A ship attitude warning method for a Beidou integrated machine

Abstract

The application provides a ship attitude warning method for a Beidou integrated machine, which comprises a ship attitude monitoring part and a warning and rescue part; wherein the ship attitude monitoring part judges whether to warn and transmits warning information to the warning and rescue part through three methods, i.e., manual warning, initial state abnormality warning of the ship and predicted attitude warning of the ship; after receiving the warning information, the warning and rescue part sends the warning information through a Beidou satellite and assists in rescue. The Beidou integrated machine installed on the ship is positioned through a satellite to meet the attitude monitoring and distress warning functions of the ship through the short message communication function of the Beidou integrated machine.

Description

TECHNICAL FIELD

[0001] The present application relates to a ship attitude warning method, in particular to a ship attitude warning method for Beidou integrated machine. BACKGROUND

[0002] With the increasing development of marine economy, the types of marine activities are gradually increasing. Due to the influence of weather and ocean currents, the marine environment is complex and changeable. How to accurately obtain the state information of the ship has become an important problem to ensure the safety of marine activities. The Beidou satellite navigation system (BDS) has realized global coverage in the available range. The Beidou short message communication function is one of the important functions of BDS. The ground terminal can directly transmit information to other ground terminals through satellite signals through the Beidou short message function, which can be used for emergency communication in emergency situations. At present, the ship usually carries Beidou integrated machine equipment. The Beidou integrated machine is a combination of navigation and communication terminal, which can provide accurate and reliable positioning, navigation, timing (PNT) information for ships according to the observation information of global satellite navigation system (GNSS) and inertial navigation system, and can realize data communication through the Beidou short message function in any area.

[0003] When encountering severe weather at sea, strong wind and waves will cause the ship to shake violently, and even cause the ship to overturn and other accidents. The existing ship attitude warning method cannot balance the attitude estimation accuracy and the warning speed. The real-time ship attitude estimation method cannot realize the prediction of the ship attitude, and the ship accident warning is not timely. The ship attitude prediction method based on IMU (Inertial Measurement Unit) data has the problem of attitude error accumulation, which will be more prominent when the ship is sailing in the open sea area. Secondly, the existing ship attitude warning method only sends an alarm to the ship operation personnel when the ship attitude is detected to be abnormal, and the alarm information cannot be effectively interacted with other objects and assist in rescue.

[0004] In the prior patent application file "A method for shipborne Beidou integrated machine tilt monitoring", the sensor observation data obtained by the shipborne Beidou integrated machine is used to calculate the position, speed and acceleration information of the Beidou integrated machine, and the tilt angle of the Beidou integrated machine is calculated based on the above information based on motion analysis, and it is judged whether the tilt angle is greater than a threshold value, then the GNSS / IMU observation information is fused by using the robust Kalman filter, and the ship motion constraint is executed, and finally the integrated machine tilt angle after information fusion is obtained, and it is judged whether it is greater than a threshold value and whether it needs to be alarmed. But this ship attitude alarm method based on motion model analysis and derivation has the problem of being too rough: first, the ship attitude is not directly obtained but is derived through ship motion analysis, and when the ship shakes violently, the attitude determination result may have a large error; second, the alarm of this method is only for the tilt of the Beidou integrated machine to the ship operation personnel, and lacks the function of communicating to the outside world, and cannot provide help for the search and rescue of the ship in distress, so this method cannot realize the distress alarm function of the ship in the open sea operation. SUMMARY

[0005] The technical problem to be solved by the present application is to provide a ship attitude alarm method for a Beidou integrated machine.

[0006] In order to solve the above technical problems, the present application discloses a ship attitude alarm method for a Beidou integrated machine, which comprises a ship attitude monitoring part and an alarm and rescue part; wherein the ship attitude monitoring part judges whether to alarm and transmits the alarm information to the alarm and rescue part through three methods of manual alarm, ship initial state abnormal alarm and ship predicted attitude alarm; the alarm and rescue part receives the alarm information, sends the alarm information through Beidou satellite, and assists in rescue.

[0007] Further, the ship attitude monitoring part specifically comprises:

[0008] Step 1, judge whether to perform manual alarm, if manual alarm is selected, directly enter the alarm and rescue part, otherwise enter step 2;

[0009] Step 2, use the observation data of the IMU embedded in the Beidou integrated machine to judge whether the initial estimated attitude of the ship appears abnormal observation value, if abnormal, enter the alarm and rescue part, otherwise enter step 3;

[0010] Step 3, determine the ship position based on pseudorange single point positioning according to the GNSS observation value;

[0011] Step 4, perform ship accurate attitude estimation by TDCP and IMU combined navigation of adaptive Kalman filter, to obtain the ship attitude estimation result based on combined navigation;

[0012] Step 5, ship attitude prediction and warning based on VMD-LSTM.

[0013] Further, step 2 uses the observation data of the IMU embedded in the Beidou integrated machine to determine whether the initial estimated attitude of the ship has abnormal observation values, specifically including:

[0014] Step 2-1, IMU observation data acquisition: according to the data obtained by the accelerometer and gyroscope in the IMU of the Beidou integrated machine, the mechanical arrangement is carried out, and the ship attitude based on IMU observation at epoch k is obtained:

[0015]

[0016] where, represents the pitch angle of the ship at epoch k, represents the roll angle of the ship at epoch k, represents the yaw angle of the ship at epoch k;

[0017] Step 2-2, check whether the ship attitude is abnormal: detect the pitch angle and roll angle of the ship, and ω tre respectively represent the maximum safe pitch angle and the maximum safe roll angle of the ship, when or , it is determined that there is an abnormal observation value, i.e. the ship attitude is in an unsafe state, and immediately enter the warning and rescue part.

[0018] Further, step 3 determines the ship position based on pseudorange single point positioning according to GNSS observation values, specifically including:

[0019] Let the number of visible satellites of the GNSS receiver in the Beidou integrated machine at the current time be The coordinates of the jth visible satellite in the Earth-Centered Earth-Fixed coordinate system are The pseudorange observation value corresponding to the jth visible star is For each visible satellite, the following pseudorange observation equation is satisfied:

[0020]

[0021] where, and are the GNSS receiver clock error and satellite clock error respectively, and represent the ionospheric error, tropospheric error and unmodeled error at the current epoch k, P k is the coordinate of the GNSS receiver in the Earth-Centered Earth-Fixed coordinate system; by simultaneously solving the pseudorange observation equations of all visible satellites, when , the position P k of the Beidou integrated machine at the current epoch is solved.

[0022] Furthermore, step 4, which describes the use of adaptive Kalman filtering in TDCP and IMU integrated navigation for accurate ship attitude estimation, specifically includes:

[0023] Step 4-1, using inter-epoch differential carrier phase TDCP to solve for the moving speed of the BeiDou integrated machine, specifically including:

[0024] Let the observed carrier phase value for each visible star at the current epoch k be... The carrier phase observation value of the previous epoch is The coordinates of visible stars in the current epoch and the previous epoch in the Earth-centered Earth-fixed coordinate system are respectively and For each visible satellite, the following TDCP observation equations are satisfied:

[0025]

[0026] Among them, carrier phase difference Let j be the unit vector between the GNSS receiver and satellite j at the current epoch. Let J be the interepoch displacement of satellite j. c is the speed of light, and λ is the carrier wavelength. For the displacement of the BeiDou integrated machine between epochs; establish the TDCP observation equations for all visible satellites, when At that time, calculate the moving speed V of the Beidou integrated machine in the current epoch k. k as follows:

[0027]

[0028] Where Δt is the sampling interval of the GNSS receiver.

[0029] Step 4-2, constructing an adaptive Kalman filter to estimate the ship's attitude, specifically including:

[0030] The recursive formula for adaptive Kalman filtering is expressed as:

[0031] X k,k-1 =Φ k,k-1 X k-1

[0032] Among them, X k,k-1 Let Φ be the one-step predicted state vector predicted from the k-1 epoch. k,k-1 Represents the state transition matrix;

[0033] X k-1 The variance-covariance matrix ∑ k,k-1 as follows:

[0034]

[0035] wherein Q k-1 is the motion model error variance matrix;

[0036] the prediction residual Res k is as follows:

[0037] Res k = Z k - H k X k,k-1

[0038] wherein Z k is the observation vector, i.e. the Beidou integrated machine moving speed V k in step 4-1, H k is the observation matrix;

[0039] the gain matrix K k is as follows:

[0040]

[0041] wherein R k is the observation noise matrix, μ k represents the adaptive factor;

[0042] The adaptive factor is constructed according to the error discrimination statistic:

[0043]

[0044] wherein M is a constant, tr(·) is the trace of a matrix, is the error discrimination statistic:

[0045]

[0046] The state update is:

[0047] X k = X k-1 + K k Res k

[0048] wherein X k is the estimated state vector; the covariance matrix ∑ k of the Kalman filter estimated state is as follows:

[0049]

[0050] wherein I is the identity matrix;

[0051] Finally, the final state estimation result X kThe ship attitude estimation result based on the integrated navigation wherein ω k and θ k respectively represent the ship attitude estimated by adaptive Kalman filtering, i.e., the ship pitch angle, the ship roll angle and the ship yaw angle.

[0052] Further, the ship attitude prediction and alarm based on VMD-LSTM in step 5 specifically comprises:

[0053] Step 5-1, constructing a ship attitude sequence: statistics of the ship attitude in a set time period, wherein the ship attitude estimation result based on the integrated navigation contains d epochs, the ship attitude sequence is A = [A k-d A k-d+1 A k-d+2 …A k ];

[0054] Step 5-2, decomposing the ship attitude sequence based on the VMD method: first, selecting parameters, including the number of decompositions m and the penalty function α; then decomposing the ship attitude sequence based on the VMD method to obtain modal components IMF1, IMF2…IMF n and residual signal ξ n ;

[0055] Further, the number of decompositions m and the penalty function α are set to 7 and 100 respectively.

[0056] Step 5-3, training and prediction based on the LSTM recurrent neural network:

[0057] Using the training data set to pre-train the LSTM model, learning the time sequence rule of the modal components and the residual signal, and adjusting the parameters of the LSTM model according to the training result;

[0058] Using the modal components and the residual signal obtained in step 5-2, the pre-trained LSTM model is used for prediction, and the prediction results are superimposed to obtain the ship attitude prediction result based on the historical ship attitude data: wherein, represents the ship pitch angle predicted based on the historical ship attitude data, represents the ship roll angle predicted based on the historical ship attitude data, represents the ship yaw angle predicted based on the historical ship attitude data.

[0059] Step 5-4, determining whether to alarm based on the ship roll angle predicted based on the historical ship attitude data:

[0060] When Or is determined to be abnormal, immediately enter the alarm rescue part, otherwise, no alarm is given and return to step 1.

[0061] Further, the alarm rescue part specifically comprises:

[0062] Step b1: alarm information broadcast based on Beidou short message: when the alarm rescue part receives the alarm information sent by the attitude monitoring part, the Beidou integrated machine on the ship first sends the pre-set receiver ID number and the ship position calculated in step 3 to the ground center station through Beidou satellite after signal encryption; the ground center station receives the signal and hands it over to the Beidou satellite for broadcasting; the receiver's Beidou integrated machine receives the alarm information sent by the alarm ship after demodulating the signal;

[0063] Step b2: continuous monitoring and broadcast of position information based on Beidou short message: after sending the alarm information for the first time, the Beidou short message communication function remains open; when no rescue response information is received, the ship position estimation based on pseudorange single point positioning is repeated, and the distress position is sent and updated in real time through the Beidou short message; when the Beidou short message containing the rescue response information is received, the sound and light equipment linked with the Beidou integrated machine is started according to the indication of the rescue information.

[0064] Beneficial effects:

[0065] 1. The application constructs the automatic alarm function of the Beidou integrated machine for the ship in distress, and only through the satellite positioning of the Beidou integrated machine installed on the ship and the Beidou short message communication function to meet the attitude monitoring and distress alarm function of the ship.

[0066] 2. The application adopts the TDCP / IMU combined (TDCP, time-differenced carrier phase) ship attitude monitoring method which can calculate the ship attitude in real time with high precision without GNSS base station, and performs short-term prediction of the attitude based on VMD-LSTM method (VMD, variational mode decomposition) according to the ship attitude information obtained by the TDCP / IMU combined navigation system.

[0067] 3. The application realizes the remote communication of the alarm information by Beidou short message, and continuously sends the real-time position of the Beidou integrated machine through the Beidou short message. BRIEF DESCRIPTION OF DRAWINGS

[0068] The above and / or other aspects of the present application will become more apparent by describing in detail the preferred embodiments thereof with reference to the attached drawings, wherein:

[0069] Figure 1 The above and / or other aspects of the present application will become more apparent by describing in detail the preferred embodiments thereof with reference to the attached drawings, wherein: DETAILED DESCRIPTION

[0070] The present application proposes a ship attitude warning method for Beidou integrated machine, which aims to build an automatic warning function for Beidou integrated machine of the ship in distress, and only through the satellite positioning of the Beidou integrated machine installed on the ship and the short message communication function of Beidou to meet the ship's attitude monitoring and distress warning function requirements. Figure 1 As shown in the figure, the specific technical solutions are as follows:

[0071] I. Ship attitude monitoring part:

[0072] Step 1: Determine whether to perform manual warning: The ship operator manually determines whether to perform manual warning. When the ship has problems such as power system interruption and power system failure, manual warning can be enabled. If manual warning is selected, it directly enters the warning and rescue module; if there is no manual warning operation, it enters the next step.

[0073] Step 2: Determine whether the initial estimated attitude of the ship appears abnormal observation value by using the observation data of the IMU embedded in the Beidou integrated machine, which includes the following steps:

[0074] (1) IMU observation data acquisition. According to the data obtained by the accelerometer and gyroscope in the IMU, the mechanical arrangement is performed to obtain the ship attitude based on IMU observation at epoch k Wherein represents the pitch angle of the ship at epoch k, represents the roll angle of the ship at epoch k, represents the yaw angle of the ship at epoch k.

[0075] (2) Check whether the ship attitude is abnormal. The pitch angle and roll angle of the ship are detected, and ω tre respectively represent the maximum safe pitch angle and the maximum safe roll angle of the ship. When the ship attitude angle exceeds these two limits, that is, or , it is considered that the ship attitude is in an unsafe state, and immediately enters the warning mode, otherwise it is considered that the current initial attitude does not appear abnormal observation value.

[0076] Step 3: Determine the ship position based on pseudorange single point positioning according to GNSS observation value: The number of visible satellites of the GNSS receiver at the current time is The coordinates of the visible stars in the Earth-Centered Earth-Fixed coordinate system are The pseudo-range observation value corresponding to each visible star is For each visible satellite, the pseudo-range observation equation can be satisfied Wherein and are the receiver clock error and the satellite clock error, respectively, indicate the ionospheric error, the tropospheric error and the un-modeled error at the current time, respectively, and P k is the coordinate of the GNSS receiver in the Earth-Centered Earth-Fixed coordinate system. When , the position P k of the Beidou integrated machine at the current epoch can be solved.

[0077] The fourth step is to perform accurate ship attitude estimation based on adaptive Kalman filtering TDCP / IMU integrated navigation, and the steps include:

[0078] (1) TDCP velocity solving. The carrier phase observation value corresponding to each visible star at the current epoch k is The carrier phase observation value at the previous epoch is The coordinates of the visible stars in the Earth-Centered Earth-Fixed coordinate system at the current epoch and the previous epoch are and For each visible satellite, the TDCP observation equation can be satisfied Wherein is the unit vector between the receiver and satellite j at the current epoch, is the inter-epoch displacement of satellite j, c is the speed of light, λ is the carrier wavelength, and is the inter-epoch displacement of the Beidou integrated machine. Therefore, when , the moving speed of the Beidou integrated machine at the current epoch k can be obtained Δt is the receiver sampling interval, which is usually selected as 0.1 seconds or 1 second.

[0079] (2) Adaptive Kalman filter construction. The recursive formula of the adaptive Kalman filter can be expressed as: X k,k-1 = Φ k,k-1 X k-1 , wherein X k,k-1 is a one-step predicted state vector predicted from the state of epoch k-1, Φ k,k-1 represents a state transition matrix. The variance covariance matrix of X k-1 is Q k-1 , which is the motion model error variance matrix. Resk = Z k -H k X k,k-1 , Z k is the observation vector, here the observation vector input is the velocity V solved by TDCP k , H k is the observation matrix, Res k is the prediction residual. K k and R k are the gain matrix and observation noise matrix respectively, μ k represents the adaptive factor, which is constructed according to the formula , where m is a constant, usually 1-1.5, tr(·) is the trace of the matrix, is the error discrimination statistic, which is expressed as (reference: Liu Zhengwu, Sun Rui, Jiang Lei. GNSS / IMU robust adaptive positioning algorithm based on pseudo-range residual and innovation [J / OL]. Journal of Beijing University of Aeronautics and Astronautics. https: / / doi.org / 10.13700 / j.bh.1001-5965.2022.0389). The state update is: X k = X k-1 + K k Res k , X k is the estimated state vector. ∑ k is the covariance matrix of Kalman filter estimation state, I is the unit vector. Finally, the final state estimation result X k at the current epoch k is obtained, which includes the ship attitude estimation result based on integrated navigation where ω k , θ k represent the more accurate ship pitch angle, roll angle and yaw angle solved by integrated navigation respectively.

[0080] Step 5: Ship attitude prediction and alarm based on VMD-LSTM, the steps include:

[0081] (1) Construct the ship attitude time sequence. Statistics of ship attitude angle in a certain period of time, which contains d epochs of integrated navigation attitude estimation results, A = [A k-d A k-d+1 A k-d+2 …A k ].

[0082] (2) Decomposition of ship attitude sequence based on VMD method. First, select appropriate parameters for VMD, including the number of decomposition q and the penalty function a. Usually set q to 7 and a to 100; then calculate the relevant analysis signal of each mode through Hilbert transform to obtain the one-way spectrum; by mixing the spectrum with the exponential tuned to the corresponding estimated center frequency, the spectrum of each mode is moved to the baseband; finally, the bandwidth of the signal is estimated according to the solution of the Gaussian smoothness of each mode signal, and the decomposition of the ship attitude sequence is completed to obtain IMF1, IMF2…IMF n Equal mode components and residual signal ξ n , complete the smoothing of ship attitude data (reference: Dragomiretskiy K, Zosso D. Variational Mode Decomposition [J]. IEEE Transactions on Signal Processing, 2014, 62(3): 531-544.).

[0083] (3) Training and prediction based on LSTM. Use LSTM model to train the time series rule of each IMF component and residual signal in advance. For the input data, LSTM calculates the output value of each neuron forward, then calculates the error value of each neuron backward, and propagates the error term to the upper layer. Finally, according to the corresponding error term, the gradient of each weight is calculated and adjusted, and the training of LSTM is finally completed to build the time series relationship of IMF components. When using, the IMF components and residual signals decomposed by VMD are used to predict the time series through the LSTM model, and the signal intensity predicted by each component is obtained. Finally, the predicted signals are superimposed to obtain the predicted ship attitude based on the historical ship attitude data

[0084] (4) Determine whether to alarm based on the predicted ship roll angle. When or , it is considered that the ship attitude is about to be dangerous, and the alarm mode is immediately entered; otherwise, it is considered that the ship attitude is temporarily safe, and no alarm is given, returning to the first step.

[0085] II. Alarm and rescue part

[0086] First step: Broadcast of alarm information based on Beidou short message. When the ship attitude is detected to be abnormal and needs to be alarmed, the Beidou integrated machine first encrypts the pre-set receiver ID number and ship position application signal, and then transmits it to the ground center station through the Beidou satellite. After receiving the communication application signal sent by the user, the ground center station decrypts and encrypts it and adds it to the broadcast text, which is broadcast to the user by the satellite. After the selected user receiver demodulates the text, it can receive the alarm information sent by the ship.

[0087] Second step: based on the Beidou short message continuous monitoring and broadcast position information. After the first sending of alarm information, the Beidou short message communication function remains open. When no rescue response information is received, the ship position estimation based on pseudorange single point positioning is repeated, and the real-time sending of distress position is updated through the Beidou short message. In order to ensure the endurance of the Beidou integrated machine alarm and rescue, low frequency positioning and communication are adopted, and positioning and Beidou short message alarm information are sent once every 5 minutes. When receiving the Beidou short message containing the rescue response information, the sound and light equipment linked with the Beidou integrated machine is started according to the indication of the rescue information, so as to facilitate the rescue personnel to search the position of the ship.

[0088] In the specific implementation, the present application provides a computer storage medium and a corresponding data processing unit, wherein the computer storage medium can store a computer program, and the computer program can run the invention content of a ship attitude alarm method for a Beidou integrated machine and part or all steps in each embodiment when executed by the data processing unit. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0089] Those skilled in the art can clearly understand that the technical solutions in the embodiments of the present application can be realized by means of a computer program and its corresponding general hardware platform. Based on such understanding, the technical solutions in the embodiments of the present application can be embodied in the form of a computer program, i.e. a software product, which can be stored in a storage medium, including a plurality of instructions for causing a device (which can be a personal computer, a server, a single-chip microcomputer, a MUU or a network device, etc.) containing a data processing unit to execute the method described in each embodiment or some parts of the embodiments of the present application.

[0090] The present application provides a thought and method for a ship attitude alarm method for a Beidou integrated machine, and there are many methods and ways to realize the technical solutions. The above description is only the preferred embodiment of the present application, and it should be pointed out that for ordinary skilled persons in the technical field, without departing from the principles of the present application, some improvements and refinements can be made, which should also be regarded as the protection scope of the present application. The components not explicitly described in the embodiments can be realized by using existing technology.

Claims

1. A ship attitude warning method for Beidou integrated machine, characterized in that, The utility model relates to a kind of ship attitude monitoring and alarm system, including: Ship attitude monitoring part and alarm rescue part;Wherein, ship attitude monitoring part is judged whether to alarm by three methods of manual alarm, ship initial state abnormal alarm and ship predicted attitude alarm, and alarm information is transmitted to alarm rescue part;After receiving alarm information, alarm rescue part sends alarm information by Beidou satellite, and assists rescue; Wherein, the ship attitude monitoring part specifically includes: Step 1, judge whether to carry out manual alarm, if manual alarm is selected, then directly enter alarm rescue part, otherwise enter step 2; Step 2, use the observation data of IMU embedded in Beidou integrated machine to judge whether the initial estimated attitude of ship appears abnormal observation value, if abnormal, then enter alarm rescue part, otherwise enter step 3; Step 3, determine the position of ship based on pseudorange single point positioning according to GNSS observation value; Step 4, use TDCP combined with IMU navigation of adaptive Kalman filter to carry out accurate attitude estimation of ship, and obtain ship attitude estimation result based on combined navigation; Step 5, ship attitude prediction and alarm based on VMD-LSTM; Step 4 uses TDCP combined with IMU navigation of adaptive Kalman filter to carry out accurate attitude estimation of ship, specifically including: Step 4-1, uses epoch difference carrier phase TDCP to solve the moving speed of Beidou integrated machine; Step 4-2, constructs adaptive Kalman filter to estimate ship attitude; Step 4-1 uses epoch difference carrier phase TDCP to solve the moving speed of Beidou integrated machine, specifically including: Let the carrier phase observation value corresponding to each visible star at the current epoch k be The carrier phase observation value at the previous epoch is The coordinates of the visible stars in the Earth-Centered Earth-Fixed coordinate system at the current epoch and the previous epoch are respectively and For each visible satellite, the following TDCP observation equation is satisfied: where the carrier phase difference is the unit vector between the current epoch GNSS receiver and satellite j, is the inter-epoch displacement of satellite j, c is the speed of light, and λ is the carrier wavelength, is the inter-epoch Beidou integrated machine displacement; the TDCP observation equation of all visible satellites is solved simultaneously, when the Beidou integrated machine moving speed V k is obtained at the current epoch k as follows: Wherein, Δt is the sampling interval of GNSS receiver.

2. The ship attitude warning method for Beidou all-in-one machine according to claim 1, characterized in that, Step 2 uses the observation data of IMU embedded in Beidou integrated machine to judge whether the initial estimated attitude of ship appears abnormal observation value, specifically including: Step 2-1, IMU observation data acquisition: according to the data obtained by accelerometer and gyroscope in IMU of Beidou integrated machine, mechanical arrangement is carried out, and epoch k based on IMU observation ship attitude is obtained: wherein, denotes the pitch angle of the ship at the epoch k, denotes the roll angle of the ship at the epoch k, denotes the yaw angle of the ship at the epoch k; Step 2-2, check if the ship attitude is abnormal: the pitch angle and roll angle of the ship are detected, and ω tre respectively represent the maximum safe pitch angle and the maximum safe roll angle of the ship, when or , it is determined that an abnormal observation value occurs, that is, the ship attitude is in an unsafe state, and immediately enters the alarm and rescue part.

3. The ship attitude warning method for Beidou all-in-one machine according to claim 2, characterized in that, Step 3 determines the position of ship based on pseudorange single point positioning according to GNSS observation value, specifically including: Let the number of visible satellites of the GNSS receiver in the Beidou integrated machine at the current time be The coordinates of the jth visible satellite in the Earth-Centered Earth-Fixed coordinate system are The pseudo-range observation value corresponding to the jth visible satellite is For each visible satellite, the following pseudo-range observation equation is satisfied: where, and are the GNSS receiver clock error and satellite clock error, respectively, and denote the ionospheric error, tropospheric error and un-modeled error at current epoch k, respectively, P k is the coordinate of GNSS receiver in the Earth-Centered Earth-Fixed coordinate system; the pseudorange observation equation of all visible satellites is solved simultaneously, and when the position P k of the current epoch of Beidou integrated machine is solved.

4. The ship attitude warning method for Beidou all-in-one machine according to claim 3, characterized in that, Step 4-2 constructs adaptive Kalman filter to estimate ship attitude, specifically including: The recursive formula of adaptive Kalman filter is expressed as: X k,k-1 = Φ k,k-1 X k-1 where X k,k-1 is a one-step predicted state vector predicted from epoch k-1 state, Φ k,k-1 denotes a state transition matrix, X k-1 is the state at epoch k-1; X k-1 covariance matrix ∑ of the variance k,k-1 as follows: where Q k-1 is the motion model error variance matrix; Predicted residual Res k As follows: Res k = Z k - H k X k,k-1 Wherein, Z k is an observation vector, that is, the Beidou integrated machine moving speed V k , H k is an observation matrix; Gain matrix K k As follows: wherein R k is an observation noise matrix, μ k denotes an adaptation factor; Adaptive factor is constructed according to error discrimination statistic: where M is a constant, tr(·) is the trace of a matrix, is the error discrimination statistic: State update is: X k = X k-1 + K k Res k where X k is the estimated state vector; the Kalman filter estimates the covariance matrix ∑ k as follows: Wherein, I is unit matrix; Finally, the final state estimation result X at the current epoch k is obtained k wherein the ship attitude estimation result based on integrated navigation is included wherein ω k and θ k respectively represent the ship attitude estimated by adaptive Kalman filtering, i.e. the ship pitch angle, roll angle and yaw angle.

5. The ship attitude warning method for Beidou all-in-one machine according to claim 4, characterized in that, Step 5 ship attitude prediction and alarm based on VMD-LSTM, specifically including: Step 5-1, constructing the ship attitude sequence: statistics of the ship attitude in the set time period, which contains d epochs of ship attitude estimation results based on integrated navigation, the ship attitude sequence is: A = [A k-d A k-d+1 A k-d+2 … A k ] Step 5-2, decompose the ship attitude sequence based on the variational mode decomposition VMD method: first, select parameters, including the number of decompositions m and the penalty function a; then, decompose the ship attitude sequence based on the variational mode decomposition VMD method to obtain mode components IMF1, IMF2…IMF n and a residual signal ξ n ; Step 5-3, based on long short-term memory recurrent neural network LSTM for training and prediction: Use training data set to pretrain long short-term memory recurrent neural network LSTM model, learn the time sequence law of modal component and residual signal, and adjust the parameters of LSTM model according to the training result; Using the modal components and residual signals obtained in step 5-2, prediction is performed through the pre-trained LSTM model, and the prediction results are superimposed to obtain the ship attitude prediction results based on the historical attitude data of the ship: wherein, represents the pitch angle of the ship predicted based on the historical attitude data of the ship, represents the roll angle of the ship predicted based on the historical attitude data of the ship, represents the yaw angle of the ship predicted based on the historical attitude data of the ship; Step 5-4, judge whether to alarm based on ship roll angle determination of historical attitude data prediction: When or then an exception is determined, and the alarm rescue part is immediately entered, otherwise, no alarm is performed, and the process returns to step 1.

6. The ship attitude warning method for Beidou all-in-one machine according to claim 5, characterized in that, The alarm rescue part specifically includes: Step b1: alarm information broadcast based on Beidou short message: when the alarm rescue part receives the alarm information sent by the attitude monitoring part, the Beidou integrated machine on the alarm ship first encrypts the pre-set receiver ID number and the ship position calculated in step 3, and then transmits it to the ground center station through the Beidou satellite; after receiving the signal, the ground center station is handed over to the Beidou satellite for broadcasting; the receiver's Beidou integrated machine demodulates the alarm information sent by the alarm ship after receiving the alarm information; Step b2: continuously monitor and broadcast position information based on Beidou short message: after sending the alarm information for the first time, the Beidou short message communication function remains open; when no rescue response information is received, the ship position estimation based on pseudorange single point positioning is repeated, and the real-time sending and updating of the distress position is carried out through the Beidou short message; when receiving the Beidou short message containing the rescue response information, the sound and light equipment linked with the Beidou integrated machine is started according to the indication of the rescue information.

7. The ship attitude warning method for Beidou all-in-one machine according to claim 6, characterized in that, The number of decompositions m and the penalty function a in step 5-2 are set to 7 and 100 respectively.

Images