameters, fused variance matrix, fused ocean current model time parameters and fused ocean current model variance parameters;

[0119] The step of obtaining the sub-filter mixing probability according to the navigation parameters and covariance matrix output by each linear Kalman sub-filter comprises:

[0120] According to the navigation parameters and covariance matrix output by each linear Kalman sub-filter, the corresponding model credibility probability is obtained; wherein the covariance matrix can be understood as the covariance matrix of the residuals existing when each linear Kalman sub-filter performs measurement update; specifically, the residuals and the corresponding covariance matrices are shown in equations (12) and (13):

[0121]

[0122]

[0123] In order to facilitate the fusion of sub-filters, it is assumed that each r i All obey the standard normal distribution, and the maximum likelihood ratio is calculated as shown in the following formula (14):

[0124]

[0125] Based on the above analysis, the model credibility probability can be updated (the probability of measuring which model is more credible during the measurement update process), and the model credibility probability of the i-th sub-filter at the k-th time is expressed as:

[0126]

[0127] Where n i,k-1 Represents the predicted probability of the i-th linear Kalman subfilter at the k-1th time, which can form a vector n, and the calculation method is as follows

[0128]

[0129] Among them, μ i,k represents the model credibility probability of the i-th linear Kalman subfilter at the k-th time; M represents the Markov chain transmission matrix; r i , C i and λ i Respectively represent the residual, corresponding covariance matrix and maximum likelihood ratio of the output of the i-th linear Kalman sub-filter; I represents the total number of linear Kalman sub-filters; n i,k-1 Represents the predicted probability of the i-th linear Kalman subfilter at the k-1th time. The initial value of the corresponding vector n can be set based on experience; for example, if there are two subfilters in total, the recommended initial value is [0.5 0.5] T , and based on experience, the recommended M matrix is:

[0130] According to the credible probabilities of each model, the sub-filter mixing probability is obtained; the sub-filter mixing probability is expressed as:

[0131]

[0132] Among them, g 11,k ,...,g 1I,k represents the superposition weight required to calculate the output of the first linear Kalman subfilter at the kth time; g I1,k ,...g II,k represents the superposition weight required to calculate the output of the I-th linear Kalman sub-filter at the k-th moment, that is, the output of the I-th sub-filter at the current moment is obtained by superimposing the outputs of all sub-filters at the previous moment; m ij represents the probability of switching from the i-th linear Kalman subfilter to the j-th linear Kalman subfilter;

[0133] The obtained sub-filter mixing probability can be used to fuse the outputs of all sub-filters. The expression of the fused ocean current information corresponding to each linear Kalman sub-filter is as follows:

[0134]

[0135] The physical meaning of equation (17) is that the state variables and covariance matrix after mixing are redistributed to each sub-filter. Redistribution means replacing the original values ​​in each sub-filter with these new state variables and covariance matrix. replace Assuming there are two sub-filters and k = 1, the initial values ​​of the state variables, variance matrix, time parameters of the ocean current model, and variance parameters of the ocean current model of sub-filters 1 and 2 are and At this point no mixing has taken place and the superscript m should be omitted.

[0136] S14, obtaining a fusion filter output according to the fused ocean current information of each linear Kalman sub-filter, and obtaining a time-varying ocean current estimation according to the fusion filter output; wherein the fusion filter output can be understood as an output value obtained by combining and superimposing the outputs of each sub-filter based on an IMM fusion algorithm; specifically, the steps of obtaining a fusion filter output according to the fused ocean current information of each linear Kalman sub-filter, and obtaining a time-varying ocean current estimation according to the fusion filter output include:

[0137] According to the model credibility probability and fusion navigation parameters of each linear Kalman sub-filter, the fusion filter output is obtained; the fusion filter output is shown in formula (18):

[0138]

[0139] in, and μ i,k They represent the fusion navigation parameters and model credibility probability of the i-th linear Kalman subfilter at the k-th moment respectively;

[0140] Obtaining an ideal ocean current value according to the fused ocean current model time parameter and the fused ocean current model variance parameter;

[0141] The ocean current estimation error in the fusion filter output is compensated to the ideal ocean current value to obtain the time-varying ocean current estimation; wherein the time-varying ocean current estimation can be understood as the estimated high-precision ocean current value, as shown in formula (19):

[0142]

[0143] Formula (19) is obtained by transforming the second formula of the previous formula (3a); It is the ideal ocean current value obtained by recursion of formula (1), but there is always a small error between this value and the actual ocean current speed; These are the last two components of the column vector output by equation (18); when the accuracy of the filter estimation is high enough, the value obtained by equation (19) can be considered to be an accurate ocean current value close to the true value.

[0144] The vector output by formula (19) can not only estimate the accurate ocean current value, but also estimate all navigation parameters or navigation information more accurately; that is, Figure 3 As shown, after the step S14 of obtaining the time-varying ocean current estimation according to the fusion filter output, the method further includes:

[0145] S15, compensating the fusion filter output to the pure inertial solution navigation parameters acquired in real time to obtain an estimated value of the integrated navigation parameter; wherein the estimated value of the integrated navigation parameter is as shown in formula (20):

[0146]

[0147] in, and They represent the estimated value of the combined navigation parameter, the pure inertial solution navigation parameter and the output of the IMM Kalman filter obtained by equation (18). This equation can be used to update navigation parameters such as attitude, speed, position and ocean current. When the accuracy of the filter estimation is high enough, It can be considered as an accurate navigation parameter close to the true value.

[0148] The embodiment of the present application realizes the construction of a preset number of ocean current sub-models based on the ocean current information of the target area based on the acquired ocean current observation data including different change periods, and then constructs the corresponding linear Kalman sub-filters, and then obtains the sub-filter mixing probability according to the navigation parameters and covariance matrix output by each linear Kalman sub-filter, and obtains the fused ocean current information of each linear Kalman sub-filter including the fused navigation parameters, the fused variance matrix, the fused ocean current model time parameters and the fused ocean current model variance parameters according to the sub-filter mixing probability, and then obtains the fused filter output according to the fused ocean current information of each linear Kalman sub-filter, and obtains the time-varying ocean current estimation method and the corresponding combined navigation parameter calibration method for time-varying ocean current valuation according to the fused filter output, and effectively solves the problem of ocean current estimation in unfamiliar waters such as deep sea by mixing and superimposing several sub-filter outputs according to the interactive multi-model according to the ocean current multi-model theory, and improves the estimation accuracy of complex ocean currents while providing reliable protection for high-precision combined navigation systems.

[0149] In one embodiment, Figure 4 As shown, a time-varying ocean current estimation system based on an interactive Kalman model is provided, the system comprising:

[0150] Model building module 1, used to establish a preset number of ocean current sub-models based on the acquired ocean current information of the target area; the ocean current information of the target area includes ocean current observation data of different change periods;

[0151] The filter construction module 2 is used to construct a corresponding linear Kalman sub-filter according to each ocean current sub-model; the measurement matrix of the linear Kalman sub-filter includes a pure inertial navigation measurement sub-matrix and an underwater speed measurement sub-matrix;

[0152] The probability fusion module 3 is used to obtain the sub-filter mixing probability according to the navigation parameters and covariance matrix output by each linear Kalman sub-filter, and obtain the fused ocean current information of each linear Kalman sub-filter according to the sub-filter mixing probability; the fused ocean current information includes the fused navigation parameter, the fused variance matrix, the fused ocean current model time parameter and the fused ocean current model variance parameter;

[0153] The ocean current estimation module 4 is used to obtain a fusion filter output according to the fusion ocean current information of each linear Kalman sub-filter, and obtain a time-varying ocean current estimation according to the fusion filter output.

[0154] In one embodiment, Figure 5 Said system further comprises:

[0155] The navigation calibration module 5 is used to compensate the fusion filter output to the pure inertial solution navigation parameters acquired in real time to obtain the estimated value of the combined navigation parameters.

[0156] For the specific definition of a time-varying ocean current estimation system based on an interactive Kalman model, please refer to the definition of a time-varying ocean current estimation method based on an interactive Kalman model mentioned above, which will not be repeated here. Each module in the above-mentioned time-varying ocean current estimation system based on an interactive Kalman model can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0157] Figure 6 FIG. 1 shows an internal structure diagram of a computer device in an embodiment, and the computer device may specifically be a terminal or a server. Figure 6 As shown, the computer device includes a processor, a memory, a network interface, a display and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a time-varying ocean current estimation method based on an interactive Kalman model is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0158] It can be understood by those skilled in the art that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the present application scheme, and does not constitute a limitation on the computer device to which the present application scheme is applied. The specific computing device may include more or fewer components than shown in the figure, or combine certain components, or have the same component arrangement.

[0159] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the steps of the above method are implemented when the processor executes the computer program.

[0160] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0161] In summary, the embodiments of the present invention provide a method, system, computer device and storage medium for estimating time-varying ocean currents based on an interactive Kalman model. The method for estimating time-varying ocean currents based on an interactive Kalman model implements a preset number of ocean current sub-models established based on the ocean current information of the target area based on the acquired ocean current observation data including different change periods, constructs a corresponding linear Kalman sub-filter, and then obtains the sub-filter mixing probability according to the navigation parameters and covariance matrix output by each linear Kalman sub-filter, and obtains the fused ocean current information of each linear Kalman sub-filter including the fused navigation parameters, the fused variance matrix, the fused ocean current model time parameters and the fused ocean current model variance parameters according to the sub-filter mixing probability, and obtains the fused filter output according to the fused ocean current information of each linear Kalman sub-filter, and obtains the time-varying ocean current estimation and the corresponding combined navigation parameter calibration according to the fused filter output. The method effectively solves the problem of ocean current estimation in unfamiliar waters such as deep seas by mixing and superimposing several sub-filter outputs according to the interactive multi-model according to the ocean current multi-model theory, improves the estimation accuracy of complex ocean currents, and provides reliable protection for high-precision combined navigation systems.

[0162] Each embodiment in this specification is described in a progressive manner, and the same or similar parts of each embodiment can be directly referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. It should be noted that the technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above-mentioned embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0163] The above-mentioned embodiments only express several preferred implementation modes of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in the technical field, several improvements and substitutions can be made without departing from the technical principles of the present invention, and these improvements and substitutions should also be regarded as the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be based on the protection scope of the claims.

Claims

1. A method for estimating time-varying ocean currents based on the interactive Kalman model, It is characterized in that The method comprises the following steps: Establishing a preset number of ocean current sub-models based on the acquired ocean current information of the target area; the ocean current information of the target area includes ocean current observation data of different change periods; According to each ocean current sub-model, a corresponding linear Kalman sub-filter is constructed; the measurement matrix of the linear Kalman sub-filter includes a pure inertial navigation measurement sub-matrix and an underwater speed measurement sub-matrix; According to the navigation parameters and covariance matrix output by each linear Kalman sub-filter, a sub-filter mixing probability is obtained, and according to the sub-filter mixing probability, fused ocean current information of each linear Kalman sub-filter is obtained; the fused ocean current information includes fused navigation parameters, fused variance matrix, fused ocean current model time parameters and fused ocean current model variance parameters; According to the fused ocean current information of each linear Kalman sub-filter, a fused filter output is obtained, and according to the fused filter output, a time-varying ocean current estimation is obtained.

2. The time-varying ocean current estimation method based on the interactive Kalman model as claimed in claim 1, It is characterized in that The ocean current submodel is a first-order Markov ocean current model; the ocean current submodel is expressed as: in, and represent the ocean current values ​​corresponding to the k+1th and kth sampling moments respectively; T C The time constant representing the relevant interval of ocean current velocity; σ C represents the root mean square of the ocean current velocity; Δt represents the time interval between data generation; w C Represents white noise that conforms to a standard normal distribution.

3. The time-varying ocean current estimation method based on the interactive Kalman model as claimed in claim 2, It is characterized in that The step of establishing a preset number of ocean current sub-models according to the ocean current information of the target area comprises: According to the ocean current information of the target area, a preset number of sub-model parameters are determined with the minimum fitting error as the optimization goal; the sub-model parameters include time parameters and variance parameters; According to the sub-model parameters, a corresponding ocean current sub-model is constructed.

4. The time-varying ocean current estimation method based on the interactive Kalman model as claimed in claim 1, It is characterized in that The step of constructing a corresponding linear Kalman sub-filter according to each ocean current sub-model comprises: The NE coordinate system is used as the navigation system, and the corresponding ocean current estimation error is obtained according to each ocean current sub-model; the ocean current estimation error is expressed as: in, and represent the ocean current estimation errors at the k-1th and kth moments, respectively; and represent the ocean current velocity and ideal ocean current velocity based on the linear Kalman subfilter and ocean current submodel, respectively; T i C represents the time parameter of the ith ocean current submodel; I represents the total number of ocean current submodels; The ocean current estimation error is used as a state variable, and the corresponding navigation parameter error is determined; the navigation parameter error is expressed as: Among them, δL and δl represent the latitude error and longitude error respectively; δv N and δv E is the velocity component in the navigation system n; and Respectively represent the roll angle error, pitch angle error and heading angle error; and represents the zero bias of the three accelerometers in the b system; ε x , ε y and ε z Indicates the three gyro biases in the b system; and They represent the northward ocean current error and the eastward ocean current error respectively; According to the real-time pure inertial navigation measurement, underwater velocity measurement and ocean current vector, a measurement equation is constructed; the measurement equation is expressed as: in, and They represent the calculated n-to-b attitude conversion matrix, accurate velocity vector, accurate ocean current vector and underwater velocity measurement output respectively; A corresponding linear Kalman sub-filter is obtained according to the navigation parameter error and the measurement equation.

5. The time-varying ocean current estimation method based on the interactive Kalman model as claimed in claim 1, It is characterized in that The step of obtaining the sub-filter mixing probability according to the navigation parameters and covariance matrix output by each linear Kalman sub-filter comprises: According to the navigation parameters and covariance matrix output by each linear Kalman sub-filter, the corresponding model credibility probability is obtained; the model credibility probability is expressed as: In the formula, Among them, μ i,k represents the model credibility probability of the i-th linear Kalman subfilter at the k-th time; n i,k-1 represents the prediction probability of the i-th linear Kalman subfilter at the k-1th time; M represents the Markov chain transmission matrix; r i , C i and λ i denote the residual, the corresponding covariance matrix and the maximum likelihood ratio output by the i-th linear Kalman subfilter, respectively; I denotes the total number of linear Kalman subfilters; According to the credible probabilities of each model, the sub-filter mixing probability is obtained; the sub-filter mixing probability is expressed as: Among them, g 11,k ,...,g 1I,k represents the superposition weight required to calculate the output of the first linear Kalman subfilter at the kth time; g I1,k ,...g II,k represents the superposition weight required to calculate the output of the I-th linear Kalman subfilter at the k-th time; m ij represents the probability of switching from the i-th linear Kalman subfilter to the j-th linear Kalman subfilter.

6. The time-varying ocean current estimation method based on the interactive Kalman model as claimed in claim 5, It is characterized in that The steps of obtaining a fusion filter output according to the fusion ocean current information of each linear Kalman sub-filter, and obtaining a time-varying ocean current estimation according to the fusion filter output include: According to the model credibility probability and fusion navigation parameters of each linear Kalman sub-filter, the fusion filter output is obtained; the fusion filter output is expressed as: in, and μ i,k They represent the fusion navigation parameters and model credibility probability of the i-th linear Kalman subfilter at the k-th moment respectively; Obtaining an ideal ocean current value according to the fused ocean current model time parameter and the fused ocean current model variance parameter; The ocean current estimation error in the fusion filter output is compensated to the ideal ocean current value to obtain the time-varying ocean current estimate.

7. The time-varying ocean current estimation method based on the interactive Kalman model as claimed in claim 1, It is characterized in that After the step of obtaining the time-varying ocean current estimation according to the fusion filter output, the method further includes: The fusion filter output is compensated to the pure inertial solution navigation parameters acquired in real time to obtain the estimated value of the combined navigation parameters.

8. A time-varying ocean current estimation system based on the interactive Kalman model, It is characterized in that The system comprises: A model building module, used to establish a preset number of ocean current sub-models based on the acquired ocean current information of the target area; the ocean current information of the target area includes ocean current observation data of different change periods; A filter construction module is used to construct a corresponding linear Kalman sub-filter according to each ocean current sub-model; the measurement matrix of the linear Kalman sub-filter includes a pure inertial navigation measurement sub-matrix and an underwater speed measurement sub-matrix; A probability fusion module is used to obtain a sub-filter mixing probability according to the navigation parameters and covariance matrix output by each linear Kalman sub-filter, and obtain fused ocean current information of each linear Kalman sub-filter according to the sub-filter mixing probability; the fused ocean current information includes fused navigation parameters, fused variance matrix, fused ocean current model time parameters and fused ocean current model variance parameters; The ocean current estimation module is used to obtain a fusion filter output according to the fusion ocean current information of each linear Kalman sub-filter, and obtain a time-varying ocean current estimation according to the fusion filter output.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, It is characterized in that When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

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

Images