AI-based ship positioning method and system

Abstract

The invention belongs to the technical field of ship positioning, and provides an AI-based ship positioning method and system. The method comprises the following steps: acquiring a first position of a ship at a first time point based on a GPS, acquiring inertial information of the ship between the first time point and a second time point, and deriving a first predicted position of the ship at the second time point based on the inertial information and the first position; obtaining a second position of the ship at a second time point based on the GPS, calculating a prediction error between the second position and the first prediction position, training and constructing a fuzzy generalized learning model, and when the GPS is invalid, predicting a position error by using the fuzzy generalized learning model according to inertial information of the ship after the GPS is invalid, based on the inertial information and position information determined before the GPS of the ship is invalid, a second predicted position of the ship is obtained through derivation, the predicted position error is compensated to the second predicted position, and an actual position is obtained; therefore, the defect that the GPS is independently used for ship positioning in the prior art can be overcome.

Description

Technical Field

[0001] The present invention relates to the field of ship positioning technology, and in particular to an AI-based ship positioning method and system. Background Art

[0002] Ship position prediction plays a key role in inland and maritime navigation warning and safety. Whether it is automatic or manual driving, the current position of the ship must be known when operating the ship to effectively avoid collisions and plan navigation paths. In the prior art, the position of the ship is generally determined by GPS. GPS has the advantage of high precision, but the attitude information of the ship cannot be obtained by using GPS equipment alone, and GPS is susceptible to electromagnetic interference and has a low data update rate. When there is no GPS signal, the ship cannot determine its current position. Summary of the invention

[0003] In view of the above technical problems, the present invention provides an AI-based ship positioning method and system to solve the defect of using GPS alone for ship positioning in the prior art.

[0004] Other features and advantages of the present invention will become apparent from the following detailed description, or may be learned in part by the practice of the present disclosure.

[0005] According to one aspect of the present invention, a ship positioning method based on AI is disclosed, the method comprising:

[0006] Acquire a first position of the ship at a first time point based on GPS, acquire inertial information of the ship between the first time point and a second time point, the inertial information at least including acceleration data and angular velocity data, and derive a first predicted position of the ship at a second time point based on the inertial information and the first position;

[0007] Acquiring a second position of the ship at the second time point based on GPS, calculating a prediction error between the second position and the first predicted position, training and constructing a fuzzy generalized learning model based on a fuzzy prediction algorithm, taking the inertial information as input and the prediction error as output, and, before training, preprocessing the inertial information based on wavelet transform denoising and SVM regression;

[0008] When GPS is valid, continuously acquiring the predicted position and the predicted error to train and update the fuzzy generalized learning model, and using the position determined by GPS as the actual position of the ship;

[0009] When GPS is invalid, the fuzzy generalized learning model is used to predict the position error based on the inertial information of the ship after the GPS is invalid, and the second predicted position of the ship is derived based on the inertial information and the position information determined before the GPS is invalid, and the predicted position error is compensated to the second predicted position to obtain the actual position of the ship.

[0010] Further, when deriving the first predicted position of the ship at the second time point, the method specifically includes:

[0011] Performing time-integrated calculations on the angular velocity and the acceleration respectively to obtain attitude changes and displacement changes of the ship;

[0012] The first position is taken as a reference point, and the first predicted position is obtained by combining the posture change and the displacement change.

[0013] Furthermore, when performing time integral calculation, error compensation is performed based on a Kalman filter or a wavelet transform denoising method.

[0014] Furthermore, the preprocessing of the inertial information based on wavelet transform denoising and SVM regression includes:

[0015] Processing the inertial information to remove noise based on a wavelet transform denoising algorithm, and performing smoothing based on a mean filter;

[0016] The smoothed inertial information is used as an input for SVM regression to train an SVM model. During training, the predicted results of the SVM model are compared with the actual results, and the optimal coefficients and parameters of the SVM model are calculated based on the traversal algorithm.

[0017] Furthermore, when the inertial information is subjected to noise removal processing based on a wavelet transform denoising algorithm, the method includes:

[0018] Performing multiple wavelet decompositions on the inertial information so that the inertial information is decomposed into different frequency components, and obtaining wavelet coefficients of different time scales and frequencies generated at each level of decomposition;

[0019] Performing threshold processing on the wavelet coefficients, treating the wavelet coefficients smaller than the threshold as noise and setting them to zero, and retaining the wavelet coefficients larger than the threshold;

[0020] The retained wavelet coefficients are reconstructed by inverse wavelet transform to obtain the denoised inertial information.

[0021] Furthermore, the SVM model uses a radial basis function as a kernel function.

[0022] Furthermore, when training and constructing the fuzzy generalized learning model, it includes:

[0023] Defining a structure of the fuzzy generalized learning model, wherein the structure of the fuzzy generalized learning model includes n fuzzy subsystems and m enhancement nodes;

[0024] defining a plurality of fuzzy rules for each of the fuzzy subsystems, wherein the fuzzy rules are used to describe the relationship between the inertial information and the position error;

[0025] A Gaussian membership function is defined for each fuzzy set, and its center and width are determined based on the K-means clustering algorithm;

[0026] calculating the activation strength of each of the fuzzy rules as a product of the Gaussian membership function of the inertial information;

[0027] Based on the activation strength and the parameters of the fuzzy subsystem, construct an output function of each of the fuzzy subsystems;

[0028] The inertial information is organized into a training data set and represented by a matrix, and the parameters of the fuzzy subsystem are trained and adjusted to obtain the fuzzy generalized learning model.

[0029] According to another aspect of the present invention, there is provided an AI-based ship positioning system, the system comprising:

[0030] an inertial sensing module, configured to obtain a first position of the ship at a first time point based on GPS, obtain inertial information of the ship between the first time point and a second time point, wherein the inertial information includes at least acceleration data and angular velocity data, and derive a first predicted position of the ship at a second time point based on the inertial information and the first position;

[0031] a training module, for obtaining a second position of the ship at the second time point based on GPS, calculating a prediction error between the second position and the first predicted position, training and constructing a fuzzy generalized learning model based on a fuzzy prediction algorithm, taking the inertial information as input and the prediction error as output, and, before training, preprocessing the inertial information based on wavelet denoising and SVM regression;

[0032] The analysis module is used to continuously acquire the predicted position and the predicted error to train and update the fuzzy generalized learning model when the GPS is valid, and use the position determined by the GPS as the actual position of the ship; when the GPS is invalid, the fuzzy generalized learning model is used to predict the position error according to the inertial information of the ship after the GPS is invalid, and derive the second predicted position of the ship based on the inertial information and the position information of the ship determined before the GPS is invalid, and compensate the predicted position error to the second predicted position to obtain the actual position of the ship.

[0033] The technical solution disclosed in this disclosure has the following beneficial effects:

[0034] The ship's position is deduced based on inertial information when there is no GPS, avoiding the risk of shipping and driving caused by long-term absence of GPS signals. In addition, based on the constructed fuzzy generalized learning model, the derived position is compensated for errors, making the derived position more accurate and solving the integral error generated in the derivation process. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a flow chart of an AI-based ship positioning method in an embodiment of this specification;

[0036] Figure 2 This is a structural block diagram of an AI-based ship positioning system implemented in an embodiment of this specification;

[0037] Figure 3 A device for implementing an AI-based ship positioning method in an embodiment of this specification;

[0038] Figure 4 A computer-readable storage medium storing an AI-based ship positioning method in an embodiment of this specification. DETAILED DESCRIPTION

[0039] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as being limited to the examples set forth herein; on the contrary, these embodiments are provided so that the present disclosure will be more comprehensive and complete, and the concepts of the example embodiments are fully conveyed to those skilled in the art. The described features, structures, or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, systems, steps, etc. may be adopted. In other cases, known technical solutions are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0040] In addition, the accompanying drawings are only schematic illustrations of the present disclosure. The same reference numerals in the figures represent the same or similar parts, and thus their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities, which do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor systems and / or microcontroller systems.

[0041] like Figure 1 As shown, the embodiment of this specification provides an AI-based ship positioning method, which is applied to a ship system, such as an AIS system (automatic identification system). The method may specifically include the following steps S101 to S104:

[0042] In step S101, a first position of a ship at a first time point is acquired based on GPS, and inertial information of the ship between the first time point and a second time point is acquired, wherein the inertial information includes at least acceleration data and angular velocity data, and a first predicted position of the ship at a second time point is derived based on the inertial information and the first position.

[0043] Among them, the ship's position at a first time point (first position) is obtained from the GPS as an initial reference point for subsequent position derivation; during the period from the first time point to the second time point, data from the inertial measurement unit (IMU) is collected, including acceleration and angular velocity, which capture the dynamic characteristics of the ship; based on the inertial information and the first position, the predicted position of the ship at the second time point is derived.

[0044] In step S102, the second position of the ship at the second time point is obtained based on GPS, and the prediction error between the second position and the first predicted position is calculated. Based on a fuzzy prediction algorithm, the inertial information is used as input and the prediction error is used as output to train and construct a fuzzy generalized learning model. Before training, the inertial information is preprocessed based on wavelet transform denoising and SVM regression.

[0045] Among them, the real position of the ship at the second time point, that is, the second position, is obtained based on GPS as a benchmark for comparison with the predicted position, and then the first predicted position derived based on the inertial information in step S101 is used to calculate the error between it and the real second position. The error mainly reflects the cumulative error of inertial navigation without relying on GPS. Then, a fuzzy generalized learning algorithm is used, inertial information is used as model input, and the predicted error is used as model output to train the model. The nonlinear processing ability of the fuzzy system and the efficient structure of generalized learning are used to construct a prediction model suitable for real-time error correction. The wavelet transform denoising is to eliminate short-term random noise in the inertial information and retain the dynamic characteristics. SVM regression corrects long-term errors through support vector regression on the basis of denoised data to ensure the accuracy of input data.

[0046] In step S103, when GPS is valid, the predicted position and the predicted error are continuously acquired to train and update the fuzzy generalized learning model, and the position determined by GPS is used as the actual position of the ship.

[0047] Among them, when the GPS signal is normal, the position provided by GPS is used as the reference of the ship's real position, and the predicted position derived based on inertial information and the predicted error between it and the real GPS position are obtained. Then the predicted position and the predicted error are continuously recorded, and these data are used to dynamically update the fuzzy generalized learning model, and the input-output mapping of the model is updated in real time to improve its prediction accuracy in nonlinear dynamic environments. In addition, the GPS position is used as the standard value of the actual position of the ship to correct the accumulated error of inertial navigation.

[0048] Specifically, when the fuzzy generalized learning model is updated in real time, the mapping relationship from input to output is expressed as:

[0049] The form of the fuzzy rule is: if x1 is A1, x2 is A2, ..., x M Yes A M ,but

[0050] Input: Inertial information x1, x2, …, x M , which are usually acceleration and angular velocity;

[0051] Output: prediction error y = δP;

[0052] Enhanced node feature extensions:

[0053] Non-linear mapping enhancement:

[0054] g(·) is the activation function, W j ,b j is the node weight and bias;

[0055] The objective function to minimize the error is:

[0056] Therefore, we can update the parameter W,α i ,μ i To dynamically update the fuzzy generalized learning model.

[0057] In step S104, when the GPS is invalid, the fuzzy generalized learning model is used to predict the position error based on the inertial information of the ship after the GPS is invalid, and the second predicted position of the ship is derived based on the inertial information and the position information determined before the GPS is invalid, and the predicted position error is compensated to the second predicted position to obtain the actual position of the ship.

[0058] Among them, when GPS is invalid, it is necessary to use inertial navigation information to deduce the position of the ship. Due to the cumulative error of the inertial navigation system (INS), the position prediction must be corrected by error compensation. Specifically, the inertial information (acceleration and angular velocity) of the ship during the period when GPS is invalid is collected, and based on the trained fuzzy generalized learning model, the inertial information is input to predict the position error during the period when GPS is invalid; using the inertial information and the known position of the ship before GPS is invalid, the second predicted position of the ship can be derived through the INS state equation; the predicted position error is applied to the second predicted position to obtain the corrected position.

[0059] The INS state equation is based on the navigation mechanism of the inertial sensor (IMU) data, including the integration of acceleration and angular velocity, to calculate the position and velocity change of the ship over a period of time:

[0060]

[0061] Attitude error formula:

[0062] Position derivation formula:

[0063] From the above formula, a derivation method from inertial information to position change is provided, so the second predicted position can be calculated.

[0064] In one embodiment, when deriving the first predicted position of the ship at the second time point, it specifically includes: performing time integral calculations on the angular velocity and the acceleration respectively to obtain the attitude change and displacement change of the ship; taking the first position as a reference point, combining the attitude change and the displacement change to obtain the first predicted position.

[0065] Among them, inertial navigation deduces the attitude change and displacement change of the ship by integrating the angular velocity and acceleration over time, and then derives the predicted position by combining the first position. When deriving, the attitude change of the ship in the body coordinate system is deduced by integrating the angular velocity over time, and the displacement change of the ship is deduced by integrating the acceleration over time. The attitude change is calculated as:

[0066] q(t)=∫ω b (t)dt;

[0067] q(t) is the attitude change of the ship, expressed by quaternion, Euler angle or direction cosine matrix (DCM);

[0068] ω b is the angular velocity vector;

[0069] Then use the direction cosine matrix Map the posture of the body coordinate system to the navigation coordinate system:

[0070]

[0071] The displacement change is calculated as:

[0072] The time integral of the acceleration calculates the velocity change:

[0073] v(t0) is the initial velocity, a n To pass Convert to the acceleration in the navigation coordinate system;

[0074] The change in displacement is then calculated by integrating the velocity over time:

[0075] Then, taking the first position as the reference, combined with the displacement change and attitude change, the first predicted position of the ship at the second time point is calculated: P 预测 (t) = P 初始 +Δr(t);

[0076] P 预测 (t) is the first predicted position at the second time point, P 初始 is the first position, and Δr(t) is the displacement change obtained by the above integration.

[0077] In one embodiment, when performing time integral calculation, error compensation is performed based on a Kalman filter or a wavelet transform denoising.

[0078] In one embodiment, when the inertial information is subjected to noise removal processing based on a wavelet transform denoising algorithm, the method includes: performing multiple wavelet decompositions on the inertial information so that the inertial information is decomposed into different frequency components, and obtaining wavelet coefficients of different time scales and frequencies generated at each level of decomposition; performing threshold processing on the wavelet coefficients, treating the wavelet coefficients less than the threshold as noise and setting them to zero, and retaining the wavelet coefficients greater than the threshold; and reconstructing the retained wavelet coefficients through an inverse wavelet transform to obtain the denoised inertial information.

[0079] Among them, the original measurement values ​​of inertial information (acceleration and angular velocity) contain short-term random noise and systematic errors. Short-term noise is mainly high-frequency noise, which can be filtered out by wavelet denoising. The original inertial information is decomposed by wavelet to obtain different frequency components:

[0080]

[0081] c j,k is the coefficient of the low-frequency component, d J,k is the coefficient of the high frequency component, φ j,k and ψ J,k is the wavelet basis function.

[0082] By setting a threshold to remove high-frequency noise, the low-frequency components can be retained, and then the inverse wavelet transform is used to reconstruct the signal to obtain the denoised inertial information.

[0083] In one embodiment, the preprocessing of the inertial information based on wavelet transform denoising and SVM regression includes: performing noise removal processing on the inertial information based on a wavelet transform denoising algorithm, and performing smoothing processing based on a mean filter; using the smoothed inertial information as an input of SVM regression to train an SVM model, and during training, comparing the predicted results of the SVM model with the actual results, and calculating the optimal coefficients and parameters of the SVM model based on a traversal algorithm.

[0084] The process of wavelet decomposition can refer to the above embodiment. In order to further improve the signal quality, the signal after wavelet denoising can also be smoothed to eliminate subtle mutations or fluctuations and improve the continuity of the signal. The mean filter formula can be expressed as:

[0085]

[0086] y i is the signal value after filtering, x j is the denoised signal value, and n is the filter window size.

[0087] After wavelet transform and mean filtering, the short-term noise in the signal is removed, but it may still contain systematic errors (such as drift and cumulative errors). SVM regression further corrects these errors by learning the dynamic characteristics of inertial information. Specifically, the inertial information after denoising and smoothing is input, and the actual motion parameters are output. SVM maps the input signal to a high-dimensional feature space by constructing a nonlinear mapping:

[0088]

[0089] Among them, K(x i ,x)=exp(-γ||x i -x|| 2 ), which is the Gaussian kernel function, is the Lagrange multiplier and b is the bias term.

[0090] In addition, when training, SVM compares the input signal with the target value (real motion parameter), calculates the error, and then uses the ergodic algorithm to adjust the parameters of SVM (kernel function parameter and regularization coefficient), where the kernel function parameter γ determines the amplitude of the function, and the regularization coefficient C controls the balance between the complexity of the model and the error. The optimization objective function of the model is:

[0091]

[0092] ∈ is the tolerance range, which is used to ignore small errors.

[0093] Traverse different combinations of γ and C, select the parameter pair that minimizes the model error on the training set and the validation set, and obtain the SVM regression model.

[0094] In another embodiment, the SVM model uses a radial basis function as a kernel function.

[0095] In one embodiment, when training and constructing the fuzzy generalized learning model, it includes: defining the structure of the fuzzy generalized learning model, the structure of the fuzzy generalized learning model includes n fuzzy subsystems and m enhanced nodes; defining multiple fuzzy rules for each of the fuzzy subsystems, the fuzzy rules are used to describe the relationship between the inertial information and the position error; defining a Gaussian membership function for each fuzzy set, and determining its center and width based on the K-means clustering algorithm; calculating the activation strength of each of the fuzzy rules as the product of the Gaussian membership function of the inertial information; constructing the output function of each of the fuzzy subsystems based on the activation strength and the parameters of the fuzzy subsystems; organizing the inertial information into a training data set, representing it in a matrix, training and adjusting the parameters of the fuzzy subsystems, and obtaining the fuzzy generalized learning model.

[0096] Based on the same idea, Figure 2As shown, the embodiment of the present specification also provides an AI-based ship positioning system, the system comprising: an inertial sensing module 201, for acquiring a first position of a ship at a first time point based on GPS, acquiring inertial information of the ship between the first time point and a second time point, the inertial information comprising at least acceleration data and angular velocity data, and deriving a first predicted position of the ship at a second time point based on the inertial information and the first position; a training module 202, for acquiring a second position of the ship at the second time point based on GPS, calculating a prediction error between the second position and the first predicted position, and taking the inertial information as input and the prediction error as output based on a fuzzy prediction algorithm. The fuzzy generalized learning model is trained and constructed, and, before training, the inertial information is preprocessed based on wavelet denoising and SVM regression; the analysis module 203 is used to continuously obtain the predicted position and the predicted error to train and update the fuzzy generalized learning model when the GPS is valid, and take the position determined by the GPS as the actual position of the ship; when the GPS is invalid, the fuzzy generalized learning model is used to predict the position error according to the inertial information of the ship after the GPS is invalid, and the second predicted position of the ship is derived based on the inertial information and the position information of the ship determined before the GPS is invalid, and the predicted position error is compensated to the second predicted position to obtain the actual position of the ship.

[0097] It can be seen from the above embodiments that the position of the ship is derived based on inertial information when there is no GPS, so as to avoid the risk of shipping and driving caused by the long-term absence of GPS signals. Moreover, based on the constructed fuzzy generalized learning model, the derived position is compensated for the error, making the derived position more accurate and solving the integral error generated in the derivation process.

[0098] Based on the same idea, the embodiment of this specification also provides an AI-based ship positioning device, such as Figure 3 shown.

[0099] The AI-based ship positioning device may be the terminal device or server provided in the above-mentioned embodiment.

[0100] The AI-based ship positioning device may have relatively large differences due to different configurations or performances, and may include one or more processors 301 and memory 302, and the memory 302 may store one or more storage applications or data. Among them, the memory 302 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) and / or a cache memory unit, and may further include a read-only storage unit. The application stored in the memory 302 may include one or more program modules (not shown in the figure), such program modules include but are not limited to: an operating system, one or more applications, other program modules and program data, each of these examples or some combination may include the implementation of a network environment. Furthermore, the processor 301 can be configured to communicate with the memory 302 to execute a series of computer executable instructions in the memory 302 on the AI-based ship positioning device. The AI-based ship positioning device may also include one or more power supplies 303, one or more wired or wireless network interfaces 304, one or more I / O interfaces (input and output interfaces) 305, one or more external devices 306 (such as keyboards, pointing devices, Bluetooth devices, etc.), and may also communicate with one or more devices that enable a user to interact with the device, and / or communicate with any device that enables the device to communicate with one or more other computing devices (such as routers, modems, etc.). Such communication may be performed through the I / O interface 305. In addition, the device may also communicate with one or more networks (such as a local area network (LAN)) through the wired or wireless interface 304.

[0101] Specifically in this embodiment, the AI-based ship positioning device includes a memory and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer executable instructions in the AI-based ship positioning device, and the one or more programs are configured to be executed by one or more processors, including the following computer executable instructions:

[0102] Acquire a first position of the ship at a first time point based on GPS, acquire inertial information of the ship between the first time point and a second time point, the inertial information at least including acceleration data and angular velocity data, and derive a first predicted position of the ship at a second time point based on the inertial information and the first position;

[0103] Acquiring a second position of the ship at the second time point based on GPS, calculating a prediction error between the second position and the first predicted position, training and constructing a fuzzy generalized learning model based on a fuzzy prediction algorithm, taking the inertial information as input and the prediction error as output, and, before training, preprocessing the inertial information based on wavelet transform denoising and SVM regression;

[0104] When GPS is valid, continuously acquiring the predicted position and the predicted error to train and update the fuzzy generalized learning model, and using the position determined by GPS as the actual position of the ship;

[0105] When GPS is invalid, the fuzzy generalized learning model is used to predict the position error based on the inertial information of the ship after the GPS is invalid, and the second predicted position of the ship is derived based on the inertial information and the position information determined before the GPS is invalid, and the predicted position error is compensated to the second predicted position to obtain the actual position of the ship.

[0106] Based on the same idea, the exemplary embodiment of the present disclosure also provides a computer-readable storage medium on which a program product capable of implementing the above method of the present specification is stored. In some possible implementations, various aspects of the present disclosure can also be implemented in the form of a program product, which includes a program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps according to various exemplary embodiments of the present disclosure described in the above "AI-based ship positioning method" section of the present specification.

[0107] refer to Figure 4 As shown, a program product 400 for implementing the above method according to an exemplary embodiment of the present disclosure is described, which can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto. In this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, system or device.

[0108] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, system or device, or any combination of the above. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0109] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, wherein readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, system, or device.

[0110] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0111] Program code for performing the operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0112] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a terminal system, or a network device, etc.) to execute the method according to the exemplary implementation of the present disclosure.

[0113] In addition, the above-mentioned figures are only schematic illustrations of the processes included in the method according to the exemplary embodiment of the present disclosure, and are not intended to be limiting. It is easy to understand that the processes shown in the above-mentioned figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0114] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the exemplary embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.

[0115] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary technical means in the art that are not disclosed in the present disclosure. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.

Claims

1. A ship positioning method based on AI, characterized in that: The method comprises: Acquire a first position of the ship at a first time point based on GPS, acquire inertial information of the ship between the first time point and a second time point, the inertial information at least including acceleration data and angular velocity data, and derive a first predicted position of the ship at a second time point based on the inertial information and the first position; Acquiring a second position of the ship at the second time point based on GPS, calculating a prediction error between the second position and the first predicted position, training and constructing a fuzzy generalized learning model based on a fuzzy prediction algorithm, taking the inertial information as input and the prediction error as output, and, before training, preprocessing the inertial information based on wavelet transform denoising and SVM regression; When GPS is valid, continuously acquiring the predicted position and the predicted error to train and update the fuzzy generalized learning model, and using the position determined by GPS as the actual position of the ship; When GPS is invalid, the fuzzy generalized learning model is used to predict the position error based on the inertial information of the ship after the GPS is invalid, and the second predicted position of the ship is derived based on the inertial information and the position information determined before the GPS is invalid, and the predicted position error is compensated to the second predicted position to obtain the actual position of the ship.

2. The AI-based ship positioning method according to claim 1, characterized in that: When deriving the first predicted position of the ship at the second time point, the method specifically includes: Performing time-integrated calculations on the angular velocity and the acceleration respectively to obtain attitude changes and displacement changes of the ship; The first position is taken as a reference point, and the first predicted position is obtained by combining the posture change and the displacement change.

3. The AI-based ship positioning method according to claim 2, characterized in that: When performing time integral calculations, error compensation is performed based on a Kalman filter or a wavelet transform denoising method.

4. The AI-based ship positioning method according to claim 1, characterized in that: The preprocessing of the inertial information based on wavelet transform denoising and SVM regression includes: Processing the inertial information to remove noise based on a wavelet transform denoising algorithm, and performing smoothing based on a mean filter; The smoothed inertial information is used as an input for SVM regression to train an SVM model. During training, the predicted results of the SVM model are compared with the actual results, and the optimal coefficients and parameters of the SVM model are calculated based on the traversal algorithm.

5. The AI-based ship positioning method according to claim 4, characterized in that: When the inertial information is subjected to noise removal processing based on a wavelet transform denoising algorithm, the method includes: Performing multiple wavelet decompositions on the inertial information so that the inertial information is decomposed into different frequency components, and obtaining wavelet coefficients of different time scales and frequencies generated at each level of decomposition; Performing threshold processing on the wavelet coefficients, treating the wavelet coefficients smaller than the threshold as noise and setting them to zero, and retaining the wavelet coefficients larger than the threshold; The retained wavelet coefficients are reconstructed by inverse wavelet transform to obtain the denoised inertial information.

6. The AI-based ship positioning method according to claim 4, characterized in that: The SVM model uses radial basis function as kernel function.

7. The AI-based ship positioning method according to claim 1, characterized in that: When training and constructing the fuzzy generalized learning model, it includes: Defining a structure of the fuzzy generalized learning model, wherein the structure of the fuzzy generalized learning model includes n fuzzy subsystems and m enhancement nodes; defining a plurality of fuzzy rules for each of the fuzzy subsystems, wherein the fuzzy rules are used to describe the relationship between the inertial information and the position error; A Gaussian membership function is defined for each fuzzy set, and its center and width are determined based on the K-means clustering algorithm; calculating the activation strength of each of the fuzzy rules as a product of the Gaussian membership function of the inertial information; Based on the activation strength and the parameters of the fuzzy subsystem, construct an output function of each of the fuzzy subsystems; The inertial information is organized into a training data set and represented by a matrix, and the parameters of the fuzzy subsystem are trained and adjusted to obtain the fuzzy generalized learning model.

8. An AI-based ship positioning system, characterized in that: The system comprises: an inertial sensing module, configured to obtain a first position of the ship at a first time point based on GPS, obtain inertial information of the ship between the first time point and a second time point, wherein the inertial information includes at least acceleration data and angular velocity data, and derive a first predicted position of the ship at a second time point based on the inertial information and the first position; a training module, for obtaining a second position of the ship at the second time point based on GPS, calculating a prediction error between the second position and the first predicted position, training and constructing a fuzzy generalized learning model based on a fuzzy prediction algorithm, taking the inertial information as input and the prediction error as output, and, before training, preprocessing the inertial information based on wavelet denoising and SVM regression; The analysis module is used to continuously acquire the predicted position and the predicted error to train and update the fuzzy generalized learning model when the GPS is valid, and use the position determined by the GPS as the actual position of the ship; when the GPS is invalid, the fuzzy generalized learning model is used to predict the position error according to the inertial information of the ship after the GPS is invalid, and derive the second predicted position of the ship based on the inertial information and the position information of the ship determined before the GPS is invalid, and compensate the predicted position error to the second predicted position to obtain the actual position of the ship.

Images