Control method and device for navigation of adjoint piloting boat, electronic equipment and medium

Abstract

The invention relates to a control method and device for accompanying piloting boat navigation, electronic equipment and a medium, and belongs to the technical field of ship path planning, and the method comprises the steps: obtaining ship cross-domain multi-modal fusion data of a target navigation water area; obtaining a feature association probability distribution database based on ship cross-domain multi-modal fusion data; extracting route data in the feature association probability distribution database to obtain a route database of a ship drawn by a destination; determining a ship intention inference model based on the feature association probability distribution database, the destination traction ship route database, the destination of the target ship and the time-space behavior data of the target ship; determining an optimal path of the target ship in the target navigation water area based on a ship intention inference model; and controlling the navigational speed, the navigational direction and the rudder angle of the target piloting boat according to the optimal path based on a PID controller. According to the invention, the sailing efficiency of the accompanying piloting boat is improved.

Description

Technical Field

[0001] This invention relates to the field of ship route planning technology, and in particular to a control method, device, electronic equipment and medium for accompanying pilot boat navigation. Background Technology

[0002] Pilotage plays a vital role in ensuring safe navigation, improving port operational efficiency, and maintaining navigational order. It is an important component and key node in my country's maritime transport and port operations. Accidents involving ships in the main channel or narrow waterways of a port can affect navigation and even render the port inoperable. The significance of pilotage lies in ensuring the safety of ships navigating within the port and port facilities, and maintaining normal port order. The safety of pilotage operations is directly dependent on the auxiliary capabilities of the pilot boat and the operating environment of the pilot. Especially in large ports, where the daily pilotage workload can reach hundreds of times, the response speed, maneuvering precision, and environmental adaptability of pilot boats have become key bottlenecks restricting port throughput capacity and shipping efficiency.

[0003] Ship pilotage relies on pilots boarding the vessel to conduct pilotage operations, which is costly and risky, and cannot meet the needs of ships entering and leaving ports in low visibility conditions such as fog or in adverse sea conditions. Moreover, traditional pilotage methods cannot guarantee the normal operation of pilotage services under conditions of larger ships and denser traffic, and cannot meet the ever-increasing demands of port development for the efficiency of ship entry and exit. In addition, although existing research has made some progress in the field of autonomous ship navigation control, it has mostly focused on route optimization for ocean-going merchant ships or simple application scenarios for unmanned inland waterway vessels, failing to fully consider the complexity of ship pilotage conditions and the special characteristics of the operation. Pilotage vessels operate in a complex and diverse navigation environment, encounter complex ship scenarios, are difficult to predict the collaborative intentions of multiple ships, and ship avoidance behaviors change frequently and dynamically.

[0004] In summary, existing technologies lack a control method for accompanying pilot boat navigation to improve the navigation efficiency of pilot boats. Summary of the Invention

[0005] In view of this, it is necessary to provide a control method, device, electronic equipment and medium for the navigation of accompanying pilot boats, so as to solve the problem of low navigation efficiency of accompanying pilot boats in the prior art.

[0006] To address the aforementioned problems, in a first aspect, the present invention provides a control method for the navigation of an accompanying pilot boat, comprising: Acquire cross-domain multimodal fusion data of ships in the target navigation waters. The cross-domain multimodal fusion data of ships includes: timestamp, latitude and longitude coordinates, speed above ground, heading above ground, heading, and rate of turn. A database of feature association probability distributions was obtained based on cross-domain multimodal fusion data of ships; Extract route data from the feature association probability distribution database to obtain a destination-to-towing ship route database; A ship intent inference model is determined based on a feature association probability distribution database, a destination-driven ship route database, the destination of the target ship, and the spatiotemporal behavior data of the target ship. Determine the optimal path for a target vessel in the target navigation area based on a vessel intent inference model; The speed, heading, and rudder angle of the target pilot boat are controlled based on the optimal path using a PID controller.

[0007] In one possible implementation, the feature association probability distribution database obtained based on cross-domain multimodal fusion data of ships includes: Ship trajectory data is obtained based on cross-domain multimodal fusion data of ships; Spatiotemporal analysis of the ship trajectory data is performed to obtain a feature correlation probability distribution database.

[0008] In one possible implementation, the step of performing spatiotemporal analysis on the ship trajectory data to obtain a feature association probability distribution database includes: The spatiotemporal correlation analysis algorithm is used to perform spatiotemporal analysis on the trajectory data of the vessels to obtain a spatiotemporal behavioral feature correlation model.

[0009] In one possible implementation, the step of extracting route data from the feature association probability distribution database to obtain a destination-to-towing vessel route database includes: Based on clustering and triangulation algorithms, route data in the feature association probability distribution database is extracted to obtain a destination-to-traction ship route database.

[0010] In one possible implementation, determining the optimal path for the target vessel in the target navigation waters based on the vessel intent inference model includes: The navigation intentions of other ships are determined by using a Bayesian theory-based ship intention inference model, which includes the other ship's route, position, heading, and speed. Determine the optimal path for the target vessel in the target navigation waters based on the navigation intentions of other vessels.

[0011] In one possible implementation, the PID controller includes: Speed ​​controller, heading controller, rudder angle controller, and track controller.

[0012] One possible implementation also includes: The heading controller and track controller are trained using a deep reinforcement learning algorithm to obtain the trained heading controller and track controller.

[0013] Secondly, the present invention also provides a control device for the navigation of an accompanying pilot boat, comprising: The multimodal fusion data acquisition module is used to acquire cross-domain multimodal fusion data of ships in the target navigation waters. The cross-domain multimodal fusion data of ships includes: timestamp, latitude and longitude coordinates, speed above ground, heading above ground, heading, and turning rate. The probability distribution database module is used to obtain a feature association probability distribution database based on cross-domain multimodal fusion data of ships; The route database module is used to extract route data from the feature association probability distribution database to obtain a destination-to-towing ship route database. The ship intent inference model determination module is used to determine the ship intent inference model based on the feature association probability distribution database, the destination-to-traction ship route database, the destination of the target ship, and the spatiotemporal behavior data of the target ship. The optimal path determination module is used to determine the optimal path of a target vessel in the target navigation waters based on the vessel intent inference model. The navigation control module is used to control the speed, heading, and rudder angle of the target pilot boat based on the optimal path using a PID controller.

[0014] Thirdly, the present invention also provides an electronic device, including a memory and a processor, wherein, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the control method for the navigation of an accompanying pilot boat as described in any of the above implementations.

[0015] Fourthly, the present invention also provides a computer-readable storage medium for storing a computer-readable program or instructions, which, when executed by a processor, can implement the steps of the control method for the navigation of an accompanying pilot boat as described in any of the above implementations.

[0016] The beneficial effects of this invention are as follows: This invention provides a control method for accompanying pilot boat navigation, acquiring cross-domain multimodal fusion data of the ship. This cross-domain multimodal fusion data includes: timestamps, latitude and longitude coordinates, ground speed, ground heading, bow direction, and turning rate. Based on the cross-domain multimodal fusion data, a feature association probability distribution database is obtained. Route data is extracted from the feature association probability distribution database to obtain a destination-to-tow ship route database. Based on the feature association probability distribution database, the destination-to-tow ship route database, the destination of the target ship, and the spatiotemporal behavior data of the target ship, a ship intention inference model is determined. This ship intention inference model is used to determine the navigation intentions of other ships to provide a basis for the navigation of this ship. Based on the ship intention inference model, the optimal path of the target ship in the target navigation waters is determined. A PID controller controls the speed, heading, and rudder angle of the target pilot boat according to the optimal path. This invention, through the ship intention inference model, can estimate the navigation intentions of other ships in the target navigation waters, thereby determining the optimal path based on the navigation intentions of other ships, avoiding unnecessary efficiency waste during the target ship's navigation, and thus improving navigation efficiency. Attached Figure Description

[0017] Figure 1 A flowchart illustrating an embodiment of the control method for accompanying pilot boat navigation provided by the present invention; Figure 2 A schematic diagram of the overall technical solution of an embodiment of the control method for accompanying pilot boat navigation provided by the present invention; Figure 3 A flowchart illustrating the technology of ship spatiotemporal behavior feature mining and intent inference, as provided by the present invention, for an embodiment of a control method for accompanying pilot boat navigation; Figure 4 A flowchart of a ship dynamic path planning decision support technology, which is an embodiment of a control method for the navigation of an accompanying pilot boat provided by the present invention; Figure 5 A flowchart of PID-based navigation control technology for an accompanying pilot boat, as provided by the present invention, is shown below; Figure 6 A schematic diagram of the DDPG algorithm framework is provided as an embodiment of the control method for accompanying pilot boat navigation provided by the present invention. Figure 7 A schematic flowchart of an embodiment of a device for controlling the navigation of an accompanying pilot boat provided by the present invention; Figure 8 A schematic diagram of an embodiment of the electronic device provided by the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0019] In the description of the embodiments of the present invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0020] The terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.

[0021] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0022] Before demonstrating the embodiments, the following terms will be explained.

[0023] This invention provides a control method, device, electronic equipment, and medium for accompanying pilot boat navigation, which are described below.

[0024] Figure 1 This is a schematic flowchart of an embodiment of the control method for the navigation of an accompanying pilot boat provided by the present invention, as shown below. Figure 1 As shown, the control method for the navigation of the accompanying pilot boat includes: S101. Obtain cross-domain multimodal fusion data of ships in the target navigation waters. The cross-domain multimodal fusion data of ships includes: timestamp, latitude and longitude coordinates, speed above ground, heading above ground, heading, and turning rate. S102. Obtain a feature association probability distribution database based on cross-domain multimodal fusion data of ships; S103. Extract route data from the feature association probability distribution database to obtain a destination-to-towing ship route database. S104. A model for determining ship intention based on a feature association probability distribution database, a destination-driven ship route database, the destination of the target ship, and the spatiotemporal behavior data of the target ship. S105. Determine the optimal path of the target vessel in the target navigation waters based on the vessel intent inference model; S106. Based on the PID controller, the speed, heading and rudder angle of the target pilot boat are controlled according to the optimal path.

[0025] Compared with existing technologies, this embodiment provides a control method for accompanying pilot boat navigation. It acquires cross-domain multimodal fusion data of the vessel, including timestamps, latitude and longitude coordinates, ground speed, ground heading, bow direction, and turning rate. Based on this data, a feature association probability distribution database is obtained. Route data is extracted from this database to obtain a destination-to-tethered vessel route database. Based on the feature association probability distribution database, the destination-to-tethered vessel route database, the target vessel's destination, and its spatiotemporal behavior data, a vessel intent inference model is determined. This model is used to determine the navigation intent of other vessels, providing a basis for the vessel's navigation. The optimal path for the target vessel in the target navigation area is determined based on this model. A PID controller controls the target pilot boat's speed, heading, and rudder angle according to the optimal path. This invention, through the vessel intent inference model, can estimate the navigation intent of other vessels in the target navigation area, thereby determining the optimal path based on their intentions. This avoids unnecessary efficiency waste during the target vessel's navigation, thus improving navigation efficiency.

[0026] In some embodiments of the present invention, such as Figure 2 The diagram shown illustrates the overall technical solution for this example, which comprises three main steps. The first step is a method for mining ship spatiotemporal behavioral characteristics and inferring intent; the second step is a method for ship dynamic path planning and decision support; and the third step is a PID-based accompanying pilot boat navigation control technology. The first step includes steps S101 to S104, the second step is S105, and the third step is S106.

[0027] In some embodiments of the present invention, such as Figure 3The diagram shows the technical flowchart for ship spatiotemporal behavior feature mining and intent inference. In step S101, cross-domain multimodal fusion data of ships in the target navigation waters are acquired through equipment such as binocular cameras, radar, and AIS carried by the escort vessel. This data includes: timestamp t, latitude and longitude coordinates (lon, lat), ground speed sog, ground heading cog, bow heading hdg, turning rate rot, etc. A spatiotemporal behavior feature association model is constructed using spatiotemporal correlation analysis methods to reveal the ship spatiotemporal behavior feature association mechanism and establish a probability distribution database of ship spatiotemporal behavior feature association based on trajectory distance traction. This provides basic support for the autonomous design of ship route planning and intent inference based on behavior features.

[0028] In some embodiments of the present invention, the feature association probability distribution database obtained based on cross-domain multimodal fusion data of ships includes: Ship trajectory data is obtained based on cross-domain multimodal fusion data of ships; Spatiotemporal analysis of the ship trajectory data is performed to obtain a feature correlation probability distribution database.

[0029] In some embodiments of the present invention, the step of performing spatiotemporal analysis on the ship trajectory data to obtain a feature association probability distribution database includes: The spatiotemporal correlation analysis algorithm is used to perform spatiotemporal analysis on the trajectory data of the vessels to obtain a spatiotemporal behavioral feature correlation model.

[0030] In a specific embodiment of the present invention, the specific steps for performing spatiotemporal analysis on the trajectory data of the vessel using a spatiotemporal correlation analysis algorithm are as follows: 1) Trajectory segmentation and feature extraction: Track segmentation: This involves dividing a long ship's navigation path into meaningful segments based on its behavioral patterns. Common segmentation points include: sharp turning points, acceleration / deceleration points, and the start and end points of berthing. This transforms a trajectory into a series of behavioral units.

[0031] Feature extraction: For each trajectory Ti, extract its macroscopic and microscopic features to form the feature vector Fi of that trajectory segment.

[0032] 2) Trajectory distance measurement and similarity calculation: Using Dynamic Time Warping (DTW), the cumulative distance is calculated by finding the optimal trajectory alignment between two time series.

[0033] Assume two trajectories: P = ( ) and Q=( ).

[0034] Construct an m×n matrix, where the matrix element (i, j) represents a point. p iand q j Euclidean distance between d ( p i , q j ).

[0035] The DTW algorithm finds the minimum cumulative path across the matrix, and its recursive definition is generally as follows:

[0036] final .

[0037] 3) Establish a spatiotemporal behavioral feature correlation model Using the distance metric described above, the pairwise similarity distances between all trajectory segments in the historical database are obtained. Based on this "distance matrix," a probabilistic model is constructed.

[0038] Clustering algorithms are used to categorize similar trajectory segments, with each cluster representing a navigation behavior pattern.

[0039] The probability of transitioning from behavior pattern C to behavior pattern T is calculated using the following correlation probability distribution function, resulting in a spatiotemporal behavior feature correlation model. Then, a database of ship spatiotemporal behavior feature correlation probability distributions based on trajectory distance traction is generated according to this model.

[0040] In some embodiments of the present invention, the step of extracting route data from the feature association probability distribution database to obtain a destination-to-traction vessel route database includes: Based on clustering and triangulation algorithms, route data in the feature association probability distribution database is extracted to obtain a destination-to-traction ship route database.

[0041] In a specific embodiment of the present invention, based on ship trajectory data, a ship route extraction method using cluster analysis and triangulation algorithms is established to create a destination-oriented ship route database, enabling autonomous design of ship planning routes based on factors such as ship category and scale level. The ship route design method based on cluster analysis is mainly used to determine the ship's planned route, clarify the ship's turning points and destination coordinates, guide the ship towards its destination, and provide support for the ship's global path planning. Specific implementation steps: Based on ship characteristics and course intentions, typical behavioral patterns are retrieved from the "Probability Distribution Database of Ship Spatiotemporal Behavior Characteristics Related to Trajectory Distance" to generate typical route segments (behavioral patterns). C r}

[0042]

[0043] Candidate pattern clusters for each stage C r Calculate its cluster center trajectory T m .

[0044]

[0045] Where D is the trajectory distance metric (DTW distance).

[0046] The navigable waterway is modeled using a triangulation algorithm, ensuring that the generated global route sketch R contains route points within the safe triangles of the triangulation network; the turning radius at turning points is then determined. R turn It is greater than the ship's minimum turning radius.

[0047] In a specific embodiment of the present invention, step S104, the construction of the ship intent inference model specifically includes the following steps: Based on the destination-driven ship route database and the ship spatiotemporal behavior feature association database, and combined with ship destination and spatiotemporal behavior data, a ship intention inference model is established. Using Bayesian theory and deep neural network methods, the intention of ship behavior features such as route, position, heading and speed is accurately inferred.

[0048] The ship intent inference model is used to accurately infer the planned routes, positions, headings, and speeds of other ships near the ship, clarify the turning points and destination coordinates of other ships and their behavioral patterns, provide support for ship encounter situation judgment, and thus support intelligent ship risk avoidance decision-making.

[0049] Specific implementation steps: 1) Extract feature vectors from real-time acquired raw ship trajectory data. F t As model input 2) Clarify the concept of "intention": Macro-level intention (strategic level): What is the ultimate destination? For example: intention to enter the port, intention to leave the port, intention to anchor.

[0050] Mid-level intentions (tactical level): What are the key actions to be taken next? For example: intention to maintain course and speed, intention to turn right, intention to cross the channel, intention to overtake.

[0051] Micro-level intentions (operational level): subtle manipulations and adjustments. For example: intending to move forward at a slight speed, or intending to slow down.

[0052] 3) Construct an intent inference model Inference based on probabilistic graphical models (Bayesian networks):

[0053] T': The intentions of the other ship to be inferred C: Real-time observed data P(T'|C'): Posterior probability, representing the probability that the ship has intention T' given the observed data C'.

[0054] P(C'|T'): Likelihood probability, representing the probability that the current data C' will be observed if the ship does indeed have intention T'. This probability is calculated from the observation model in the database of destination-to-towing ship routes.

[0055] P(T'): Prior probability, representing the probability that the ship has intention T' in the absence of any observed data C'. This probability is based on historical statistics from a database of destination-to-towing vessel routes.

[0056] P(C'): Evidence factor.

[0057] Inference based on deep neural networks (DNN): Input the feature sequence of another ship over a period of time The algorithm is trained using a high-quality tag database constructed from a database of destination-driven ship routes, outputting vectors representing the probabilities of various intentions. Finally, a weighted fusion of the Bayesian and deep neural network decisions yields the optimal intention inference result.

[0058] In some embodiments of the present invention, determining the optimal path of a target vessel in the target navigation waters based on the vessel intent inference model includes: The navigation intentions of other ships are determined by using a Bayesian theory-based ship intention inference model, which includes the other ship's route, position, heading, and speed. Determine the optimal path for the target vessel in the target navigation waters based on the navigation intentions of other vessels.

[0059] In a specific embodiment of the present invention, after determining the navigation intention of another vessel, the target vessel takes evasive action based on that intention. First, the target vessel determines the navigable waterway within the target waters.

[0060] Specifically, the first step is to identify navigational obstructions and shallow water areas near the ship's route based on data such as electronic nautical charts, ship draft, and tides, combined with the ship's characteristics and the specific conditions of the water environment, and to determine the range and boundaries of navigable waters.

[0061] Electronic chart processing: The charts are vectorized and displayed in the S57 international standard format. The charts show anchorage markings, and the color scheme is well-defined. The size and shape of the icons and symbols are clear and standardized.

[0062] Clear Display of Elements: Standard elements on electronic nautical charts are clearly displayed, showing chart elements such as hazards, obstacles, and shallows in standard display mode. The chart interface offers three color themes: daytime, dusk, and nighttime. For different modes and backgrounds, the display color of vessel information is made more prominent.

[0063] Water depth data display: Water depth data display automatically selects the isobaths using the electronic chart function and displays the chart with reference to the selected safe isobath spacing.

[0064] Based on data such as electronic nautical charts, ship draft, and tides, the scope and boundaries of navigable waters are determined.

[0065] Safe water depth for ship navigation: , For electronic nautical chart water depth; For high tide; For the ship's draft.

[0066] Utilizing the safe water depth for ship navigation The system searches for the boundaries of areas in the electronic nautical chart model that do not meet the safe depth requirements for ship navigation, constructs no-navigation zones, and then determines the boundaries of navigable waters.

[0067] Step 2: Determine the planned route for the target vessel Identify the target vessel's turning point and endpoint: Based on the nautical chart's water depth data and the navigable water area, and according to the specific characteristics of the target vessel and the water environment conditions, mark the navigational obstacles and shallow water areas near the target vessel's route, select a route with sufficient safe water depth to ensure that the vessel has enough margin of safety; and clarify the vessel's navigable water area, avoiding narrow and complex channels and busy traffic areas as much as possible, taking into account the vessel's maneuverability and turning radius, and planning a reasonable turning point.

[0068] Determining if Collision Avoidance Maneuvers are Required: Based on vessel route data, it is determined whether an encounter with nearby vessels is possible and whether collision avoidance maneuvers are necessary. For scenarios where collision avoidance maneuvers are not required, the target vessel maintains its speed and heading, continuing along the predetermined route. For scenarios requiring collision avoidance, based on the route and behavioral characteristics of the target vessel and other vessels, as well as the surrounding water environment, an optimized function based on the gravitational field of the route and the repulsive field of the hazard source is created using an improved artificial potential field method. This function, combined with the vessel's navigation risk, determines the intelligent collision avoidance algorithm.

[0069] Taking into account collision avoidance rules and ship maneuverability, this paper studies a dynamic path search method for ships using a hyperheuristic algorithm, and completes the real-time planning and output of the ship's dynamic path through human-computer interaction.

[0070] Based on the nautical chart ranging function, the distance and bearing between any two points and the cumulative distance between multiple points are displayed; based on the ship collision avoidance rules and the characteristics of the port area navigation environment, the expert system method is used to study the ship dynamic path planning framework and clarify the ship dynamic path planning mechanism; all ships are set with arbitrary destinations, and the arrival time of the ship is calculated and updated in real time according to the ship's real-time speed and heading.

[0071] By comparing with actual routes, the model and method are validated and corrected to achieve dynamic route planning and decision-making for ships in complex environments.

[0072] In some embodiments of the present invention, the PID controller includes: Speed ​​controller, heading controller, rudder angle controller, and track controller.

[0073] In a specific example of the present invention, after determining the optimal path of the target vessel, a PID (proportion integration differentiation) control method is used to design a speed controller, heading controller, rudder angle controller, and track controller for the escort vessel.

[0074] 1) Speed ​​controller: Input the trajectory to plan its desired speed Actual measured speed u Speed ​​error

[0075] Input the speed error into the PID controller

[0076] , , The proportional, integral, and derivative coefficients of the heading controller.

[0077] 2) Heading Controller: Input the trajectory to plan its desired speed Actual measured speed Speed ​​error

[0078] Input the speed error into the PID controller.

[0079]

[0080] 3) Rudder Angle Controller: This controller processes command signals... This is converted into actual servo signals.

[0081] 4) Track Controller: Given the current position (x, y), the system continuously calculates the Cross-Track Error (XTE) from the current position to the planned path. The PID controller calculates the desired heading adjustment required to eliminate this error based on the magnitude of the XTE and outputs the desired speed. and expected speed .

[0082] In addition, this embodiment also uses the Deep Deterministic Policy Gradient (DDPG) deep reinforcement learning algorithm as the basis for unmanned vessel control. The deep reinforcement learning algorithm is used to train the heading and trajectory controllers of the unmanned vessel, and then the trained controller model is calibrated using real data to improve the disturbance rejection capability of the control system.

[0083] DDPG (Deep Deterministic Policy Gradient Algorithm) is a reinforcement learning algorithm applicable to continuous action spaces.

[0084] The ship control problem is modeled as a reinforcement learning environment, and the specific steps are as follows: Status: including speed error Heading error Lateral error (XTE), rudder angle, thrust, and historical status information, etc.

[0085] Action: This refers to the PID parameters that need to be optimized, for example... .

[0086] Reward: A carefully designed function to encourage good behavior from the controller.

[0087]

[0088] A negative reward function means that the greater the error and the greater the steering error, the greater the penalty. These are the weighting coefficients.

[0089] Controller training The parameters are randomly initialized using the Xavier method. After training using the DDPG algorithm, the model is continuously trained until a satisfactory model is obtained. Finally, the trained model is used in the navigation control of the accompanying pilot boat. The overall control scheme is verified using randomly generated routes, and its generalization and robustness are evaluated.

[0090] In some embodiments of the present invention, it further includes: The heading controller and track controller are trained using a deep reinforcement learning algorithm to obtain the trained heading controller and track controller.

[0091] To better implement the control method for accompanying pilot boat navigation in the embodiments of the present invention, based on the control method for accompanying pilot boat navigation, correspondingly, as follows: Figure 7 As shown, this embodiment of the invention also provides a control device for the navigation of an accompanying pilot boat. The control device 700 for the navigation of an accompanying pilot boat includes: The multimodal fusion data acquisition module 701 is used to acquire cross-domain multimodal fusion data of ships in the target navigation waters. The cross-domain multimodal fusion data of ships includes: timestamp, latitude and longitude coordinates, speed above ground, heading above ground, heading, and turning rate. The probability distribution database determination module 702 is used to obtain a feature association probability distribution database based on cross-domain multimodal fusion data of ships; The route database determination module 703 is used to extract route data from the feature association probability distribution database to obtain a destination-to-towing ship route database. The ship intent inference model determination module 704 is used to determine the ship intent inference model based on the feature association probability distribution database, the destination-to-traction ship route database, the destination of the target ship, and the spatiotemporal behavior data of the target ship. The optimal path determination module 705 is used to determine the optimal path of the target vessel in the target navigation waters based on the vessel intent inference model; The navigation control module 706 is used to control the speed, heading, and rudder angle of the target pilot boat based on the optimal path using a PID controller.

[0092] The control device 700 for accompanying pilot boat navigation provided in the above embodiments can realize the technical solution described in the above embodiments of the control method for accompanying pilot boat navigation. The specific implementation principle of each module or unit can be found in the corresponding content in the above embodiments of the control method for accompanying pilot boat navigation, which will not be repeated here.

[0093] like Figure 8 As shown, the present invention also provides an electronic device 800. The electronic device 800 includes a processor 801, a memory 802, and a display 803. Figure 8 Only some components of the electronic device 800 are shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead.

[0094] In some embodiments, processor 801 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 802 or process data, such as a control method for the navigation of an accompanying pilot boat in this invention.

[0095] In some embodiments, processor 801 may be a single server or a group of servers. The server group may be centralized or distributed. In some embodiments, processor 801 may be local or remote. In some embodiments, processor 801 may be implemented on a cloud platform. In some embodiments, the cloud platform may include private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, internal cloud, multi-cloud, etc., or any combination thereof.

[0096] In some embodiments, memory 802 may be an internal storage unit of electronic device 800, such as a hard disk or memory of electronic device 800. In other embodiments, memory 802 may also be an external storage device of electronic device 800, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on electronic device 800.

[0097] Furthermore, the memory 802 may include both internal storage units of the electronic device 800 and external storage devices. The memory 802 is used to store application software and various types of data installed on the electronic device 800.

[0098] In some embodiments, display 803 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 803 is used to display information from electronic device 800 and to display a visual user interface. Components 801-803 of electronic device 800 communicate with each other via a system bus.

[0099] In one embodiment, when processor 801 executes a control program for the navigation of an accompanying pilot boat stored in memory 802, the following steps can be implemented: Acquire cross-domain multimodal fusion data of ships in the target navigation waters. The cross-domain multimodal fusion data of ships includes: timestamp, latitude and longitude coordinates, speed above ground, heading above ground, heading, and rate of turn. A database of feature association probability distributions was obtained based on cross-domain multimodal fusion data of ships; Extract route data from the feature association probability distribution database to obtain a destination-to-towing ship route database; A ship intent inference model is determined based on a feature association probability distribution database, a destination-driven ship route database, the destination of the target ship, and the spatiotemporal behavior data of the target ship. Determine the optimal path for a target vessel in the target navigation area based on a vessel intent inference model; The speed, heading, and rudder angle of the target pilot boat are controlled based on the optimal path using a PID controller.

[0100] It should be understood that when the processor 801 executes a control program for the navigation of an accompanying pilot boat stored in the memory 802, in addition to the functions mentioned above, it can also perform other functions, as detailed in the description of the corresponding method embodiments above.

[0101] Furthermore, the embodiments of the present invention do not specifically limit the type of electronic device 800 mentioned. Electronic device 800 can be a mobile phone, tablet computer, personal digital assistant (PDA), wearable device, laptop computer, or other portable electronic device. Exemplary embodiments of portable electronic devices include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic device can also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, electronic device 800 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0102] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0103] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A control method for accompanying navigation of a pilot boat, characterized by, include: Acquire cross-domain multimodal fusion data of ships in the target navigation waters. The cross-domain multimodal fusion data of ships includes: timestamp, latitude and longitude coordinates, speed above ground, heading above ground, heading, and rate of turn. A database of feature association probability distributions was obtained based on cross-domain multimodal fusion data of ships; Extract route data from the feature association probability distribution database to obtain a destination-to-towing ship route database; A ship intent inference model is determined based on a feature association probability distribution database, a destination-driven ship route database, the destination of the target ship, and the spatiotemporal behavior data of the target ship. Determine the optimal path for a target vessel in the target navigation area based on a vessel intent inference model; The speed, heading, and rudder angle of the target pilot boat are controlled based on the optimal path using a PID controller.

2. The control method for the navigation of an accompanying pilot boat according to claim 1, characterized in that, The feature association probability distribution database obtained based on cross-domain multimodal fusion data of ships includes: Ship trajectory data is obtained based on cross-domain multimodal fusion data of ships; Spatiotemporal analysis of the ship trajectory data is performed to obtain a feature correlation probability distribution database.

3. The control method for the navigation of an accompanying pilot boat according to claim 2, characterized in that, The process of performing spatiotemporal analysis on the ship trajectory data to obtain a feature association probability distribution database includes: The spatiotemporal correlation analysis algorithm is used to perform spatiotemporal analysis on the trajectory data of the vessels to obtain a spatiotemporal behavioral feature correlation model.

4. The control method for the navigation of an accompanying pilot boat according to claim 1, characterized in that, The step of extracting route data from the feature association probability distribution database to obtain a destination-to-towing vessel route database includes: Based on clustering and triangulation algorithms, route data in the feature association probability distribution database is extracted to obtain a destination-to-traction ship route database.

5. The control method for the navigation of an accompanying pilot boat according to claim 1, characterized in that, The determination of the optimal path for a target vessel in the target navigation waters based on the vessel intent inference model includes: The navigation intentions of other ships are determined by using a Bayesian theory-based ship intention inference model, which includes the other ship's route, position, heading, and speed. Determine the optimal path for the target vessel in the target navigation waters based on the navigation intentions of other vessels.

6. The control method for the navigation of an accompanying pilot boat according to claim 1, characterized in that, The PID controller includes: Speed ​​controller, heading controller, rudder angle controller, and track controller.

7. The control method for the navigation of an accompanying pilot boat according to claim 1, characterized in that, Also includes: The heading controller and track controller are trained using a deep reinforcement learning algorithm to obtain the trained heading controller and track controller.

8. A control device for the navigation of an accompanying pilot boat, characterized in that, include: The multimodal fusion data acquisition module is used to acquire cross-domain multimodal fusion data of ships in the target navigation waters. The cross-domain multimodal fusion data of ships includes: timestamp, latitude and longitude coordinates, speed above ground, heading above ground, heading, and turning rate. The probability distribution database determination module is used to obtain a feature association probability distribution database based on cross-domain multimodal fusion data of ships; The route database determination module is used to extract route data from the feature association probability distribution database to obtain a destination-to-towing vessel route database. The ship intent inference model determination module is used to determine the ship intent inference model based on the feature association probability distribution database, the destination-to-traction ship route database, the destination of the target ship, and the spatiotemporal behavior data of the target ship. The optimal path determination module is used to determine the optimal path of a target vessel in the target navigation waters based on the vessel intent inference model. The navigation control module is used to control the speed, heading, and rudder angle of the target pilot boat based on the optimal path using a PID controller.

9. An electronic device, characterized in that, Including memory and processor, among which, The memory is used to store programs; The processor, coupled to the memory, is used to execute the program stored in the memory to implement the steps in the control method for the navigation of an accompanying pilot boat as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the control method for the navigation of an accompanying pilot boat as described in any one of claims 1 to 7.

Images