Electromagnetic array positioning method and device based on lstm-ekf fusion and readable medium

Abstract

The application discloses an electromagnetic array positioning method and device based on LSTM-EKF fusion and a readable medium, and relates to the electromagnetic positioning field.The method comprises the following steps: calculating the relative distance between each electromagnetic transmitting module and an electromagnetic receiving module at the current time through a magnetic dipole model; acquiring the position coordinates of each electromagnetic transmitting module, constructing input data together with the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time, and inputting the input data into a trained position coordinate prediction model of the electromagnetic receiving module to output the prediction value of the position coordinates of the electromagnetic receiving module at the current time; initializing the estimated value of a state vector and a covariance matrix thereof to obtain an initial estimated value of the state vector and an initial covariance matrix, and inputting the initial estimated value of the state vector and the initial covariance matrix into an EKF algorithm to output the final estimated value of the state vector.The application solves the problems of the existing electromagnetic positioning technology, such as noise sensitivity, susceptibility to metal interference, significant decline in positioning accuracy and limited working range.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of electromagnetic positioning, in particular to an electromagnetic array positioning method and device based on LSTM-EKF fusion and readable medium. BACKGROUND

[0002] As a high-precision non-contact positioning technology, the principle of electromagnetic positioning technology is to realize spatial positioning through the coupling effect between the alternating electromagnetic field generated by the electromagnetic transmitting module and the electromagnetic receiving module. It can be widely applied to medical surgery navigation, industrial robot trajectory control and other scenes. Due to its non-vision dependence, strong anti-shielding and other advantages, it has become the mainstream positioning scheme in the motion tracking field. However, it has the bottleneck problems of being sensitive to noise, being easily interfered by metal to cause significant decline in positioning accuracy and limited working range, which may limit its further development and application. SUMMARY

[0003] The present application aims to solve the above-mentioned technical problems by providing an electromagnetic array positioning method and device based on LSTM-EKF fusion and readable medium.

[0004] In a first aspect, the present application provides an electromagnetic array positioning method based on LSTM-EKF fusion, comprising the following steps:

[0005] In an interference-free environment, a position coordinate prediction model of the electromagnetic receiving module based on the LSTM model is constructed and trained to obtain a trained position coordinate prediction model of the electromagnetic receiving module;

[0006] The relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time is calculated by the magnetic dipole model. The position coordinates of each electromagnetic transmitting module are obtained and combined with the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time to construct input data. The input data is input into the trained position coordinate prediction model of the electromagnetic receiving module to output the predicted value of the position coordinates of the electromagnetic receiving module at the current time.

[0007] A state vector composed of the position coordinates and velocity of the electromagnetic receiving module is defined. The estimated value of the state vector and its covariance matrix are initialized to obtain the initial estimated value of the state vector and the initial covariance matrix and input into the EKF algorithm. An observation equation is constructed according to the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time and the predicted value of the position coordinates of the electromagnetic receiving module at the current time. The final estimated value of the state vector is output by iterative optimization in the prediction stage and the update stage. The positioning result of the electromagnetic receiving module at the current time is obtained in the final estimated value of the state vector.

[0008] As a preferred option, a genetic algorithm is used to optimize the parameters of the LSTM model during the training process of the position coordinate prediction model of the electromagnetic receiving module.

[0009] As a preferred option, the observation equation is expressed as:

[0010] ;

[0011] in, Let be the joint observation vector corresponding to the estimated state vector at time k. This represents the estimated value of the state vector at time k. This represents the observation matrix at time k. This represents the observation noise at time k. The observed noise follows an independent Gaussian distribution and exists separately. , , This represents the noise in network observations. The noise in the magnetic dipole observation is represented by T, which represents the matrix transpose. This represents a 3×3 identity matrix. Indicates obedience, Indicates a Gaussian distribution; and The standard deviations of network observations and magnetic dipole observations are respectively represented and calculated as follows: the error between the predicted and actual coordinates of the electromagnetic receiving module output by the trained electromagnetic receiving module position coordinate prediction model is calculated, and the standard deviation of network observations is obtained; the error between the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module obtained by the magnetic dipole model and the actual distance is calculated, and the standard deviation of magnetic dipole observations is obtained.

[0012] make And combined with the observation equation, the observation noise at time k is calculated. The observation equation will include the known observation noise at time k. of It participates in the update phase calculation of the EKF algorithm, whereby... Let be the distance vector formed by the relative distances between each electromagnetic transmitting module and the electromagnetic receiving module at time k. This represents the coordinate vector formed by the predicted values ​​of the electromagnetic receiving module's position coordinates at time k.

[0013] Preferably, when the number of electromagnetic transmission modules is 4, then , ,in, Let be the relative distance between the i-th electromagnetic transmitting module and the electromagnetic receiving module at time k. B is the magnetic field intensity sensed by the electromagnetic receiving module, is a characteristic parameter of the electromagnetic transmitting module, is the position coordinate of the electromagnetic receiving module in the prediction result of the estimation value of the state vector at time k, is the position coordinate of the i-th electromagnetic transmitting module, i = 1, 2, 3, 4.

[0014] As preferred, in the prediction stage, a state transition matrix is constructed, and the prediction result of the state vector estimation at time k and the prediction result of the covariance matrix are predicted according to the state transition matrix and the estimation value of the state vector at time k-1, as shown in the following formula:

[0015] ;

[0016] ;

[0017] wherein, represents the covariance matrix at time k-1, represents the prediction result of the estimation value of the state vector at time k, represents the prediction result of the covariance matrix at time k, is the state transition matrix, , is the time interval, represents the process noise covariance matrix, represents the process noise, which is subject to a Gaussian distribution with a mean of zero and a variance of the process noise covariance matrix ;

[0018] In the update stage, the Kalman gain at time k is calculated according to the prediction result of the covariance matrix at time k, the observation matrix and the residual covariance matrix, as shown in the following formula:

[0019] ;

[0020] wherein, represents the Kalman gain at time k, represents the residual covariance matrix;

[0021] The residual vector at time k is calculated according to the joint observation vector at time k, as shown in the following formula:

[0022] ;

[0023] wherein, represents the residual vector at time k, represents the prediction result of the estimation value of the state vector at time k corresponding to the joint observation vector;

[0024] The estimation value of the state vector at the k moment is updated according to the prediction result of the estimation value of the state vector at the k moment, the residual vector and the Kalman gain, as shown in the following formula:

[0025] ;

[0026] wherein, represents the estimation value of the state vector at the k moment;

[0027] The covariance matrix at the k moment is updated according to the prediction result of the covariance matrix at the k moment, the observation matrix and the Kalman gain, as shown in the following formula:

[0028] ;

[0029] wherein, represents a unit matrix, represents the covariance matrix at the k moment.

[0030] Preferably, the input data of the position coordinate prediction model of the electromagnetic receiving module is ; the state vector is represented as ; the position coordinate of the electromagnetic receiving module is represented as ; and the velocity of the electromagnetic receiving module is represented as .

[0031] In a second aspect, the present application provides an electromagnetic array positioning method based on LSTM-EKF fusion, comprising:

[0032] A model construction module is configured to construct a position coordinate prediction model of the electromagnetic receiving module based on an LSTM model and train the model in an interference-free environment, to obtain a trained position coordinate prediction model of the electromagnetic receiving module;

[0033] A prediction module is configured to calculate the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current moment by a magnetic dipole model, obtain the position coordinates of each electromagnetic transmitting module, construct input data together with the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current moment, input the input data into the trained position coordinate prediction model of the electromagnetic receiving module, and output the prediction value of the position coordinates of the electromagnetic receiving module at the current moment.

[0034] The iterative optimization module is configured to define a state vector composed of position coordinates of the electromagnetic receiving module and velocities in respective coordinate directions, initialize an estimated value of the state vector and a covariance matrix thereof, obtain an initial estimated value of the state vector and an initial covariance matrix and input into the EKF algorithm, construct an observation equation according to relative distances between each electromagnetic transmitting module and the electromagnetic receiving module at the current moment and a predicted value of the position coordinates of the electromagnetic receiving module at the current moment, and output an obtained final estimated value of the state vector by iterative optimization in a prediction stage and an update stage, and obtain a positioning result of the electromagnetic receiving module at the current moment in the final estimated value of the state vector.

[0035] In a third aspect, the present application provides an electronic device, comprising one or more processors; a storage device configured to store one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any of the implementation manners of the first aspect.

[0036] In a fourth aspect, the present application provides a computer readable storage medium, which stores a computer program, when the computer program is executed by a processor, the method described in any of the implementation manners of the first aspect is implemented.

[0037] In a fifth aspect, the present application provides a computer program product, comprising a computer program, when the computer program is executed by a processor, the method described in any of the implementation manners of the first aspect is implemented.

[0038] Compared with the prior art, the present application has the following beneficial effects:

[0039] (1) The electromagnetic array positioning method based on LSTM-EKF fusion provided by the present application can obtain the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current moment and the position coordinates of each electromagnetic transmitting module by the magnetic dipole model, and predict the predicted value of the position coordinates of the electromagnetic receiving module by the position coordinate prediction model of the electromagnetic receiving module based on the LSTM model, and the parameters of the LSTM model are optimized by the genetic algorithm in the training process of the position coordinate prediction model of the electromagnetic receiving module, which can effectively improve the robustness and reduce the workload and subjectivity of manual parameter adjustment.

[0040] (2) The electromagnetic array positioning method based on LSTM-EKF fusion provided in the application can utilize the standard deviation of network observation and the standard deviation of magnetic dipole observation to construct observation noise in an observation equation, and can utilize the relative distance between each electromagnetic transmitting module and electromagnetic receiving module at the current time calculated by a magnetic dipole model and the predicted value of the position coordinates of the electromagnetic receiving module at the current time predicted by a trained position coordinate prediction model of the electromagnetic receiving module to construct a joint observation vector, and update an observation matrix in each iteration of the EKF algorithm to accurately obtain geometric constraints in the EKF algorithm.

[0041] (3) The electromagnetic array positioning method based on LSTM-EKF fusion provided in the application realizes stable estimation of the position coordinates of the electromagnetic receiving module under the iteration optimization of the prediction stage and the update stage in the EKF algorithm, effectively suppresses the performance degradation problem caused by nonlinear spatial noise and motion uncertainty, and makes the positioning accuracy higher. BRIEF DESCRIPTION OF DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor under the premise of the drawings.

[0043] Figure 1 The flowchart of the electromagnetic array positioning method based on LSTM-EKF fusion of the embodiments of the application;

[0044] Figure 2 The flowchart of the electromagnetic array positioning method based on LSTM-EKF fusion of the embodiments of the application;

[0045] Figure 3 The schematic diagram of the experimental scene of the electromagnetic array positioning method based on LSTM-EKF fusion of the embodiments of the application;

[0046] Figure 4 The schematic diagram of the electromagnetic array positioning device based on LSTM-EKF fusion of the embodiments of the application;

[0047] Figure 5 The hardware structure schematic diagram of the electronic device provided by the embodiments of the application. DETAILED DESCRIPTION

[0048] In order to make the objects, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of the present application.

[0049] Figure 1 An electromagnetic array positioning method based on LSTM-EKF fusion provided by an embodiment of the present application is shown, which comprises the following steps:

[0050] S1, in an interference-free environment, a position coordinate prediction model of an electromagnetic receiving module based on an LSTM model is constructed and trained to obtain a trained position coordinate prediction model of the electromagnetic receiving module.

[0051] In specific embodiments, a genetic algorithm is used to optimize the parameters of the LSTM model in the training process of the position coordinate prediction model of the electromagnetic receiving module.

[0052] Specifically, referring to Figure 2 , in the electromagnetic array positioning method of the embodiments of the present application, one electromagnetic receiving module and more than or equal to 3 electromagnetic transmitting modules are taken as the research objects. In one example, the number of electromagnetic transmitting modules is 4.

[0053] In the four corners of a 2m x 2m x 2m closed test space, four frequency-split excitation electromagnetic transmitting modules are respectively arranged, and the electromagnetic receiving module is fixed on a trolley which can move along a preset track on the track. In the case of ensuring no additional metal interference, a high-precision optical positioning device Vicon (positioning accuracy of 1mm) is used as the distance true value in the embodiments of the present application, the three-dimensional coordinates of the electromagnetic receiving module are collected in real time by the infrared camera array of the high-precision optical positioning device, and the true distance values between the electromagnetic receiving module and each electromagnetic transmitting module are calculated. At the beginning of the experiment, the trolley is controlled to move at a constant speed (speed of 0.25m / s) along the track, the magnetic field intensity induced by each electromagnetic transmitting module is recorded synchronously, and the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module calculated by the magnetic dipole model ( ) and the true distance values between each electromagnetic transmitting module and the electromagnetic receiving module collected by Vicon ( ) are recorded synchronously. At least 3000 groups of data are collected to ensure the reliability of the statistical results. The ranging error : of each group of data is calculated; statistical analysis is performed on all ranging errors , and the mean value (theoretically should be close to 0, if there is a system deviation, it needs to be calibrated first) and the standard deviation are calculated, which is the standard deviation of the magnetic dipole observation .

[0054] In the non-interference environment, a training data set generated by the movement of the electromagnetic receiving module along the preset track is collected, the data set contains input data and corresponding real coordinate values, the sample size in the training data set is not less than 5000 groups. The LSTM model optimized by using the genetic algorithm is used as the position coordinate prediction model of the electromagnetic receiving module, the mean square error (MSE) between the predicted value and the real coordinate value of the position coordinate of the electromagnetic receiving module is used as the loss function to train the position coordinate prediction model of the electromagnetic receiving module until the model converges. The LSTM model can effectively process the long-term dependence problem of time series data, in the training process, the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time and the position coordinate of each electromagnetic transmitting module constitute a vector as the input data of the position coordinate prediction model of the electromagnetic receiving module, and the predicted value of the position coordinate of the electromagnetic receiving module is used as the output data, the position coordinate prediction model of the electromagnetic receiving module is trained, after the training is completed, on the independent validation set, the three-dimensional error between the predicted value of the position coordinate of the electromagnetic receiving module output by the trained position coordinate prediction model of the electromagnetic receiving module , , and the real coordinate value collected by Vicon is calculated , all the errors in the validation set are counted, and the standard deviation is calculated, that is, the standard deviation of network observation .

[0055] S2, the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time is calculated by the magnetic dipole model; the position coordinates of each electromagnetic transmitting module are obtained and the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time is constructed to form input data, the input data is input into the trained position coordinate prediction model of the electromagnetic receiving module, and the predicted value of the position coordinate of the electromagnetic receiving module at the current time is output.

[0056] In specific embodiments, the input data of the position coordinate prediction model of the electromagnetic receiving module is ; the state vector is represented as ; , which represents the position coordinate of the electromagnetic receiving module, , which represents the velocity of the electromagnetic receiving module.

[0057] Specifically, refer to Figure 3The trained position coordinate prediction model of the electromagnetic receiving module is deployed, the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time is calculated by using a magnetic dipole model, and input data composed of the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time and the position coordinates of each electromagnetic transmitting module are input into the trained position coordinate prediction model of the electromagnetic receiving module, so that the prediction value of the position coordinates of the electromagnetic receiving module at the current time is obtained; the prediction value of the position coordinates of the electromagnetic receiving module at the current time is further used in the subsequent EKF algorithm together with the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time.

[0058] S3, a state vector composed of the position coordinates and the speed of the electromagnetic receiving module is defined, an estimated value of the state vector and a covariance matrix thereof are initialized, an initial estimated value of the state vector and an initial covariance matrix are obtained and input into the EKF algorithm, an observation equation is constructed according to the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time and the prediction value of the position coordinates of the electromagnetic receiving module at the current time, and a final estimated value of the state vector is output by using iterative optimization in the prediction stage and the update stage, so that the positioning result of the electromagnetic receiving module at the current time is obtained in the final estimated value of the state vector.

[0059] In specific embodiments, the observation equation is represented as:

[0060] ;

[0061] wherein, is a joint observation vector corresponding to the estimated value of the state vector at the k time, represents the estimated value of the state vector at the k time, represents the observation matrix at the k time, represents the observation noise at the k time, the observation noise is subject to an independent Gaussian distribution, and there are , , represents the noise of network observation, represents the noise of magnetic dipole observation, and T represents matrix transposition, represents a 3x3 unit matrix, represents subject to, represents a Gaussian distribution; and respectively, are calculated in the following ways: the standard deviation of the network observation is calculated by calculating the error between the predicted value of the position coordinates of the electromagnetic receiving module output by the trained position coordinate prediction model of the electromagnetic receiving module and the real coordinate value; the standard deviation of the magnetic dipole observation is calculated by calculating the error between the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module obtained by the magnetic dipole model and the real distance value;

[0062] Let and combine it with the observation equation to calculate the observation noise at time k , which contains the known observation noise at time k in the observation equation participate in the calculation of the update stage of the EKF algorithm, wherein is a distance vector composed of the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at time k, is a coordinate vector composed of the predicted value of the position coordinates of the electromagnetic receiving module at time k.

[0063] In specific embodiments, when the number of electromagnetic transmitting modules is 4, then , wherein, is the relative distance between the i-th electromagnetic transmitting module and the electromagnetic receiving module at time k, the relative distance , B is the magnetic field intensity induced by the electromagnetic receiving module, is a characteristic parameter of the electromagnetic transmitting module, is the position coordinates of the electromagnetic receiving module in the predicted result of the estimate of the state vector at time k, is the position coordinates of the i-th electromagnetic transmitting module, i=1,2,3,4.

[0064] In specific embodiments, in the prediction stage, a state transition matrix is constructed, and the predicted result of the state vector estimate and the predicted result of the covariance matrix at time k are predicted according to the state transition matrix and the estimate of the state vector at time k-1, as shown in the following formula:

[0065] ;

[0066] ;

[0067] wherein, is the covariance matrix at time k-1, is the predicted result of the estimate of the state vector at time k, is the predicted result of the covariance matrix at time k, is a state transition matrix,​ , is a time interval, denotes a process noise covariance matrix, denotes a process noise, which is subject to a Gaussian distribution with a mean of zero and a variance of the process noise covariance matrix ;

[0068] In the update stage, the Kalman gain at the kth time is calculated according to the prediction result of the covariance matrix at the kth time, the observation matrix and the residual covariance matrix, as shown in the following formula:

[0069] ;

[0070] wherein, denotes the Kalman gain at the kth time, denotes the residual covariance matrix;

[0071] The residual vector at the kth time is calculated according to the joint observation vector at the kth time, as shown in the following formula:

[0072] ;

[0073] wherein, denotes the residual vector at the kth time, denotes the prediction result of the estimation value of the state vector at the kth time corresponding to the joint observation vector;

[0074] The estimation value of the state vector at the kth time is updated according to the prediction result of the estimation value of the state vector at the kth time, the residual vector and the Kalman gain, as shown in the following formula:

[0075] ;

[0076] wherein, denotes the estimation value of the state vector at the kth time;

[0077] The covariance matrix at the kth time is updated according to the prediction result of the covariance matrix at the kth time, the observation matrix and the Kalman gain, as shown in the following formula:

[0078] ;

[0079] wherein, denotes a unit matrix, denotes the covariance matrix at the kth time.

[0080] Specifically, in the EKF algorithm, the state vector is defined as wherein, denotes the position coordinates of the electromagnetic receiving module, represents the speed of the electromagnetic receiving module. The state transition equation is constructed using the uniform motion model, and the process noise covariance matrix characterizes the unmodeled motion uncertainty. The process noise covariance matrix is determined in combination with the physical model, experimental data and engineering experience. The distance vector composed of the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at time k calculated using the magnetic dipole model and the predicted value of the position coordinates of the electromagnetic receiving module at time k output by the trained position coordinate prediction model of the electromagnetic receiving module are taken as double observation inputs to construct a joint observation vector, the observation noise is constructed using the standard deviation of the network observation and the standard deviation of the magnetic dipole observation, and the observation equation is established; the state vector and its covariance matrix are optimized through continuous iteration of the prediction stage and the update stage, and according to the initial estimate of the state vector and the initial covariance matrix, the final estimate of the state vector and the final covariance matrix can be robustly updated. The extended Kalman filter fusion algorithm can continuously optimize the positioning result and output a more accurate final estimate of the state vector, which includes the estimate of the position coordinates of the electromagnetic receiving module and the estimate of the speed.

[0081] The effect of the application will be described below using a specific experimental process.

[0082] The experimental equipment includes a Vicon optical positioning system, an electromagnetic transmitting module and an electromagnetic receiving module, wherein the Vicon optical positioning system is used to obtain the real coordinates and attitude of the tracked module; the electromagnetic transmitting module is used to generate an alternating electromagnetic field; the electromagnetic receiving module is used to induce the alternating electromagnetic field in space; and the position coordinate estimate of the electromagnetic receiving module is predicted using the electromagnetic array positioning method based on LSTM-EKF fusion proposed in the embodiments of the application. In the experimental process, the hyperparameter set of the LSTM network is optimized by a genetic algorithm, specifically including: the hidden layer dimension is [64, 128]; the network depth is optionally {2, 3, 4}; the learning rate η is in the range of [10-4, 10-3]; and the Dropout rate is in the range of [0, 0.3].

[0083] In the experimental process, the trilateration algorithm, the LSTM algorithm and the electromagnetic array positioning method based on LSTM-EKF fusion proposed in the embodiments of the application are compared respectively, and the results are shown in Table 1. The positioning accuracy of the electromagnetic array positioning method based on LSTM-EKF fusion proposed in the embodiments of the application is improved by 55.7% compared with the trilateration algorithm and by 50.1% compared with the LSTM algorithm, the trajectory stability is maintained under the condition of fast operation and interference, and the maximum dynamic tracking error is controlled within 12.5 mm.

[0084] Table 1 Error analysis of different algorithms

[0085]

[0086] Further reference Figure 4 As an implementation of the method shown in the above figures, the present application provides an embodiment of an electromagnetic array positioning device based on LSTM-EKF fusion, which corresponds to the method embodiment shown in Figure 1 The device can be applied to various electronic devices.

[0087] The embodiment of the present application provides an electromagnetic array positioning method based on LSTM-EKF fusion, comprising:

[0088] The model construction module 1 is configured to construct and train a position coordinate prediction model of an electromagnetic receiving module based on an LSTM model in an interference-free environment to obtain a trained position coordinate prediction model of the electromagnetic receiving module.

[0089] The prediction module 2 is configured to calculate the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time through a magnetic dipole model, obtain the position coordinates of each electromagnetic transmitting module, and construct input data with the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time, input the input data into the trained position coordinate prediction model of the electromagnetic receiving module, and output the prediction value of the position coordinates of the electromagnetic receiving module at the current time.

[0090] The iterative optimization module 3 is configured to define a state vector composed of the position coordinates of the electromagnetic receiving module and the speed in each coordinate direction, initialize the estimated value of the state vector and its covariance matrix to obtain an initial estimated value of the state vector and an initial covariance matrix and input them into the EKF algorithm, construct an observation equation with the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time and the prediction value of the position coordinates of the electromagnetic receiving module at the current time, and output the final estimated value of the state vector through iterative optimization in the prediction stage and the update stage, and obtain the positioning result of the electromagnetic receiving module at the current time in the final estimated value of the state vector.

[0091] Figure 5 The hardware structure diagram of the electronic device provided by the embodiment of the present application is shown in FIG. 1. Figure 5As shown, the electronic device of the embodiment includes a processor 501 and a memory 502; the memory 502 is configured to store computer execution instructions; the processor 501 is configured to execute the computer execution instructions stored in the memory to implement each step performed by the electronic device in the above embodiment. For details, refer to the related description in the foregoing method embodiment.

[0092] Optionally, the memory 502 can be independent or integrated with the processor 501.

[0093] When the memory 502 is independently arranged, the electronic device further includes a bus 503 configured to connect the memory 502 and the processor 501.

[0094] The embodiment of the application further provides a computer storage medium, and the computer storage medium stores computer execution instructions; when the processor 501 executes the computer execution instructions, the method described above is implemented.

[0095] The embodiment of the application further provides a computer program product, and the computer program product includes a computer program; when the computer program is executed by the processor 501, the method described above is implemented.

[0096] In the embodiments of the application, it should be understood that the disclosed device and method can be implemented in other ways. For example, the device embodiments described above are only schematic; for example, the division of the modules is only a logical function division, and there can be another division manner in actual implementation; for example, a plurality of modules can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the modules shown or discussed can be indirect coupling or communication connection through some interfaces, devices or modules, and can be electrical, mechanical or other forms.

[0097] The modules illustrated as separate components can or can not be physically separate, and the components illustrated as modules can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. Some or all of the modules can be selected according to actual needs to implement the embodiment scheme.

[0098] In addition, each functional module in each embodiment of the application can be integrated in one processing unit, or each module can be physically present independently, or two or more modules can be integrated in one unit. The unit formed by the above modules can be realized in the form of hardware, or in the form of hardware plus software function unit.

[0099] The integrated modules realized in the form of software function modules can be stored in a computer readable storage medium. The software function modules are stored in a storage medium and include a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or the processor 501 to execute part of the steps of the various embodiments of the method.

[0100] It should be understood that the processor 501 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), etc. The general-purpose processor can be a microprocessor or the processor 501 can also be any conventional processor 501, etc. The steps of the method disclosed in combination with the application can be directly embodied as the execution of the processor 501 in hardware, or be executed by a combination of hardware and software modules in the processor 501.

[0101] The memory 502 can include a high-speed RAM memory, and can also include a non-volatile storage NVM, for example at least one disk memory, and can also be a U disk, a mobile hard disk, a read-only memory, a magnetic disk or an optical disk, etc.

[0102] The bus 503 can be an industry standard architecture (ISA), a peripheral component interconnect (PCI) bus, or an extended industry standard architecture (EISA) bus, etc. The bus 503 can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, the bus 503 in the drawings of the present application does not limit only one bus 503 or one type of bus 503.

[0103] The storage medium can be realized by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk. The storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0104] An example storage medium is coupled to the processor 501 such that the processor 501 can read information from, and can write information to, the storage medium. Of course, the storage medium can be a part of the processor 501. Consistent with the teachings provided herein, the processor 501 and the storage medium can be located in a ASIC. The teachings provided herein can be used in other electronic devices such as a computer system, a consumer electronics device, or any other device that includes a processor and a storage medium.

[0105] Those skilled in the art can understand that all or part of the steps of the foregoing method embodiments can be completed by a program instruction related to hardware. The foregoing program can be stored in a computer readable storage medium. When the program is executed, the steps of the foregoing method embodiments are executed; and the foregoing storage medium includes various media capable of storing program codes, such as ROM, RAM, magnetic disk, or optical disk.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can still be modified, or some or all of the technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. An electromagnetic array positioning method based on LSTM-EKF fusion, characterized in that, The method comprises the following steps: In an interference-free environment, a position coordinate prediction model of an electromagnetic receiving module based on an LSTM model is constructed and trained to obtain a trained position coordinate prediction model of the electromagnetic receiving module; The relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time is calculated by a magnetic dipole model; the position coordinates of each electromagnetic transmitting module are obtained and used to construct input data together with the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time; the input data is input into the trained position coordinate prediction model of the electromagnetic receiving module to output a predicted value of the position coordinates of the electromagnetic receiving module at the current time; A state vector composed of the position coordinates and the speed of the electromagnetic receiving module is defined, and an estimated value of the state vector and a covariance matrix thereof are initialized to obtain an initial estimated value of the state vector and an initial covariance matrix and input into an EKF algorithm; an observation equation is constructed according to the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time and the predicted value of the position coordinates of the electromagnetic receiving module at the current time, and the final estimated value of the state vector is output by iterative optimization in the prediction stage and the update stage, and the positioning result of the electromagnetic receiving module at the current time is obtained in the final estimated value of the state vector.

2. The electromagnetic array positioning method based on LSTM-EKF fusion according to claim 1, characterized in that, The genetic algorithm is used to optimize the parameters of the LSTM model in the training process of the position coordinate prediction model of the electromagnetic receiving module.

3. The LSTM-EKF fusion based electromagnetic array positioning method of claim 1, wherein, The observation equation is expressed as: ； wherein, is a joint observation vector corresponding to the estimated value of the state vector at time k, represents an estimated value of a state vector at time k, represents an observation matrix at time k, represents an observation noise at time k, , the observation noise is subject to an independent Gaussian distribution, and there are , , represents a noise of network observation, represents a noise of magnetic dipole observation, and T represents a matrix transpose, represents a 3x3 unit matrix, represents subject to, represents a Gaussian distribution; and respectively represent a standard deviation of network observation and a standard deviation of magnetic dipole observation, and are calculated in the following manner: the error between the predicted value of the position coordinates of the electromagnetic receiving module output by the trained position coordinate prediction model of the electromagnetic receiving module and the real coordinate value is calculated to obtain the standard deviation of the network observation; the error between the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module obtained by the magnetic dipole model and the real distance value is calculated to obtain the standard deviation of the magnetic dipole observation. Let and combined with the observation equation, the observation noise at time k is calculated The observation equation contains the known observation noise at time k of participate in the calculation of the update phase of the EKF algorithm, wherein, is the distance vector composed of the relative distance between each electromagnetic transmitting module and electromagnetic receiving module at time k, is the coordinate vector composed of the predicted value of the position coordinates of the electromagnetic receiving module at time k.

4. The electromagnetic array positioning method based on LSTM-EKF fusion according to claim 3, characterized in that, When the number of the electromagnetic transmitting modules is 4, then , wherein, is the relative distance between the i-th electromagnetic transmitting module and the electromagnetic receiving module at the k-th moment, the relative distance , B is the magnetic field intensity sensed by the electromagnetic receiving module, is the characteristic parameter of the electromagnetic transmitting module, represents the position coordinates of the electromagnetic receiving module in the prediction result of the estimation value of the state vector at the k-th moment, is the position coordinates of the i-th electromagnetic transmitting module, i = 1, 2, 3, 4.

5. The electromagnetic array positioning method based on LSTM-EKF fusion according to claim 3, characterized in that, In the prediction stage, a state transition matrix is constructed, and the predicted result of the state vector estimation and the predicted result of the covariance matrix at time k are predicted according to the state transition matrix and the estimated value of the state vector at time k-1, as shown in the following formula: ； ； wherein, denotes a covariance matrix at time k-1, denotes a prediction result of an estimated value of a state vector at time k, denotes a prediction result of a covariance matrix at time k, is a state transition matrix, , is a time interval, denotes a process noise covariance matrix, denotes a process noise, which is subject to a Gaussian distribution with a mean of zero and a variance of a process noise covariance matrix . In the update stage, the Kalman gain at time k is calculated according to the predicted result of the covariance matrix at time k, the observation matrix and the residual covariance matrix, as shown in the following formula: ； wherein, denotes the Kalman gain at time k, denotes the residual covariance matrix; The residual vector at time k is calculated according to the joint observation vector at time k, as shown in the following formula: ； wherein, represents a residual vector at time k, represents a prediction result of an estimated value of a state vector at time k the corresponding joint observation vector; The estimated value of the state vector at time k is updated according to the predicted result of the estimated value of the state vector at time k, the residual vector and the Kalman gain, as shown in the following formula: ； wherein, represents the estimated value of the state vector at time k; The covariance matrix at time k is updated according to the predicted result of the covariance matrix at time k, the observation matrix and the Kalman gain, as shown in the following formula: ； wherein, denotes the identity matrix, denotes the covariance matrix at time k.

6. The electromagnetic array positioning method based on LSTM-EKF fusion according to claim 4, characterized in that, The input data of the position coordinate prediction model of the electromagnetic receiving module is ; the state vector is represented as ; represents the position coordinate of the electromagnetic receiving module, represents the speed of the electromagnetic receiving module.

7. An electromagnetic array positioning method based on LSTM-EKF fusion, characterized in that, The method comprises: The model construction module is configured to construct and train a position coordinate prediction model of an electromagnetic receiving module based on an LSTM model in an interference-free environment to obtain a trained position coordinate prediction model of the electromagnetic receiving module; The prediction module is configured to calculate the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time by a magnetic dipole model; obtain the position coordinates of each electromagnetic transmitting module and construct input data together with the relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at the current time; input the input data into the trained position coordinate prediction model of the electromagnetic receiving module to output a predicted value of the position coordinates of the electromagnetic receiving module at the current time; The iterative optimization module is configured to define a state vector composed of position coordinates of the electromagnetic receiving module and velocities in respective coordinate directions, initialize an estimated value of the state vector and a covariance matrix thereof, obtain an initial estimated value of the state vector and an initial covariance matrix and input into the EKF algorithm, construct an observation equation according to a relative distance between each electromagnetic transmitting module and the electromagnetic receiving module at a current time and a predicted value of the position coordinates of the electromagnetic receiving module at the current time, and output an obtained final estimated value of the state vector by using iterative optimization in a prediction stage and an update stage, and obtain a positioning result of the electromagnetic receiving module at the current time in the final estimated value of the state vector.

8. An electronic device, comprising: one or more processors; storage configured to store one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1-6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the method according to any one of claims 1-6.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the method according to any one of claims 1-6. The computer program is executed by the processor to implement the method according to any one of claims 1-6.

Images