An intelligent ship piloting method and system based on RTK and inertial navigation fusion

Abstract

The application relates to an intelligent ship pilotage method and system based on RTK and inertial navigation fusion, which comprises the following steps: deploying and calibrating an RTK base station at a high point of a wharf, generating a differential correction quantity in real time and broadcasting the differential correction quantity to a shipborne terminal; collecting shore feature points by using an RTK surveying and mapping device, constructing a continuous and smooth three-dimensional shore curve model by using a cubic spline interpolation algorithm, and generating a shore visual map with geographic coordinates by fusing wharf geographic information; establishing a ship coordinate system, converting inertial navigation attitude data to the coordinate system by using a rotation matrix and a translation vector, fusing the RTK position data, and outputting high-precision ship position and posture; transmitting the fused data through a 5G network, converting the data to a local plane coordinate system at a shore terminal, generating a real-time dynamic visual interface, pushing the interface to a tugboat terminal through an RTSP protocol, and assisting in pilotage operation, so that high-precision perception, visual guidance and shore-ship cooperation in a ship berthing process are realized, and the safety and intelligent level of pilotage are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent ship pilotage technology, specifically to an intelligent ship pilotage method and system based on RTK and inertial navigation fusion. Background Technology

[0002] With the continuous advancement of intelligent shipping and smart port construction, pilotage operations during the berthing phase are developing towards higher precision, automation, and collaboration. Pilotage operations, especially in complex waters or under adverse weather conditions, place extremely high demands on the relative position perception between the vessel and the shoreline, dynamic attitude monitoring, and shore-to-ship / tugboat collaborative control. Traditional pilotage methods, which mainly rely on human experience, paper charts, VHF voice communication, and ordinary Global Positioning System (GPS) for vessel positioning and command, are no longer sufficient to meet the demands of modern intelligent ships for safe, efficient, and precise berthing.

[0003] The existing pilotage system has several significant shortcomings: 1) Insufficient positioning accuracy. Traditional ship positioning generally uses standard GPS or DGPS technology, with positioning accuracy typically within the range of 5–10 meters. This is insufficient to determine the centimeter-level relative position between the ship and the dock edge, fender facilities, or adjacent ships. During berthing, even minor positional deviations can lead to collisions between the ship and the dock, posing a significant safety hazard. Although differential positioning technologies (such as RTK) have been applied in surveying, autonomous driving, and other fields, a standardized closed loop of collaborative work between shore-based reference stations and shipborne mobile stations has not yet been established in pilotage systems, resulting in the ineffective implementation of high-precision positioning capabilities. 2) Coarse shoreline geographic information modeling. Current pilotage operations rely heavily on shoreline information derived from paper charts or low-resolution remote sensing imagery, which is static and approximate data. This data fails to reflect details such as the actual concave and convex structures of the dock edge, fender distribution, temporary construction areas, or buoy locations. 3) Traditional modeling methods often employ linear interpolation or simplified geometric fitting, making it difficult to accurately reproduce the continuous curvature changes of complex shorelines. This results in the inability to accurately predict the shortest distance to the shoreline during berthing, increasing operational risks. 4) Lagging and low visibility in shore-to-ship and tugboat interaction. Existing communication mainly relies on VHF radio voice calls or 4G public network transmission. Voice communication is inefficient and prone to misunderstandings, while 4G networks have a transmission delay of 50-100ms and unstable bandwidth, making it difficult to support real-time data synchronization. Tugboat operators usually cannot directly obtain the precise coordinates, heading, or attitude information of the vessel to be tugboat, and can only make indirect judgments through voice commands, resulting in delayed collaborative response and seriously affecting berthing efficiency and safety. 5) Weak multi-source sensor fusion capability. Single RTK positioning is prone to loss of lock when satellite signals are blocked (such as by viaducts or gantry cranes), while inertial navigation systems (IMUs), although capable of high-frequency output, suffer from drift errors that accumulate over time. Most existing systems do not deeply fuse RTK and IMU data and lack effective filtering algorithms (such as Kalman filtering) for state estimation and error compensation, resulting in a significant decrease in the continuity and reliability of sensing data in dynamic and complex environments.

[0004] In summary, existing pilotage technologies have significant shortcomings in positioning accuracy, shoreline modeling, information interaction, and data fusion, making it difficult to support the closed-loop control requirements of intelligent vessels during the berthing phase, encompassing "precise perception—dynamic positioning—real-time collaboration." Therefore, there is an urgent need for an intelligent pilotage method that integrates high-precision positioning, refined modeling, low-latency communication, and multi-sensor fusion to improve the safety, efficiency, and automation level of berthing operations. Summary of the Invention

[0005] To address the problems of insufficient positioning accuracy, coarse shoreline information modeling, lagging shore-to-ship and tugboat collaborative interaction, and poor reliability of sensing data in dynamic environments currently existing in intelligent vessel pilotage (especially during berthing), this invention provides an intelligent vessel pilotage method based on RTK and inertial navigation fusion. This method can achieve centimeter-level dynamic positioning of the vessel, high-precision 3D shoreline modeling, and real-time attitude perception, significantly improving the spatial perception capability and collaborative efficiency of the pilotage process. This invention also relates to an intelligent vessel pilotage system based on RTK and inertial navigation fusion.

[0006] The technical solution of the present invention is as follows:

[0007] A smart ship pilotage method based on RTK and inertial navigation fusion, characterized by the following steps:

[0008] The differential correction calculation steps are as follows: A GNSS receiver is deployed at the highest point of the dock as an RTK base station and its position coordinates are obtained. Then, the position coordinates of the RTK base station are calibrated by a surveying instrument to obtain the calibrated position coordinates. Based on the calibrated position coordinates and the satellite observation coordinates received by the RTK base station in real time, the differential correction is calculated in real time and broadcast to the shipborne terminal through a wireless link.

[0009] The steps for constructing the curve model and generating the visualization map are as follows: Using an RTK surveying device, the three-dimensional coordinates of multiple feature points are collected along the edge of the wharf shoreline. Based on the three-dimensional coordinates of the multiple feature points, a three-dimensional shoreline curve model that satisfies the continuity of position, the continuity of the first derivative, and the continuity of the second derivative is constructed using a cubic spline interpolation algorithm. Based on the three-dimensional shoreline curve model and the pre-acquired basic geographic information of the wharf, a shoreline visualization map with geographic coordinates is generated.

[0010] The steps for fusing shipboard inertial navigation and RTK are as follows: Using the midpoint of the ship's deck as the origin and the bow direction as the reference point... Axis, starboard direction The axis, perpendicular to the deck plane and upward is A ship coordinate system is established. The differential correction is received in real-time by an RTK rover station mounted on the ship. Based on the differential correction and the original observation coordinates received in real-time by the RTK rover station, the corrected ship position coordinates in the ship coordinate system are calculated. The ship attitude angles in its own inertial navigation system are obtained using an inertial navigation device mounted on the ship. An angle measurement device is then used to measure the rotation angle of the inertial navigation system relative to the ship coordinate system, thereby generating a rotation matrix. A distance measurement device is used to measure the offset distance of the origin of the inertial navigation system relative to the origin of the ship coordinate system, thereby generating a translation vector. Based on the rotation matrix and translation vector, a coordinate transformation method is used to transform the ship attitude angles in the inertial navigation system to those in the ship coordinate system. A Kalman filter algorithm is then used to fuse the transformed ship attitude angles with the corrected ship position coordinates, outputting the fused ship pose data.

[0011] The steps for generating and displaying the real-time dynamic visualization interface are as follows: The fused ship pose data is encapsulated into JSON format using the MQTT protocol and transmitted to the shore-based processing center via a 5G network. After receiving the ship pose data, the shore-based processing center uses a coordinate transformation method to convert the ship pose data from the ship coordinate system to the local plane coordinate system corresponding to the shoreline visualization map. Based on the ship pose data in the local plane coordinate system, ship icons are dynamically drawn in the shoreline visualization map to generate a real-time dynamic visualization interface. This interface is then pushed to the display terminal on the tugboat via the RTSP protocol to assist the tugboat in adjusting its thrust and direction, thereby enabling intelligent ship pilotage.

[0012] Preferably, in the curve model construction and visualization map generation step, the use of cubic spline interpolation algorithm to construct the three-dimensional shoreline curve model specifically includes:

[0013] At any two adjacent feature points and The formed interval [ Within [the range], define a cubic polynomial as the difference function:

[0014]

[0015] The coefficients for each interval are solved by imposing the following constraints on the difference function. :

[0016] Positional continuity: ;

[0017] The first derivative is continuous: ;

[0018] The second derivative is continuous: ;

[0019] Natural boundary conditions: ;

[0020] Based on the above constraints, establish a system of linear equations, and solve for the coefficients in all intervals using the system of linear equations. This generates a continuous and smooth three-dimensional shoreline curve model.

[0021] Preferably, in the shipborne inertial navigation and RTK fusion step, the data fusion using the Kalman filter algorithm specifically includes:

[0022] Constructing state vectors ,in,( ( ) represents the true position coordinates of the ship to be estimated. () represents the actual attitude angle of the ship to be estimated;

[0023] Construct observation vectors ,in,( ) represents the corrected ship position coordinates obtained from the RTK rover station calculation. ( ) represents the ship's attitude angles obtained by the inertial navigation system;

[0024] Using the Kalman filter algorithm, the observation vector For the state vector Recursive filtering calculations are performed to output the optimal estimated fused ship pose data.

[0025] Preferably, in the shipborne inertial navigation and RTK fusion step, the pre-acquired installation deviation angle of the inertial navigation equipment is used to correct the deviation of the heading angle in the fused ship attitude data, so as to obtain the corrected true heading angle.

[0026] Preferably, in the step of generating and displaying the real-time dynamic visualization interface, dynamically drawing ship icons in the shoreline visualization map specifically includes:

[0027] The ship is represented by an arrow-shaped graphic symbol, and the direction of the arrow-shaped graphic symbol is determined according to the corrected true heading angle in the ship's attitude data. Then, the arrow-shaped graphic symbol is rendered with different colors according to the roll angle in the ship's attitude data. Specifically, green is used when the roll angle is less than 3 degrees, yellow is used when the roll angle is between 3 and 5 degrees, and red is used when the roll angle is greater than 5 degrees.

[0028] An intelligent ship pilotage system based on RTK and inertial navigation fusion is characterized by comprising, in sequence, a differential correction calculation module, a curve model construction and visualization map generation module, a shipborne inertial navigation and RTK fusion module, and a real-time dynamic visualization interface generation and display module.

[0029] The differential correction calculation module deploys a GNSS receiver as an RTK base station at the highest point of the dock and obtains its position coordinates. Then, it calibrates the position coordinates of the RTK base station using a surveying instrument to obtain calibrated position coordinates. Based on the calibrated position coordinates and the satellite observation coordinates received by the RTK base station in real time, it calculates the differential correction in real time and broadcasts it to the shipborne terminal via a wireless link.

[0030] The curve model construction and visualization map generation module uses RTK surveying equipment to collect the three-dimensional coordinates of multiple feature points along the edge of the wharf shoreline. Based on the three-dimensional coordinates of the multiple feature points, a cubic spline interpolation algorithm is used to construct a three-dimensional shoreline curve model that satisfies positional continuity, first derivative continuity, and second derivative continuity. Based on the three-dimensional shoreline curve model and the pre-acquired basic geographic information of the wharf, a shoreline visualization map with geographic coordinates is generated.

[0031] The shipborne inertial navigation and RTK fusion module uses the midpoint of the ship's deck as the origin and the bow direction as the reference point. Axis, starboard direction The axis, perpendicular to the deck plane and upward is A ship coordinate system is established. The differential correction is received in real-time by an RTK rover station mounted on the ship. Based on the differential correction and the original observation coordinates received in real-time by the RTK rover station, the corrected ship position coordinates in the ship coordinate system are calculated. The ship attitude angles in its own inertial navigation system are obtained using an inertial navigation device mounted on the ship. An angle measurement device is then used to measure the rotation angle of the inertial navigation system relative to the ship coordinate system, thereby generating a rotation matrix. A distance measurement device is used to measure the offset distance of the origin of the inertial navigation system relative to the origin of the ship coordinate system, thereby generating a translation vector. Based on the rotation matrix and translation vector, a coordinate transformation method is used to transform the ship attitude angles in the inertial navigation system to those in the ship coordinate system. A Kalman filter algorithm is then used to fuse the transformed ship attitude angles with the corrected ship position coordinates, outputting the fused ship pose data.

[0032] The real-time dynamic visualization interface generation and display module uses the MQTT protocol to encapsulate the fused ship posture data into JSON format and transmits it to the shore-based processing center via a 5G network. After receiving the ship posture data, the shore-based processing center uses a coordinate transformation method to convert the ship posture data from the ship coordinate system to the local plane coordinate system corresponding to the shoreline visualization map. Based on the ship posture data in the local plane coordinate system, the system dynamically draws ship icons in the shoreline visualization map, generates a real-time dynamic visualization interface, and pushes the real-time dynamic visualization interface to the display terminal on the tugboat via the RTSP protocol for display, so as to assist the tugboat in adjusting thrust and direction and realize intelligent ship pilotage.

[0033] Preferably, in the curve model construction and visualization map generation module, the construction of the three-dimensional shoreline curve model using the cubic spline interpolation algorithm specifically includes:

[0034] At any two adjacent feature points and The formed interval [ Within [the range], define a cubic polynomial as the difference function:

[0035]

[0036] in, The coefficients for each interval are solved by imposing the following constraints on the difference function. :

[0037] Positional continuity: ;

[0038] The first derivative is continuous:

[0039] The second derivative is continuous:

[0040] Natural boundary conditions:

[0041] Based on the above constraints, establish a system of linear equations, and solve for the coefficients in all intervals using the system of linear equations. This generates a continuous and smooth three-dimensional shoreline curve model.

[0042] Preferably, in the shipborne inertial navigation and RTK fusion module, the data fusion using the Kalman filter algorithm specifically includes:

[0043] Constructing state vectors ,in,( ( ) represents the true position coordinates of the ship to be estimated. () represents the actual attitude angle of the ship to be estimated;

[0044] Construct observation vectors ,in,( ) represents the corrected ship position coordinates obtained from the RTK rover station calculation. ( ) represents the ship's attitude angles obtained by the inertial navigation system;

[0045] Using the Kalman filter algorithm, the observation vector For the state vector Recursive filtering calculations are performed to output the optimal estimated fused ship pose data.

[0046] Preferably, in the shipborne inertial navigation and RTK fusion module, the pre-acquired installation deviation angle of the inertial navigation equipment is used to correct the deviation of the heading angle in the fused ship attitude data, so as to obtain the corrected true heading angle.

[0047] Preferably, in the real-time dynamic visualization interface generation and display module, dynamically drawing ship icons in the shoreline visualization map specifically includes:

[0048] The ship is represented by an arrow-shaped graphic symbol, and the direction of the arrow-shaped graphic symbol is determined according to the corrected true heading angle in the ship's attitude data. Then, the arrow-shaped graphic symbol is rendered with different colors according to the roll angle in the ship's attitude data. Specifically, green is used when the roll angle is less than 3 degrees, yellow is used when the roll angle is between 3 and 5 degrees, and red is used when the roll angle is greater than 5 degrees.

[0049] The technical effects of this invention are as follows:

[0050] This invention provides an intelligent ship pilotage method based on RTK and inertial navigation fusion. First, a GNSS receiver is deployed at a high point on the dock as an RTK base station to obtain its position coordinates. This effectively avoids obstruction of satellite signals and multipath interference from surrounding buildings or facilities, significantly improving the quality and stability of satellite observation data received by the base station. Then, the position coordinates of the RTK base station are calibrated using surveying instruments to obtain calibrated position coordinates. This eliminates systematic positioning errors caused by installation deviations, geodetic height transformation errors, or coordinate system inconsistencies, ensuring that the base station coordinates achieve centimeter-level or even millimeter-level accuracy, providing a highly reliable benchmark for subsequent differential correction. Based on this, the differential correction amount is calculated in real time and broadcast to the shipborne terminal via a wireless link. This allows the subsequent RTK rover station to effectively eliminate common error sources such as ionospheric delay, tropospheric delay, and satellite orbit errors during the calculation process, thereby achieving centimeter-level high-precision dynamic positioning and providing a reliable positioning basis for ship pilotage operations. Then, RTK mapping equipment was used to collect the 3D coordinates of multiple feature points, and a cubic spline interpolation algorithm was used to construct a 3D shoreline curve model that satisfies positional continuity, first derivative continuity, and second derivative continuity. This ensures that the generated shoreline curve model is geometrically smooth and without abrupt changes, realistically restoring the continuous curvature changes of the wharf edge and effectively preventing the "sawtooth effect" or navigation misjudgment caused by broken line connections. Based on the 3D shoreline curve model and the pre-acquired basic geographic information of the wharf, a high-precision 3D shoreline visualization map with geographic coordinates was generated. This effectively solved the problem of rough shoreline modeling, improved the geometric realism of the map, and enhanced its information richness and readability, providing high-fidelity base map support for subsequent ship dynamic positioning and visualization guidance. A ship coordinate system is then established, and differential correction values ​​are received in real time via an RTK rover station mounted on the ship. The corrected ship position coordinates in the ship coordinate system are then calculated, achieving centimeter-level high-precision positioning. Simultaneously, the inertial navigation system continuously outputs the ship's attitude angles (heading, pitch, roll) in its own inertial navigation coordinate system, exhibiting good short-term stability and high-frequency response capabilities. Furthermore, the rotation angle of the inertial navigation coordinate system relative to the ship coordinate system is measured using an angle measurement device (such as an electronic compass), generating a precise rotation matrix. Finally, the distance between the origins of the two coordinate systems is measured using a distance measurement device (such as a laser rangefinder). The offset distance is used to generate a translation vector, thereby achieving a precise transformation of inertial navigation data from its body coordinate system to the ship coordinate system, eliminating systematic errors caused by sensor installation deviations. Based on this, a Kalman filter algorithm is used to fuse the transformed attitude angles with the RTK corrected position, giving full play to the high precision advantage of RTK and the continuity advantage of inertial navigation, effectively suppressing the performance degradation of a single sensor under signal loss or dynamic disturbances, solving the problem of low reliability of single devices, and outputting high-precision, high-frequency, and highly robust fused ship attitude data, significantly improving the reliability of state perception during pilotage.Finally, the MQTT protocol was used to encapsulate the fused ship pose data into JSON format and transmit it to the shore-based processing center via a 5G network. Leveraging MQTT's lightweight, low-latency, and reconnection-supporting characteristics, reliable transmission of ship status data was ensured in complex wireless environments. Simultaneously, the high bandwidth and low latency of the 5G network ensured real-time data transmission, with end-to-end latency below 20ms. Upon receiving the data, the shore-based processing center transformed the ship pose data from the ship's coordinate system to a local planar coordinate system corresponding to the shoreline visualization map (e.g., a northeast-sky coordinate system with a fixed point at the dock as the origin), achieving spatial alignment between the ship's position and the shoreline map, facilitating subsequent processing. Visualization: Based on the transformed local coordinates, ship icons (such as rectangles with the bow pointing or real outlines) are dynamically drawn on the shoreline visualization map, reflecting the ship's position, heading, and attitude in real time, generating an intuitive real-time dynamic visualization interface; then, the interface is pushed to the display terminal in the tugboat's bridge via the RTSP protocol in the form of a video stream, enabling the tugboat operator to simultaneously grasp the relative position relationship between the ship and the shoreline, the berthing angle, and the safe distance, and adjust the thrust magnitude and direction in a timely manner, significantly improving the response speed and operational accuracy of tugboat collaborative operations, realizing low-latency shore-ship data interaction and tugboat collaboration, solving the problem of interaction lag, and ultimately achieving safe, efficient, and intelligent ship pilotage guidance.

[0051] This invention constructs a high-precision RTK differential reference, generates a continuous and smooth 3D shoreline visualization map, achieves coordinate alignment and data fusion between shipborne inertial navigation and RTK, and establishes a low-latency shore-ship collaborative visualization channel, forming a complete intelligent ship pilotage system. Compared with traditional pilotage methods that rely on manual experience or single GNSS positioning, this invention has the following significant comprehensive technical effects: 1) Centimeter-level positioning accuracy: Through the fusion of RTK differential correction and Kalman filtering of shipborne inertial navigation data, high-precision estimation of ship attitude is achieved, with position accuracy better than ±5cm and bow angle accuracy better than 0.1°, which is more than 100 times higher than traditional GPS positioning. This completely solves the risk of collision caused by positioning deviation during berthing and has industry-leading centimeter-level perception capabilities. 2) High-precision shoreline modeling: A 3D shoreline curve model is constructed using over 100 shoreline feature points collected in the field and combined with a cubic spline interpolation algorithm. The modeling error is less than 0.5m, accurately reproducing the wharf edge contour, fender position, and curvature changes. This effectively fills the gap in local detail accuracy of traditional nautical charts, providing a high-fidelity geographic base map for visualization guidance. 3) Low-latency collaborative interaction: Through 5G private network transmission, the fused pose data and real-time visualization interface achieve end-to-end communication latency of less than 20ms, a reduction of over 60% compared to 4G networks. This ensures information synchronization between the shore-based processing center and the tugboat terminal, enabling tugboat operators to monitor vessel dynamics in real time, significantly improving collaborative response speed, and reducing the measured berthing accident rate by 30%. 4) High-reliability data: Through multi-source data fusion of inertial navigation and RTK, the system maintains continuous data output even under complex conditions such as satellite signal obstruction (e.g., wharf viaducts, bridge areas) or severe ship vibration. Compared to solutions relying solely on RTK, data availability and system reliability are improved by over 80%. In summary, this invention not only achieves comprehensive breakthroughs in positioning accuracy, modeling fidelity, real-time transmission, and system robustness, but also constructs a closed loop of "perception-modeling-fusion-interaction" through multi-technology collaboration, forming an intelligent ship pilotage guidance system with high uniqueness and engineering practical value, significantly improving the safety, efficiency, and intelligence level of berthing operations.

[0052] Furthermore, the pre-acquired inertial navigation equipment installation deviation angle is used to compensate for and correct the heading angle in the fused ship attitude data, resulting in a corrected true heading angle. This effectively eliminates systematic heading deviations caused by improper installation of the inertial navigation equipment (such as an angle with the ship's bow and stern lines), avoiding persistent angle misjudgments in subsequent navigation and berthing guidance. By calibrating and storing this installation deviation angle during system initialization, the heading output can be compensated in real time during data processing, ensuring that the corrected heading angle accurately reflects the ship's true heading. This significantly enhances the accuracy of ship attitude perception during berthing, especially in direction-sensitive scenarios such as low-speed approach to the dock or tugboat collaborative operations, greatly reducing the risk of collisions caused by heading errors and improving the reliability and safety of the pilot guidance system.

[0053] This invention also relates to an intelligent ship pilotage system based on RTK and inertial navigation fusion. This system corresponds to the aforementioned intelligent ship pilotage method based on RTK and inertial navigation fusion. It can be understood as a system that implements the aforementioned intelligent ship pilotage method based on RTK and inertial navigation fusion. It includes a differential correction calculation module, a curve model construction and visualization map generation module, a shipborne inertial navigation and RTK fusion module, and a real-time dynamic visualization interface generation and display module connected in sequence. Each module works in concert to construct a high-precision RTK differential reference, generate a continuous and smooth three-dimensional shoreline visualization map, realize coordinate alignment and data fusion of shipborne inertial navigation and RTK, and establish a low-latency shore-ship collaborative visualization channel, thus forming a complete intelligent ship pilotage guidance system. This solution achieves centimeter-level high-precision, high-frequency, and high-continuity perception of ship position and attitude information, effectively overcoming problems such as signal blockage, multipath interference, and inertial navigation drift in complex port environments. Through 3D shoreline modeling and geographic information fusion, it provides a high-fidelity, quantifiable visual guidance base map, significantly improving spatial awareness during berthing. Based on a real-time data transmission and visualization push mechanism using MQTT+5G+RTSP, it constructs an efficient collaborative operation closed loop between the shore, ship, and tugboat, enabling tugboat operators to monitor ship dynamics in real time and precisely adjust thrust direction and magnitude. This solution deeply integrates high-precision positioning, multi-source fusion, 3D modeling, and real-time visualization technologies, achieving intelligent, visualized, and collaborative guidance during ship berthing. It not only significantly improves the safety and efficiency of pilotage operations but also reduces reliance on pilot experience, demonstrating significant engineering application value and promising prospects for wider application. Attached Figure Description

[0054] Figure 1 This is a flowchart of the intelligent ship pilotage method based on the fusion of RTK and inertial navigation of the present invention. Detailed Implementation

[0055] The present invention will now be described with reference to the accompanying drawings.

[0056] This invention relates to an intelligent ship pilotage method based on RTK and inertial navigation fusion, the flowchart of which is shown below. Figure 1 As shown, the steps are as follows:

[0057] I. Differential Correction Calculation Steps: Deploy a GNSS receiver as an RTK base station at the highest point of the dock and obtain its position coordinates. Then, calibrate the position coordinates of the RTK base station using surveying instruments to obtain calibrated position coordinates. Based on the calibrated position coordinates and the real-time satellite observation coordinates received by the RTK base station (the RTK base station receives GPS satellite signals and obtains its own satellite observation coordinates), calculate the differential correction in real time and broadcast it to the shipborne terminal via a wireless link.

[0058] This step provides a centimeter-level differential positioning reference for the deployment and differential correction of the shore-based RTK base station, addressing the issue of insufficient accuracy in ordinary GPS. Specifically, a high-precision GNSS receiver (supporting BeiDou, GPS, etc.) is first deployed as an RTK base station at an unobstructed high point (such as the top of a lighthouse) in the shore-side dock area. The power supply, lightning rod, and wireless data transmission module (such as a UHF radio or 4G / 5G network module) are connected, and the sampling frequency is set to 1. The location coordinates of the RTK base station are obtained through multi-frequency, multi-system (GPS / BDS / Galileo) observations. Since the location coordinates may contain installation errors or geodetic height conversion deviations, the location coordinates of the RTK base station are calibrated using surveying instruments (such as a total station) to obtain the calibrated, accurate location coordinates. Then, based on the calibrated position coordinates and the real-time satellite observation coordinates received by the RTK base station, Real-time calculation of differential correction amount Calculate according to the following formula:

[0059]

[0060]

[0061]

[0062] Finally, the difference correction amount is... by The frequency is broadcast to the shipborne terminal (shipborne RTK mobile station) via a wireless link (such as UHF) to achieve coordinate correction.

[0063] II. Curve Model Construction and Visual Map Generation Steps: Use RTK mapping equipment to collect the three-dimensional coordinates of multiple feature points along the edge of the wharf shoreline. Based on the three-dimensional coordinates of the multiple feature points, use a cubic spline interpolation algorithm to construct a three-dimensional shoreline curve model that satisfies positional continuity, first derivative continuity, and second derivative continuity. Based on the three-dimensional shoreline curve model and the pre-acquired basic geographic information of the wharf, generate a shoreline visualization map with geographic coordinates.

[0064] This step involves 100-point shoreline surveying and cubic spline interpolation modeling to construct a high-precision 3D shoreline visualization map, addressing the issue of coarse shoreline modeling. Specifically, firstly, surveyors use handheld RTK surveying equipment to measure the 3D coordinates of one feature point every 5-10 meters along the edge of the wharf (from the starting point to the end point), ensuring coverage of key locations such as wharf curves, fenders, and corners, recording the 3D coordinates of a total of 100 feature points. ,in, =1,2,...,100 , For planar coordinates, The elevation information is obtained, and the 3D coordinates of 100 measured feature points are stored as a CSV file. Then, the 3D coordinates of the 100 feature points are imported into MATLAB or Python, and the coefficients are solved using a cubic spline interpolation algorithm. Thus, a continuous three-dimensional shoreline curve model satisfying positional continuity, first derivative continuity, and second derivative continuity is constructed. The specific process is as follows:

[0065] 1) Definition of interpolation function: Interpolation function is defined for any two adjacent feature points. and The formed interval [ Within [the function], define a cubic polynomial as the difference function. The expression is as follows:

[0066]

[0067] in, , , , These are the coefficients to be solved for on the interval; (The positions are consecutive).

[0068] 2) Continuity condition setting: requires the entire interpolation function to be continuous. The following continuity constraints must be met:

[0069] Positional continuity: The interpolation function for each interval must pass through the measurement point at the node, i.e., satisfy:

[0070] (Continuous position)

[0071] First derivative continuity: The first derivative (slope) of the function changes smoothly at the nodes, i.e., it satisfies:

[0072] ;

[0073] Second derivative continuity: The second derivative (curvature) of the function changes smoothly at the nodes, i.e., it satisfies:

[0074] ;

[0075] 3) Boundary condition setting: To ensure a unique solution to the system of equations, boundary conditions must be applied. Here, natural boundary conditions are used, which require that the second derivative at the start and end points of the entire curve be zero.

[0076]

[0077] Natural boundary conditions ensure that the curvature of the curve is gentle at the endpoints, avoiding abrupt changes in curvature at the endpoints.

[0078] 4) Solving the system of equations and generating the model: Based on the above constraints, establish a system of linear equations, and solve for the coefficients in all intervals using the system of linear equations. This generates a continuous and smooth three-dimensional shoreline curve model that closely matches the real shoreline.

[0079] Finally, the generated 3D shoreline curve model is fused with the pre-acquired basic geographic information of the wharf (such as berth number, fender location, etc.), and a shoreline visualization map with geographic coordinates is generated using an OpenGL or ArcGIS platform (supporting zooming, panning, and displaying coordinate scale).

[0080] III. Steps for fusing shipborne inertial navigation and RTK: Using the midpoint of the ship's deck as the origin and the bow direction as... Axis, starboard direction The axis, perpendicular to the deck plane and upward is A ship coordinate system is established. The differential correction is received in real-time via an RTK rover station mounted on the ship. Based on the differential correction and the original observation coordinates received in real-time by the RTK rover station, the corrected ship position coordinates in the ship coordinate system are calculated. The ship attitude angles in its own inertial navigation system are obtained via an inertial navigation device mounted on the ship. An angle measurement device is used to measure the rotation angle of the inertial navigation system relative to the ship coordinate system, thereby generating a rotation matrix. A distance measurement device is used to measure the offset distance of the origin of the inertial navigation system relative to the origin of the ship coordinate system, thereby generating a translation vector. Based on the rotation matrix and translation vector, a coordinate transformation method is used to transform the ship attitude angles in the inertial navigation system to those in the ship coordinate system. A Kalman filter algorithm is then used to fuse the transformed ship attitude angles with the corrected ship position coordinates, outputting the fused ship attitude data.

[0081] This step involves shipborne inertial navigation and RTK fusion sensing, which can acquire high-precision coordinates and heading of the ship in real time, solving the problem of low reliability of single equipment. Specifically, it first uses the midpoint of the ship's deck as the origin and the heading as the reference point. Axis, starboard direction The axis, perpendicular to the deck plane and upward is The ship's coordinate system is established by the axis. A pilot carries a portable inertial navigation system (IMU, 100Hz sampling rate, <5kg weight, 8-hour endurance), an RTK rover (supporting differential signal reception from an RTK base station), and a 220V power bank onto the ship. After boarding, the inertial navigation antenna and RTK rover are fixed 1m ahead of the ship's centerline using a specially designed "installation mold (with antenna installation positions marked)," ensuring the antenna coincides with the origin of the ship's coordinate system. Then, the RTK rover receives differential correction data from the shore-based RTK base station in real time. Based on this differential correction and the rover's own original observation coordinates (the rover receives satellite signals from the same GPS system as the RTK base station to obtain its own original observation coordinates), the corrected ship position coordinates in the ship's coordinate system are calculated. The ship's attitude angles in its own inertial navigation coordinate system are obtained through the installed inertial navigation equipment. ), that is, roll angle Pitch angle Heading angle .

[0082] And use angle measuring devices (such as electronic compasses or attitude sensors) to measure the rotation angle (around) of the inertial navigation coordinate system relative to the ship's coordinate system. (axis), and then generate a rotation matrix. ; and use distance measuring equipment (such as a laser rangefinder) to measure the offset distance of the inertial navigation coordinate system origin relative to the ship coordinate system origin, and then generate a translation vector. Based on rotation matrix Translation vector The ship's attitude angles in the inertial navigation coordinate system are transformed to the ship's coordinate system using a coordinate transformation method, as shown in the following formula:

[0083]

[0084] in, The coordinates are in the inertial navigation system of the inertial navigation device itself. These are the coordinates in the transformed ship coordinate system.

[0085] The Kalman filter algorithm is then used to fuse the transformed ship attitude angles with the corrected ship position coordinates, outputting the fused ship attitude data, which specifically includes:

[0086] First construct the state vector ,in,( ( ) represents the true position coordinates of the ship to be estimated. () represents the actual attitude angle of the ship to be estimated; It is the transpose matrix;

[0087] Reconstruct the observation vector ,in,( ) represents the corrected ship position coordinates obtained from the RTK rover station calculation. ( ) represents the ship's attitude angles obtained by the inertial navigation system;

[0088] Finally, the Kalman filter algorithm is used to apply the observed vector. For the state vector Recursive filtering calculations are performed, and the optimal estimated fused ship attitude data is output at a frequency of 20Hz.

[0089] After obtaining the fused ship attitude data, slight deviations in the IMU installation angle are considered. Therefore, it is necessary to analyze the heading angle in the fused ship pose data. Real-time compensation and correction are performed, that is, the installation deviation angle of the inertial navigation equipment is obtained in advance through static calibration. (Pre-calibration, The heading angle is compensated in real time, and the corrected true heading angle is output to eliminate systematic errors. (i.e., the actual bow direction of the ship) is shown in the following formula:

[0090]

[0091] IV. Real-time Dynamic Visualization Interface Generation and Display Steps: The fused ship pose data is encapsulated into JSON format using the MQTT protocol and transmitted to the shore-based processing center via a 5G network. After receiving the ship pose data, the shore-based processing center uses a coordinate transformation method to convert the ship pose data from the ship coordinate system to the local plane coordinate system corresponding to the shoreline visualization map. Based on the ship pose data in the local plane coordinate system, ship icons are dynamically drawn in the shoreline visualization map to generate a real-time dynamic visualization interface. This interface is then pushed to the display terminal on the tugboat via the RTSP protocol to assist the tugboat in adjusting its thrust and direction, thus enabling intelligent ship pilotage.

[0092] This step can be understood as 5G private network transmission and tugboat visualization, enabling low-latency shore-to-ship data interaction and tugboat collaboration, solving the problem of interaction lag. Specifically, firstly, a 5G industrial router is connected to the dock's 5G private network (previously tested with the operator to ensure coverage of the dock and a surrounding 5km range). The MQTT protocol (supporting reconnection after network interruption) is used to transmit the fused ship pose data. The data is encapsulated in JSON format and transmitted to the shore-based processing center (i.e., the shore-based workstation IP address) via a dedicated 5G industrial IoT network (100Mbps bandwidth, <20ms latency) at a frequency of 20Hz. The shore-based processing center (configured with an i7 processor and 16GB of memory) runs monitoring software. After receiving the ship's pose data, it uses a coordinate transformation method (coordinate transformation formula) to transform the ship's pose data from the ship's coordinate system to the local planar coordinate system corresponding to the shoreline visualization map. , ), as shown in the following formula:

[0093]

[0094]

[0095] in, The coordinates of the origin of the local planar coordinate system corresponding to the shoreline visualization map.

[0096] Then, based on the ship pose data in the local plane coordinate system, ship icons are dynamically drawn in the shoreline visualization map, generating a real-time dynamic visualization interface. That is, the ship is represented by an arrow-shaped graphic symbol, and the corrected true heading angle is based on the ship pose data. Determine the direction of the arrow-shaped graphic symbol (i.e., use an arrow to indicate the actual bow direction of the ship). Then, based on the roll angle in the ship's attitude data... The size is adjusted, and the arrow-shaped graphic symbol is rendered in different colors (i.e., the roll angle is marked with different colors). (); where green is used when the roll angle is less than 3 degrees, yellow is used when the roll angle is between 3 and 5 degrees, and red is used when the roll angle is greater than 5 degrees, i.e., green < 3°, yellow 3-5°, and red > 5°.

[0097] The shore-based workstation then pushes the generated real-time dynamic visualization interface (1920×1080 resolution) to the tugboat's bridge display terminal (15-inch touchscreen, 20Hz refresh rate) via a 5G private network using the RTSP streaming protocol. The tugboat operator can view the vessel's position, attitude, and relative distance to the shoreline in real time (the interface displays the "Shortest Distance Between Vessel and Shoreline" value), thus assisting the tugboat in adjusting thrust and direction. After the vessel berths, the pilot shuts down the onboard equipment, and the shore-based workstation automatically stops receiving data, transmitting the vessel's trajectory (during the pilotage process) to the shore. Time-series data, shoreline models, and tugboat interaction logs are stored as database files for subsequent review and optimization, thus completing the intelligent vessel pilotage.

[0098] This invention also relates to an intelligent ship pilotage system based on RTK and inertial navigation fusion. This system corresponds to the aforementioned intelligent ship pilotage method based on RTK and inertial navigation fusion, and can be understood as a system implementing the aforementioned method. The system includes, in sequence, a differential correction calculation module, a curve model construction and visualization map generation module, a shipborne inertial navigation and RTK fusion module, and a real-time dynamic visualization interface generation and display module. Specifically,

[0099] The differential correction calculation module deploys a GNSS receiver as an RTK base station at the highest point of the dock and obtains its position coordinates. Then, it calibrates the position coordinates of the RTK base station using a surveying instrument to obtain calibrated position coordinates. Based on the calibrated position coordinates and the satellite observation coordinates received by the RTK base station in real time, it calculates the differential correction in real time and broadcasts it to the shipborne terminal via a wireless link.

[0100] The curve model construction and visualization map generation module uses RTK surveying equipment to collect the three-dimensional coordinates of multiple feature points along the edge of the wharf shoreline. Based on the three-dimensional coordinates of the multiple feature points, a cubic spline interpolation algorithm is used to construct a three-dimensional shoreline curve model that satisfies positional continuity, first derivative continuity, and second derivative continuity. Based on the three-dimensional shoreline curve model and the pre-acquired basic geographic information of the wharf, a shoreline visualization map with geographic coordinates is generated.

[0101] The shipborne inertial navigation and RTK fusion module uses the midpoint of the ship's deck as the origin and the bow direction as the reference point. Axis, starboard direction The axis, perpendicular to the deck plane and upward is A ship coordinate system is established. The differential correction is received in real-time by an RTK rover station mounted on the ship. Based on the differential correction and the original observation coordinates received in real-time by the RTK rover station, the corrected ship position coordinates in the ship coordinate system are calculated. The ship attitude angles in its own inertial navigation system are obtained using an inertial navigation device mounted on the ship. An angle measurement device is then used to measure the rotation angle of the inertial navigation system relative to the ship coordinate system, thereby generating a rotation matrix. A distance measurement device is used to measure the offset distance of the origin of the inertial navigation system relative to the origin of the ship coordinate system, thereby generating a translation vector. Based on the rotation matrix and translation vector, a coordinate transformation method is used to transform the ship attitude angles in the inertial navigation system to those in the ship coordinate system. A Kalman filter algorithm is then used to fuse the transformed ship attitude angles with the corrected ship position coordinates, outputting the fused ship pose data.

[0102] The real-time dynamic visualization interface generation and display module uses the MQTT protocol to encapsulate the fused ship posture data into JSON format and transmits it to the shore-based processing center via a 5G network. After receiving the ship posture data, the shore-based processing center uses a coordinate transformation method to convert the ship posture data from the ship coordinate system to the local plane coordinate system corresponding to the shoreline visualization map. Based on the ship posture data in the local plane coordinate system, the system dynamically draws ship icons in the shoreline visualization map, generates a real-time dynamic visualization interface, and pushes the real-time dynamic visualization interface to the display terminal on the tugboat via the RTSP protocol for display, so as to assist the tugboat in adjusting thrust and direction and realize intelligent ship pilotage.

[0103] Preferably, in the curve model construction and visualization map generation module, the construction of a three-dimensional shoreline curve model using a cubic spline interpolation algorithm specifically includes:

[0104] At any two adjacent feature points and The formed interval [ Within [the range], define a cubic polynomial as the difference function:

[0105]

[0106] in, The coefficients for each interval are solved by imposing the following constraints on the difference function. :

[0107] Positional continuity: ;

[0108] The first derivative is continuous:

[0109] The second derivative is continuous:

[0110] Natural boundary conditions:

[0111] Based on the above constraints, establish a system of linear equations, and solve for the coefficients in all intervals using the system of linear equations. This generates a continuous and smooth three-dimensional shoreline curve model.

[0112] Preferably, in the shipborne inertial navigation and RTK fusion module, the data fusion using the Kalman filter algorithm specifically includes:

[0113] Constructing state vectors ,in,( ( ) represents the true position coordinates of the ship to be estimated. () represents the actual attitude angle of the ship to be estimated;

[0114] Construct observation vectors ,in,( ) represents the corrected ship position coordinates obtained from the RTK rover station calculation. ( ) represents the ship's attitude angles obtained by the inertial navigation system;

[0115] Using the Kalman filter algorithm, the observation vector For the state vector Recursive filtering calculations are performed to output the optimal estimated fused ship pose data.

[0116] Preferably, in the shipborne inertial navigation and RTK fusion module, the pre-acquired installation deviation angle of the inertial navigation equipment is used to correct the deviation of the heading angle in the fused ship attitude data, so as to obtain the corrected true heading angle.

[0117] Preferably, in the real-time dynamic visualization interface generation and display module, dynamically drawing ship icons in the shoreline visualization map specifically includes:

[0118] The ship is represented by an arrow-shaped graphic symbol, and the direction of the arrow-shaped graphic symbol is determined according to the corrected true heading angle in the ship's attitude data. Then, the arrow-shaped graphic symbol is rendered with different colors according to the roll angle in the ship's attitude data. Specifically, green is used when the roll angle is less than 3 degrees, yellow is used when the roll angle is between 3 and 5 degrees, and red is used when the roll angle is greater than 5 degrees.

[0119] This invention provides an objective and scientific intelligent ship pilotage method and system based on RTK and inertial navigation fusion. By constructing a high-precision RTK differential reference, generating a continuous and smooth 3D shoreline visualization map, achieving coordinate alignment and data fusion between shipborne inertial navigation and RTK, and establishing a low-latency shore-ship collaborative visualization channel, a complete intelligent ship pilotage guidance system is formed. Through the deep integration of high-precision positioning, multi-source fusion, 3D modeling, and real-time visualization technologies, intelligent, visualized, and collaborative guidance of the ship berthing process is achieved. This not only significantly improves the safety and efficiency of pilotage operations but also reduces reliance on pilot experience, demonstrating significant engineering application value and promising prospects for widespread adoption.

[0120] It should be noted that the specific embodiments described above enable those skilled in the art to more fully understand the present invention, but do not limit the present invention in any way. Therefore, although the present invention has been described in detail with reference to the accompanying drawings and embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the present invention. In short, all technical solutions and improvements that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the present invention patent.

Claims

1. An intelligent ship piloting method based on RTK and inertial navigation fusion, characterized in that, Includes the following steps: Differential correction calculation steps: Deploy a GNSS receiver as an RTK base station at the commanding height of the dock and obtain its position coordinates. Then, calibrate the position coordinates of the RTK base station using surveying instruments to obtain the calibrated position coordinates. Based on the calibrated position coordinates and the satellite observation coordinates received in real time by the RTK base station, the differential correction amount is calculated in real time and broadcast to the shipborne terminal through a wireless link. Curve model construction and visualization map generation steps: Use RTK surveying equipment to collect the three-dimensional coordinates of multiple feature points along the edge of the wharf shoreline. Based on the three-dimensional coordinates of the multiple feature points, use cubic spline interpolation algorithm to construct a three-dimensional shoreline curve model that satisfies positional continuity, first derivative continuity and second derivative continuity. Based on the three-dimensional shoreline curve model and the pre-acquired basic geographic information of the wharf, a shoreline visualization map with geographic coordinates is generated. The steps for fusing shipboard inertial navigation and RTK are as follows: Using the midpoint of the ship's deck as the origin and the bow direction as the reference point... Axis, starboard direction The axis, perpendicular to the deck plane and upward is A ship coordinate system is established; the differential correction amount is received in real time by an RTK mobile station set up on the ship; the corrected ship position coordinates in the ship coordinate system are calculated based on the differential correction amount and the original observation coordinates received in real time by the RTK mobile station; the ship attitude angle in its own inertial navigation coordinate system is obtained by an inertial navigation device set up on the ship; and the rotation angle of the inertial navigation coordinate system relative to the ship coordinate system is measured by an angle measuring device, thereby generating a rotation matrix. The offset distance between the origin of the inertial navigation coordinate system and the origin of the ship coordinate system is measured using a distance measurement device, and then a translation vector is generated. Based on the rotation matrix and translation vector, the ship attitude angle in the inertial navigation coordinate system is transformed to the ship coordinate system using a coordinate transformation method. Then, the Kalman filter algorithm is used to fuse the transformed ship attitude angle with the corrected ship position coordinates to output the fused ship pose data. The steps for generating and displaying the real-time dynamic visualization interface are as follows: The fused ship pose data is encapsulated into JSON format using the MQTT protocol and transmitted to the shore-based processing center via a 5G network. After receiving the ship pose data, the shore-based processing center uses a coordinate transformation method to convert the ship pose data from the ship coordinate system to the local plane coordinate system corresponding to the shoreline visualization map. Based on the ship pose data in the local plane coordinate system, ship icons are dynamically drawn in the shoreline visualization map to generate a real-time dynamic visualization interface. This interface is then pushed to the display terminal on the tugboat via the RTSP protocol to assist the tugboat in adjusting its thrust and direction, thereby enabling intelligent ship pilotage.

2. The intelligent ship pilotage method based on RTK and inertial navigation fusion according to claim 1, characterized in that, In the steps of constructing the curve model and generating the visualization map, the use of cubic spline interpolation algorithm to construct the three-dimensional shoreline curve model specifically includes: At any two adjacent feature points and The formed interval [ Within [the range], define a cubic polynomial as the difference function: ； The coefficients for each interval are solved by imposing the following constraints on the difference function. : Positional continuity: ; The first derivative is continuous: ; The second derivative is continuous: ; Natural boundary conditions: ; Based on the above constraints, establish a system of linear equations, and solve for the coefficients in all intervals using the system of linear equations. This generates a continuous and smooth three-dimensional shoreline curve model.

3. The intelligent ship pilotage method based on RTK and inertial navigation fusion according to claim 1, characterized in that, The data fusion step of the shipborne inertial navigation and RTK is specifically performed using the Kalman filter algorithm, including: Constructing state vectors ,in,( ( ) represents the true position coordinates of the ship to be estimated. () represents the actual attitude angle of the ship to be estimated; Construct observation vectors ,in,( ) represents the corrected ship position coordinates obtained from the RTK rover station calculation. ( ) represents the ship's attitude angles obtained by the inertial navigation system; Using the Kalman filter algorithm, the observation vector For the state vector Recursive filtering calculations are performed to output the optimal estimated fused ship pose data.

4. The intelligent ship pilotage method based on RTK and inertial navigation fusion according to claim 3, characterized in that, In the process of fusing shipborne inertial navigation and RTK, the pre-acquired installation deviation angle of the inertial navigation equipment is used to correct the deviation of the heading angle in the fused ship attitude data, so as to obtain the corrected true heading angle.

5. The intelligent ship pilotage method based on RTK and inertial navigation fusion according to claim 4, characterized in that, In the step of generating and displaying the real-time dynamic visualization interface, dynamically drawing ship icons in the shoreline visualization map specifically includes: The ship is represented by an arrow-shaped graphic symbol, and the direction of the arrow-shaped graphic symbol is determined according to the corrected true heading angle in the ship's attitude data. Then, the arrow-shaped graphic symbol is rendered with different colors according to the roll angle in the ship's attitude data. Specifically, green is used when the roll angle is less than 3 degrees, yellow is used when the roll angle is between 3 and 5 degrees, and red is used when the roll angle is greater than 5 degrees.

6. An intelligent ship pilotage system based on RTK and inertial navigation fusion, characterized in that, It includes, in sequence, a differential correction calculation module, a curve model construction and visualization map generation module, a shipborne inertial navigation and RTK fusion module, and a real-time dynamic visualization interface generation and display module. The differential correction calculation module deploys a GNSS receiver as an RTK base station at the highest point of the dock and obtains its position coordinates. Then, it calibrates the position coordinates of the RTK base station using a surveying instrument to obtain calibrated position coordinates. Based on the calibrated position coordinates and the satellite observation coordinates received by the RTK base station in real time, it calculates the differential correction in real time and broadcasts it to the shipborne terminal via a wireless link. The curve model construction and visualization map generation module uses RTK surveying equipment to collect the three-dimensional coordinates of multiple feature points along the edge of the wharf shoreline. Based on the three-dimensional coordinates of the multiple feature points, a cubic spline interpolation algorithm is used to construct a three-dimensional shoreline curve model that satisfies positional continuity, first derivative continuity, and second derivative continuity. Based on the three-dimensional shoreline curve model and the pre-acquired basic geographic information of the wharf, a shoreline visualization map with geographic coordinates is generated. The shipborne inertial navigation and RTK fusion module uses the midpoint of the ship's deck as the origin and the bow direction as the reference point. Axis, starboard direction The axis, perpendicular to the deck plane and upward is A ship coordinate system is established; the differential correction amount is received in real time by an RTK mobile station set up on the ship; the corrected ship position coordinates in the ship coordinate system are calculated based on the differential correction amount and the original observation coordinates received in real time by the RTK mobile station; the ship attitude angle in its own inertial navigation coordinate system is obtained by an inertial navigation device set up on the ship; and the rotation angle of the inertial navigation coordinate system relative to the ship coordinate system is measured by an angle measuring device, thereby generating a rotation matrix. The offset distance between the origin of the inertial navigation coordinate system and the origin of the ship coordinate system is measured using a distance measurement device, and then a translation vector is generated. Based on the rotation matrix and translation vector, the ship attitude angle in the inertial navigation coordinate system is transformed to the ship coordinate system using a coordinate transformation method. Then, the Kalman filter algorithm is used to fuse the transformed ship attitude angle with the corrected ship position coordinates to output the fused ship pose data. The real-time dynamic visualization interface generation and display module uses the MQTT protocol to encapsulate the fused ship posture data into JSON format and transmits it to the shore-based processing center via a 5G network. After receiving the ship posture data, the shore-based processing center uses a coordinate transformation method to convert the ship posture data from the ship coordinate system to the local plane coordinate system corresponding to the shoreline visualization map. Based on the ship posture data in the local plane coordinate system, the system dynamically draws ship icons in the shoreline visualization map, generates a real-time dynamic visualization interface, and pushes the real-time dynamic visualization interface to the display terminal on the tugboat via the RTSP protocol for display, so as to assist the tugboat in adjusting thrust and direction and realize intelligent ship pilotage.

7. The intelligent ship pilotage system based on RTK and inertial navigation fusion according to claim 6, characterized in that, The curve model construction and visualization map generation module uses a cubic spline interpolation algorithm to construct a three-dimensional shoreline curve model, specifically including: At any two adjacent feature points and The formed interval [ Within [the range], define a cubic polynomial as the difference function: ， in, The coefficients for each interval are solved by imposing the following constraints on the difference function. : Positional continuity: ; The first derivative is continuous: ; The second derivative is continuous: ; Natural boundary conditions: ; Based on the above constraints, establish a system of linear equations, and solve for the coefficients in all intervals using the system of linear equations. This generates a continuous and smooth three-dimensional shoreline curve model.

8. The intelligent ship pilotage system based on RTK and inertial navigation fusion according to claim 6, characterized in that, The shipborne inertial navigation and RTK fusion module employs a Kalman filter algorithm for data fusion, specifically including: Constructing state vectors ,in,( ( ) represents the true position coordinates of the ship to be estimated. () represents the actual attitude angle of the ship to be estimated; Construct observation vectors ,in,( ) represents the corrected ship position coordinates obtained from the RTK rover station calculation. ( ) represents the ship's attitude angles obtained by the inertial navigation system; Using the Kalman filter algorithm, the observation vector For the state vector Recursive filtering calculations are performed to output the optimal estimated fused ship pose data.

9. The intelligent ship pilotage system based on RTK and inertial navigation fusion according to claim 8, characterized in that, In the shipborne inertial navigation and RTK fusion module, the pre-acquired installation deviation angle of the inertial navigation equipment is used to correct the deviation of the heading angle in the fused ship attitude data, so as to obtain the corrected true heading angle.

10. The intelligent ship pilotage system based on RTK and inertial navigation fusion according to claim 9, characterized in that, The real-time dynamic visualization interface generation and display module, specifically includes dynamically drawing ship icons in the shoreline visualization map, which includes: The ship is represented by an arrow-shaped graphic symbol, and the direction of the arrow-shaped graphic symbol is determined according to the corrected true heading angle in the ship's attitude data. Then, the arrow-shaped graphic symbol is rendered with different colors according to the roll angle in the ship's attitude data. Specifically, green is used when the roll angle is less than 3 degrees, yellow is used when the roll angle is between 3 and 5 degrees, and red is used when the roll angle is greater than 5 degrees.

Images