Motion prediction method for ship-borne landing platform

Abstract

The invention discloses a motion prediction method for a ship-borne landing platform, and relates to the technical field of unmanned aerial vehicle landing, and the method comprises the following steps: S1, obtaining historical data, S2, estimating initial parameters, S3, analyzing the order of a model, S4, collecting real-time data, S5, updating model parameters, and S6, outputting prediction data. According to the method, the model parameter estimation values corresponding to different moments are calculated according to the sample data by using the recursive least square method, so that the finally determined model parameters can be more accurate, it is ensured that motion prediction during landing of the unmanned aerial vehicle can fit reality, the estimation error value corresponding to the current moment is analyzed by using the prediction analysis algorithm, and the prediction accuracy of the unmanned aerial vehicle is improved. The optimal model order is calculated according to the estimated error value, the model parameters and the order value are transmitted to the time sequence model, the model is used for accurately estimating the motion amount of the unmanned aerial vehicle on the ship-borne landing platform at the next moment, and the calculated amount required by motion prediction is simplified to a certain extent.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle landing, and in particular to a motion prediction method for a shipborne landing platform. Background Art

[0002] Shipborne drones have the characteristics of fast flight speed and wide field of view. They can be combined with various types of ships to expand the mission capabilities of ships. Compared with land-based landing platforms, ships will produce rolling, pitching and sinking movements during navigation due to the influence of factors such as wind, water flow and speed, resulting in single-sided landing, heavy landing and landing outside the landing area of ​​drones. In order to improve the safety of drone landing on shipborne platforms, it is necessary to estimate the platform movement before landing. In the invention patent with application number CN201910212992.7, a method and system for autonomous landing of drones based on hybrid navigation of GPS and image recognition is disclosed. The method includes: controlling the unmanned ship to stop moving according to the received landing command. , and transmit the GPS navigation coordinates of the landing pad on the unmanned ship to the drone in real time; control the drone to fly above the auxiliary landing platform and wait for landing; process the image of the landing mark taken by the drone to identify the center coordinates of the landing mark image; obtain the position, attitude angle and coordinate difference of the camera center relative to the landing mark based on the center coordinates of the captured image, the altitude data of the drone and its coordinate position; control the drone to land based on the position, attitude angle of the drone camera center relative to the landing mark and the attitude angle of the airborne gimbal camera relative to the drone. The present invention provides a high-precision, high-reliability, low-cost autonomous landing method for a rotary-wing drone, which is particularly suitable for landing a drone on an unmanned ship. "

[0003] The above-mentioned prior art solves the problem that the UAV cannot land accurately and autonomously, but in the analysis process, the model parameters at different times are not calculated, resulting in that the model parameters are not updated in time at different times, so that the final prediction result may be significantly different from the actual situation, and when estimating the movement amount of the UAV on the ship-borne landing platform, the required amount of calculation is large, and the prediction result cannot be obtained in the shortest time. Summary of the invention

[0004] The object of the present invention is to provide a motion prediction method for a shipborne landing platform to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a motion prediction method for a shipborne landing platform, comprising the following steps:

[0006] S1. Obtain historical data: query the database according to the ship number, extract the ship-borne IMU data and satellite positioning data corresponding to the number, and make these data into sample data;

[0007] S2. Estimation of initial parameters: After obtaining multiple sets of sample data, use the recursive least squares method to estimate the model parameters at different times based on these sample data, and input the estimation results into the time series model;

[0008] S3. Analyze the model order: After extracting the estimated model parameters, use the prediction analysis algorithm to calculate the sample data and model parameters to obtain the estimated error value, and use the BIC criterion function to calculate the optimal model order according to the estimated error value;

[0009] S4. Collecting real-time data: using inertial measurement equipment and satellite positioning devices to collect the onboard IMU data and satellite positioning data of the current ship in real time, and transferring these data to the storage device;

[0010] S5. Update model parameters: after retrieving the collected real-time data in the storage device, the data is analyzed using the recursive least squares method, and the obtained parameters are transmitted to the time series model;

[0011] S6. Output prediction data: After determining the parameters and order of the time series model, use the time series model to estimate the movement of the UAV on the shipborne landing platform at the next moment based on the collected real-time data, model parameters and noise sequence, and transmit the estimation results to the visualization interface.

[0012] Preferably, the step S1 specifically includes the following steps:

[0013] S101, after obtaining the current ship number, query the database according to the ship number to obtain all shipborne IMU data and satellite positioning data related to the ship number;

[0014] S102, combining the shipborne IMU data and the satellite positioning data at the same time, setting the number of groups of acquired data, and sorting all the data according to the recording time.

[0015] Preferably, the step S1 specifically further comprises the following steps:

[0016] S103, extracting data from the last group of data in the sequence in sequence according to the number of groups, and making the extracted shipborne IMU data and satellite positioning data into sample data.

[0017] Preferably, in step S2, the time series model is specifically:

[0018] y(t)=A(t)Y T (t)+e(t)

[0019] Wherein, y(t) represents the motion of the UAV landing platform at time t, and the motion specifically refers to the angle and displacement, A(t) represents the model parameters at time t, Y(t) represents the historical motion before time t, and e(t) represents the noise sequence;

[0020] The recursive least squares method is specifically:

[0021]

[0022] Among them, A(t) represents the model parameters at time t, Y(j) represents the historical motion before time j, A(t+1) represents the model parameters at time t+1, y(t+1) represents the motion of the UAV landing platform at time t+1, Y(t+1) represents the historical motion before time t+1, P(t+1) represents the covariance matrix corresponding to the parameters at time t+1, P(t) represents the covariance matrix corresponding to the parameters at time t, and M(t+1) represents the gain matrix corresponding to the parameters at time t+1.

[0023] Preferably, the step S3 specifically includes the following steps:

[0024] S301, after extracting the estimated model parameters, use the prediction analysis algorithm to calculate the sample data at the current moment and the model parameters at the previous moment to obtain the estimated error value corresponding to the current moment;

[0025] S302: After calculating corresponding estimated error values ​​according to multiple groups of sample data, the estimated error sum of squares is obtained by analyzing each estimated error value.

[0026] Preferably, the step S3 specifically further comprises the following steps:

[0027] S303: After obtaining the estimated error sum of squares, the BIC criterion function is used to select the order of the time series model according to the estimated error sum of squares and sample data to determine the optimal model order.

[0028] Preferably, in step S301, the prediction analysis algorithm is specifically:

[0029] err(t+1)=y(t+1)-Y(t+1)A T (t)

[0030] Among them, err(t+1) represents the estimated error value at time t+1, y(t+1) represents the movement of the UAV landing platform at time t+1, Y(t+1) represents the historical movement before time t+1, and A(t) represents the model parameters at time t.

[0031] Preferably, in step S303, the BIC criterion function is specifically:

[0032]

[0033] Among them, t represents time, N represents the total sample size, n represents the number of samples, and p,i represents parameters. represents the sum of squared errors of the estimates.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] The present invention uses the recursive least squares method to calculate the estimated values ​​of model parameters corresponding to different moments according to sample data, so that the model parameters finally determined can be more accurate. Compared with the traditional method, this method can provide estimated values ​​for a longer time, improve the safety of the landing of the UAV on the landing platform, and ensure that the motion prediction of the UAV during landing can be in line with reality. The prediction analysis algorithm is used to analyze the estimated error value corresponding to the current moment, and the optimal model order is calculated according to the estimated error value. The model parameters and order values ​​are transmitted to the time series model, and the model is used to accurately estimate the motion amount of the UAV on the ship-borne landing platform at the next moment, which simplifies the calculation amount required for motion prediction to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 Provide an overall method flow chart for an embodiment of the present invention;

[0037] Figure 2 A flow chart of a method for obtaining historical data provided by an embodiment of the present invention;

[0038] Figure 3 A flow chart of a method for analyzing model order provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] See also Figure 1-3 The present invention provides a technical solution: a motion prediction method for a shipborne landing platform, comprising the following steps:

[0041] S1. Obtain historical data: query the database according to the ship number, extract the ship-borne IMU data and satellite positioning data corresponding to the number, and make these data into sample data;

[0042] S2. Estimation of initial parameters: After obtaining multiple sets of sample data, use the recursive least squares method to estimate the model parameters at different times based on these sample data, and input the estimation results into the time series model;

[0043] S3. Analyze the model order: After extracting the estimated model parameters, use the prediction analysis algorithm to calculate the sample data and model parameters to obtain the estimated error value, and use the BIC criterion function to calculate the optimal model order according to the estimated error value;

[0044] S4. Collecting real-time data: using inertial measurement equipment and satellite positioning devices to collect the onboard IMU data and satellite positioning data of the current ship in real time, and transferring these data to the storage device;

[0045] S5. Update model parameters: after retrieving the collected real-time data in the storage device, the data is analyzed using the recursive least squares method, and the obtained parameters are transmitted to the time series model;

[0046] S6. Output prediction data: After determining the parameters and order of the time series model, use the time series model to estimate the movement of the UAV on the shipborne landing platform at the next moment based on the collected real-time data, model parameters and noise sequence, and transmit the estimation results to the visualization interface.

[0047] Step S1 specifically includes the following steps:

[0048] S101, after obtaining the current ship number, query the database according to the ship number to obtain all shipborne IMU data and satellite positioning data related to the ship number;

[0049] S102, combining the shipborne IMU data and the satellite positioning data at the same time, setting the number of groups of acquired data, and sorting all the data according to the recording time;

[0050] Step S1 specifically also includes the following steps:

[0051] S103, extracting data from the last group of data in the sequence in sequence according to the number of groups, and making the extracted shipborne IMU data and satellite positioning data into sample data.

[0052] Preferably, in step S2, the time series model is specifically:

[0053] y(t)=A(t)Y T (t)+e(t)

[0054] Among them, y(t) represents the motion of the UAV landing platform at time t, and the motion specifically refers to the angle and displacement, A(t) represents the model parameters at time t, Y(t) represents the historical motion before time t, and e(t) represents the noise sequence;

[0055] The recursive least squares method is as follows:

[0056]

[0057] Among them, A(t) represents the model parameters at time t, Y(j) represents the historical motion before time j, A(t+1) represents the model parameters at time t+1, y(t+1) represents the motion of the UAV landing platform at time t+1, Y(t+1) represents the historical motion before time t+1, P(t+1) represents the covariance matrix corresponding to the parameters at time t+1, P(t) represents the covariance matrix corresponding to the parameters at time t, and M(t+1) represents the gain matrix corresponding to the parameters at time t+1;

[0058] Step S3 specifically includes the following steps:

[0059] S301, after extracting the estimated model parameters, use the prediction analysis algorithm to calculate the sample data at the current moment and the model parameters at the previous moment to obtain the estimated error value corresponding to the current moment;

[0060] S302, after calculating corresponding estimated error values ​​according to multiple groups of sample data, using each estimated error value to analyze and obtain the estimated error sum of squares;

[0061] Step S3 specifically also includes the following steps:

[0062] S303, after obtaining the estimated error sum of squares, use the BIC criterion function to select the order of the time series model according to the estimated error sum of squares and sample data to determine the optimal model order;

[0063] In step S301, the prediction analysis algorithm is specifically:

[0064] err(t+1)=y(t+1)-Y(t+1)A T (t)

[0065] Among them, err(t+1) represents the estimated error value at time t+1, y(t+1) represents the movement of the UAV landing platform at time t+1, Y(t+1) represents the historical movement before time t+1, and A(t) represents the model parameters at time t;

[0066] In step S303, the BIC criterion function is specifically:

[0067]

[0068] Among them, t represents time, N represents the total sample size, n represents the number of samples, and p,i represents parameters. represents the sum of squared errors of the estimates.

[0069] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0070] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A motion prediction method for a shipborne landing platform, characterized in that: The method comprises the following steps: S1. Obtain historical data: query the database according to the ship number, extract the ship-borne IMU data and satellite positioning data corresponding to the number, and make these data into sample data; S2. Estimation of initial parameters: After obtaining multiple sets of sample data, use the recursive least squares method to estimate the model parameters at different times based on these sample data, and input the estimation results into the time series model; S3. Analyze the model order: After extracting the estimated model parameters, use the prediction analysis algorithm to calculate the sample data and model parameters to obtain the estimated error value, and use the BIC criterion function to calculate the optimal model order according to the estimated error value; S4. Collecting real-time data: using inertial measurement equipment and satellite positioning devices to collect the onboard IMU data and satellite positioning data of the current ship in real time, and transferring these data to the storage device; S5. Update model parameters: after retrieving the collected real-time data in the storage device, the data is analyzed using the recursive least squares method, and the obtained parameters are transmitted to the time series model; S6. Output prediction data: After determining the parameters and order of the time series model, use the time series model to estimate the movement of the UAV on the shipborne landing platform at the next moment based on the collected real-time data, model parameters and noise sequence, and transmit the estimation results to the visualization interface.

2. The motion prediction method of a shipborne landing platform according to claim 1, characterized in that: The step S1 specifically includes the following steps: S101, after obtaining the current ship number, query the database according to the ship number to obtain all shipborne IMU data and satellite positioning data related to the ship number; S102, combining the shipborne IMU data and the satellite positioning data at the same time, setting the number of groups of acquired data, and sorting all the data according to the recording time.

3. The motion prediction method of a shipborne landing platform according to claim 2, characterized in that: The step S1 specifically further comprises the following steps: S103, extracting data from the last group of data in the sequence in sequence according to the number of groups, and making the extracted shipborne IMU data and satellite positioning data into sample data.

4. The motion prediction method of a shipborne landing platform according to claim 1, characterized in that: In step S2, the time series model is specifically: y(t)=A(t)Y T (t)+e(t) Among them, y(t) represents the movement of the UAV landing platform at time t, and the movement specifically refers to the angle and displacement, A(t) represents the model parameters at time t, Y(t) represents the historical movement before time t, and e(t) represents the noise sequence.

5. The motion prediction method of a shipborne landing platform according to claim 1, characterized in that: The step S3 specifically comprises the following steps: S301, after extracting the estimated model parameters, use the prediction analysis algorithm to calculate the sample data at the current moment and the model parameters at the previous moment to obtain the estimated error value corresponding to the current moment; S302: After calculating corresponding estimated error values ​​according to multiple groups of sample data, the estimated error sum of squares is obtained by analyzing each estimated error value.

6. The motion prediction method of a shipborne landing platform according to claim 5, characterized in that: The step S3 specifically further comprises the following steps: S303: After obtaining the estimated error sum of squares, the BIC criterion function is used to select the order of the time series model according to the estimated error sum of squares and sample data to determine the optimal model order.

7. The motion prediction method of a shipborne landing platform according to claim 5, characterized in that: In step S301, the prediction analysis algorithm is specifically: err(t+1)=y(t+1)-Y(t+1)A T (t) Among them, err(t+1) represents the estimated error value at time t+1, y(t+1) represents the movement of the UAV landing platform at time t+1, Y(t+1) represents the historical movement before time t+1, and A(t) represents the model parameters at time t.

8. The motion prediction method of a shipborne landing platform according to claim 6, characterized in that: In step S303, the BIC criterion function is specifically: Among them, t represents time, N represents the total sample size, n represents the number of samples, and p,i represents parameters. represents the sum of squared errors of the estimates.

Images