Intelligent decision-making method based on maritime collision avoidance rule

Abstract

The invention discloses an intelligent decision-making method based on a maritime collision avoidance rule, and belongs to the technical field of intelligent ship collision avoidance decision-making, and the method comprises the following steps: S10, setting a global route of a ship autonomous cruise task; s20, obtaining ship positioning and environment perception prediction data; s30, acquiring a ship local reference trajectory through a local trajectory calculation module; s40, constructing a ship navigation encounter situation model; s50, constructing a multi-element composite collision avoidance risk assessment model; s60, constructing a behavior decision tree model based on a marine collision avoidance rule; s70, a trajectory planning module performs trajectory planning based on a behavior decision tree model result; s80, establishing a monitoring continuous decision operation mechanism; according to the invention, on the premise of guaranteeing the safety and efficiency of ship navigation, on the premise of complying with the International Marine Collision Regulation, the interactive collision avoidance behavior between the autonomous ship and the manned ship in a complex and changeable sea surface environment is explored, so that a rapid and reasonable collision avoidance decision is expected to be realized, and the navigation safety and efficiency are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent ship collision avoidance decision-making, and in particular relates to an intelligent decision-making method based on maritime collision avoidance rules. Background Art

[0002] With the rapid development of artificial intelligence and unmanned driving technology, unmanned ships have become an important research field in intelligent maritime transportation, among which ship collision avoidance decision-making is a crucial research content. Effective collision avoidance decision-making can ensure the navigation safety of ships and reduce loss of life and property and environmental pollution.

[0003] However, in the marine environment, wind, current, surge, and waves are highly time-varying, greatly affecting the safety of ship navigation. Secondly, offshore waters have strong structured navigation characteristics, with many types of navigation lanes and a large amount of navigation aid information, such as external environmental factors such as light buoys, beacons, channel buildings, navigation lights, and unruly small fishing boats. Therefore, intelligent decision-making methods based on maritime collision avoidance rules are needed to solve the above problems. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent decision-making method based on maritime collision avoidance rules to solve the problems raised in the above background technology.

[0005] To achieve the above-mentioned object, the present invention provides the following technical solution: an intelligent decision-making method based on maritime collision avoidance rules, comprising the following steps: S10. Setting the global route of the ship's autonomous cruise mission; S20, obtaining ship positioning and environment perception prediction data; S30, obtaining a local reference trajectory of the ship through a local trajectory calculation module; S40, constructing a ship navigation encounter situation model; S50. Construct a multivariate composite collision avoidance risk assessment model; S60. Construct a behavioral decision tree model based on maritime collision avoidance rules; S70, the trajectory planning module performs trajectory planning based on the behavior decision tree model results; S80. Establish a continuous monitoring and decision-making operation mechanism.

[0006] As a preferred solution, in "step S10", setting the global route of the ship autonomous cruising mission in step S10 includes the following steps: S101, marking the cruise mission route of the intelligent ship on the electronic map in sequence, and recording the positions on the marked cruise mission route to obtain a point queue of the cruise mission route; S102: Convert the cruise mission route point queue into a Cartesian coordinate system through a latitude, longitude and Cartesian coordinate conversion module; S103 , performing waypoint compensation on two adjacent points in the global route point queue in the Cartesian coordinate system using a sample point interpolation algorithm to obtain a desired point queue, and using the desired point queue as the global route.

[0007] As a preferred solution, in "step S20", the step S20 of obtaining ship positioning and environment perception prediction data includes the following steps: S201. The vessel obtains the current GPS longitude and latitude in real time through its own global positioning system, and converts it into a Cartesian coordinate system through a latitude and longitude & Cartesian coordinate system conversion module; S202: The vessel obtains its current speed in real time through its own configured speed calculation module; S203, the vessel obtains the current heading angle in real time through its own heading calculation module; S204: The ship obtains obstacle boundary data through the perception prediction module.

[0008] As a preferred solution, in "step S30", the step S30 of obtaining the local reference trajectory of the ship through the local trajectory calculation module includes the following steps: S301: Calling a trajectory planning module to generate a current local reference path based on the global route.

[0009] As a preferred solution, in "step S40", the step S40 of constructing the ship navigation encounter situation model includes the following steps: S401. The International Regulations for Preventing Collisions at Sea provide three types of encounter scenarios: head-on encounter, crossing encounter, and overtaking. However, in actual navigation, ships have different rights of way for different encounter situations, so it is necessary to classify the encounter scenarios in detail. S402. Based on the International Regulations for Preventing Collisions at Sea, a model of ship navigation encounter areas is established. Six encounter areas are divided based on the ship's orientation: bow, forward of starboard beam, aft of starboard beam, stern, aft of port beam, and forward of port beam. The areas are divided clockwise, with the ship's hull position as the center and the bow heading as arc 0. The coordinates of the boundary points of each encounter area are calculated using the arc area boundary point calculation tool. S403: Based on the ship navigation encounter area model, in each encounter area, in combination with the International Regulations for Preventing Collisions at Sea, a ship navigation encounter scenario situation model is set.

[0010] As a preferred solution, in "step S50", the step S50 constructs a multivariate composite collision avoidance risk assessment model including the following steps: S501, constructing a local reference path 2DBox detection frame set based on the local reference path data; S502, constructing a DCPA spatial collision avoidance risk detection model; S503: Construct a TCPA time collision avoidance risk detection model.

[0011] As a preferred solution, in "step S60", the step S60 constructs a behavior decision tree model based on maritime collision avoidance rules, including the following steps: S601, constructing a behavior decision tree feature state; S602, constructing a behavior decision tree classification result; S603: Build a behavior decision tree classification model, set corresponding decision results under different feature states to form training data, and train the behavior decision tree classification model; S604, setting the current data feature state set of the ship; S605: Based on the current data feature state set of the ship, predict the current decision tree result of the ship through the behavior decision tree classification model.

[0012] As a preferred solution, in "step S70", the trajectory planning module in step S70 performs trajectory planning based on the behavior decision tree model result, including the following steps: S701, if the decision result of the behavior decision tree model is , set the current maximum speed limit of the ship to 0, and plan the stop trajectory with the ship's maximum deceleration parameter through the trajectory planning module; S702, if the decision result of the behavior decision tree model is , based on the spatial closest collision risk distance and minimum collision avoidance time obtained in the multivariate composite collision avoidance risk assessment model constructed in the steps, the maximum speed boundary of the ship's avoidance is calculated; S703, if the decision result of the behavior decision tree model is , restore the maximum speed parameters of the ship itself, and plan a right-circling and overtaking trajectory through the trajectory planning module; S704, if the decision result of the behavior decision tree model is , restore the maximum speed parameters of the ship itself, and maintain the original channel trajectory through the trajectory planning module; S705: Send the trajectory planning module result to the ship control module for execution.

[0013] As a preferred solution, in "step S80", the step S80 of establishing a monitoring continuous decision-making operation mechanism includes the following steps: S801. Establish a continuous monitoring decision-making operation program, continuously run from step S20 to step S70 in a loop, and adjust the decision-making planning results in real time until the autonomous cruise global route navigation is completed.

[0014] Compared with the prior art, the present invention has the following beneficial effects: Under the premise of ensuring the safety and efficiency of ship navigation and complying with the International Regulations for Preventing Collisions at Sea, this invention explores the interactive collision avoidance behavior between autonomous ships and manned ships in complex and changing sea environments, in order to achieve rapid and reasonable collision avoidance decisions, thereby improving navigation safety and efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a schematic diagram of the overall process of the present invention; Figure 2 Schematic diagram of the process of step S10 of the present invention; Figure 3 Schematic diagram of the process of step S20 of the present invention; Figure 4 Schematic diagram of the process of step S30 of the present invention; Figure 5 Schematic diagram of the process of step S50 of the present invention; Figure 6 Schematic diagram of the process of step S60 of the present invention; Figure 7 Schematic diagram of the process of step S70 of the present invention; Figure 8 Schematic diagram of the process of step S80 of the present invention. DETAILED DESCRIPTION

[0016] The present invention will be further described below with reference to the embodiments.

[0017] The following examples are intended to illustrate the present invention but are not intended to limit the scope of protection of the present invention. The conditions in the examples may be further adjusted according to specific conditions. Simple improvements to the method of the present invention within the scope of the present invention are also within the scope of protection claimed in the present invention.

[0018] See also Figure 1-8 The present invention provides an intelligent decision-making method based on maritime collision avoidance rules, comprising the following steps: S10. Setting the global route of the ship's autonomous cruise mission; S20, obtaining ship positioning and environment perception prediction data; S30, obtaining a local reference trajectory of the ship through a local trajectory calculation module; S40, constructing a ship navigation encounter situation model; S50. Construct a multivariate composite collision avoidance risk assessment model; S60. Construct a behavioral decision tree model based on maritime collision avoidance rules; S70, the trajectory planning module performs trajectory planning based on the behavior decision tree model results; S80. Establish a continuous monitoring and decision-making operation mechanism.

[0019] In "Step S10", step S10 sets the global route of the ship autonomous cruise mission, including the following steps: S101, marking the cruise mission route of the intelligent ship on the electronic map in sequence, and recording the positions on the marked cruise mission route to obtain a point queue of the cruise mission route; S102: Convert the cruise mission route point queue into a Cartesian coordinate system through a latitude, longitude and Cartesian coordinate conversion module; S103, performing waypoint compensation on two adjacent points in the global route point queue in the Cartesian coordinate system using a sample point interpolation algorithm to obtain a desired point queue, and taking the desired point queue as the global route, that is, .

[0020] In "Step S20", step S20 obtains ship positioning and environment perception prediction data including the following steps: S201, the ship obtains the current GPS latitude and longitude in real time through its own global positioning system, recorded as , converted to the Cartesian coordinate system through the latitude and longitude & Cartesian coordinate system conversion module, recorded as ; S202: The ship obtains the current speed in real time through its own configured speed calculation module, which is recorded as ; S203, the ship obtains the current heading angle in real time through its own heading calculation module, which is recorded as ; S204: The ship obtains obstacle boundary data through the perception prediction module, including two types of data: One type is static obstacle boundary line data, which is composed of line geometry, specifically ; One type is obstacle data, which includes static obstacles and dynamic obstacles and is composed of polygonal geometry: ; The data on movement disorders are as follows: All obstacle boundaries at all times ; Obstacle speed , obstacle direction .

[0021] In "step S30", step S30 obtains the local reference trajectory of the ship through the local trajectory calculation module, including the following steps: S301, call the trajectory planning module to generate the current local reference path based on the global route. The local reference path is obtained by interpolation of the time dimension and is recorded as ,in, Represents a local reference path The coordinates of the first point at the moment are the current position of the ship. Represents a local reference path The coordinates of the sample point at the moment, Represents the coordinates of the end point of the local reference path.

[0022] In "step S40", step S40 constructs a ship navigation encounter situation model including the following steps: S401. The International Regulations for Preventing Collisions at Sea provide three encounter scenarios: head-on encounter, crossing encounter, and overtaking. However, in actual navigation, ships have different rights of way for different encounter situations, so it is necessary to classify the encounter scenarios in detail. S402. Based on the International Regulations for Preventing Collisions at Sea, a ship navigation encounter area model is established. Six encounter areas are divided according to the ship's position, namely, bow, starboard side ahead, starboard side behind, stern, port side behind, and port side ahead. With the hull position as the center, the parameters is the radius, the bow heading is arc 0, divided clockwise, For the bow range, It is the range just forward of the starboard side. It is the range behind the starboard side. The stern area, It is the range behind the port side. The coordinates of the boundary points of each encounter area are calculated using the arc area boundary point calculation tool: The coordinate set of the boundary points of the bow encounter area is: ; The coordinate set of the boundary points of the encounter area on the starboard side is: ; The coordinate set of the boundary points of the area encountered after the starboard side is: ; The coordinate set of the boundary points of the stern encounter area is: ; The coordinate set of the boundary points of the area encountered after the port side is: ; The coordinate set of the boundary points of the port side encounter area is: ; S403. Based on the ship navigation encounter area model, in each encounter area, a ship navigation encounter scenario model is set in combination with the International Regulations for Preventing Collisions at Sea, which is divided into: HO encounter scenario, CR crossing encounter scenario, OT overtaking scenario, and FO following scenario; Set own ship's heading angle to 、The ship's speed is , the heading angle of the target obstacle ship is and the speed is ; Calculate the absolute value of the heading angle difference between the own ship and the target obstacle ship ; Encounter scene The specific calculation formula of the model is as follows: , As shown in the above formula, when the angle difference between the main ship and the obstacle ship satisfies , is the HO encounter scenario; when the angle difference between the main ship and the obstacle ship meets , which is a CR intersection scenario; when the angle difference between the main ship and the obstacle ship meets , And the speed meets , which is the OT overtaking scenario; when the angle difference between the main ship and the obstacle ship meets , and the speed meets , for FO following scene.

[0023] In step S50, building a multivariate composite collision avoidance risk assessment model includes the following steps: S501, constructing a local reference path 2DBox detection frame set based on the local reference path data; Based on the local reference path data, constructing the local reference path 2DBox detection frame includes the following steps: S5011. Set the ship's safety detection shape (horizontal and vertical parameters). Generally, the length and width of the ship itself are increased by a certain safety buffer distance, which is recorded as ; S5012, calculate local reference path The orientation angle of the sample point at each moment t; Step Local Reference Path The orientation angle of the sample point at each time t includes the following steps: Select reference point: The time point is used as the position where the tangential angle will be calculated; Determine adjacent points: The time point is taken as the adjacent point of the reference point, i.e., the next point on the local reference path; Calculate the direction vector: Using adjacent coordinates, calculate the direction vector (vector), which represents the direction from the reference point to the next point. The calculation method of the direction vector is: ,in are the coordinates of the reference point, are the coordinates of adjacent points; calculate the orientation angle: use the inverse tangent function to calculate the tangent angle: ; For local reference paths Every sample point at time t Execute the above steps to get the local reference path towards the corner array: ; S5013, based on local reference path coordinate points and the local reference path heading angle , traverse the local path The sample point at the moment, based on the coordinates of the point , facing angle and ship safety shape , calculate the 2DBox bounding box coordinates corresponding to the sample point , including the following steps: Calculate the coordinates of each vertex of the bounding box separately: For the first corner (upper left corner) coordinates: , , For the second corner (upper right corner) coordinates: , , For the third corner (lower right corner) coordinates: , , For the fourth corner (lower left corner) coordinates: , , Therefore, the calculated sample points correspond to the bounding box coordinates: .

[0024] S5014, traverse the local reference path At the sample point at time t, execute the above steps to generate the 2DBox bounding box set corresponding to the local reference path: ; S502, constructing a DCPA spatial collision avoidance risk detection model; Steps: Constructing the DCPA spatial collision avoidance risk detection model includes the following steps: S5021, traverse the 2DBox detection frame generated at each time t along the local reference path , perform collision detection with all boundary lines and obstacle geometry data: The bounding box of the sample point is calculated by the line and polygon intersection calculation module and the boundary line whether they intersect; The bounding box of the sample point is calculated by the intersection calculation module of the two polygons Bounding boxes of various obstacles Do they intersect? If they are moving obstacles, only The moment bounding box performs intersection judgment; If they intersect, the position of this sample point is considered to be the nearest collision avoidance danger point in space and is recorded as , and then the sample point traversal ends.

[0025] S5022, based on local reference path The closest collision avoidance danger point with space Calculate the minimum collision avoidance distance in space: , in The nearest collision avoidance point The moment you are in, for The coordinates corresponding to the point, for Click the next time point The corresponding coordinates, That is the minimum space collision avoidance danger distance; If the local reference path is completed All sample points are traversed, and no spatial nearest collision danger point is generated , then set the minimum space collision avoidance danger distance to positive infinity; S5023. In combination with the International Regulations for Preventing Collisions at Sea, the smaller the spatial collision risk distance, the greater the spatial collision risk of the ship. A membership function for the spatial collision risk of the ship is constructed. for: ; Where: The minimum safe encounter distance parameter is usually 20 times the ship length. The cumulative s length parameter of the local reference path is usually the sensing range distance + 12 times the ship length.

[0026] S503, constructing a TCPA time collision avoidance risk detection model; Steps: Constructing the TCPA time collision avoidance risk detection model includes the following steps: S5031. Construct a detector for the closest collision risk time between a single moving obstacle and the vessel: Traversing the local reference path at each t moment to generate the 2DBox detection frame , and the bounding box of the moving obstacle at each moment t Perform collision detection: Calculate the bounding box of the sample point through the intersection calculation module of the two polygons Bounding box of moving obstacle at time t If they intersect, then the sample point t is considered to be the closest collision danger point between the ship and the moving obstacle, and is recorded as , and then the sample point traversal ends; If all sample points of the local reference path are traversed and no closest collision danger time point is generated, the closest collision danger time point between the ship and the moving obstacle is set to positive infinity.

[0027] S5032. Construct a detector for the encounter scene between a single moving obstacle and the vessel: If the closest collision avoidance danger time point of the moving obstacle is positive infinity after the closest collision avoidance danger time detection, it is considered that the moving obstacle has no encounter relationship with the ship; Determine the encounter area to which the dynamic obstacle belongs: Based on the steps to build the ship navigation encounter situation model, the boundary set of the ship encounter area is obtained: the bow encounter area , starboard side ahead of the meeting area 、The area you will encounter after being on the starboard side , stern encounter area , the area you will encounter after being on the port side and the port side abeam meeting area ; The calculation module of whether the point is inside the polygon calculates whether the coordinates of the moving obstacle position are within a certain encounter area; If the moving obstacle is inside a certain encounter area, then the encounter area is set as the encounter area to which the moving obstacle belongs for subsequent decision-making; If the dynamic obstacle is not within all the encounter areas, it is considered that the dynamic obstacle has no encounter relationship with the ship; Determine the encounter scenario to which the dynamic obstacle belongs: Based on the encounter scenario situation model obtained by constructing the ship navigation encounter situation model in the steps, calculate the encounter scenario to which the dynamic obstacle belongs. The encounter scenario result is one of the scenarios of encounter, crossing, overtaking and following.

[0028] S5033. Traverse all moving obstacle data to generate the nearest collision avoidance danger time, the encounter area, and the encounter scene for each moving obstacle; calculate the minimum collision avoidance danger time for all moving obstacles: ; In the formula Represents the latest collision avoidance danger time calculated for each moving obstacle; S5034. In combination with the International Regulations for Preventing Collisions at Sea, the smaller the minimum collision risk time, the greater the ship's time collision risk, and the ship's spatial collision risk membership function is constructed. for: ; Where: is the minimum safe encounter time parameter, is the parameter of the local reference path at the final state at time t.

[0029] S504, based on the membership function of the spatial collision avoidance risk model and membership function of time collision avoidance risk model , construct a multivariate composite collision avoidance risk assessment model, the specific formula is as follows: , Where, is the spatial collision avoidance risk weight, The result of space collision avoidance risk, is the time collision avoidance risk weight, is the time collision avoidance risk result, is the final collision avoidance risk detection result; ,The higher the collision avoidance risk, the more dangerous the current ship navigation state is; In "step S60", step S60 constructs a behavior decision tree model based on maritime collision avoidance rules, including the following steps: S601, constructing a behavior decision tree feature state; setting the decision tree feature state to: S6011: Is the situation in an emergency collision avoidance scenario? ; S6012: Whether to slow down to avoid the scene, expressed as: ; S6013, whether to overtake and bypass the scene, expressed as .

[0030] S602: Construct a behavior decision tree classification result; set the decision tree classification result to: S6021, the stationary decision result, indicates that the ship should decelerate to 0 at the maximum deceleration and remain stationary, recorded as ; S6022, the result of the avoidance decision, indicates that the ship needs to maintain its course and slow down to avoid, recorded as ; S6023, overtaking decision result, indicates that the ship needs to overtake and bypass at this time. According to the International Regulations for Preventing Collisions at Sea, when two power-driven vessels meet on opposite or nearly opposite courses and there is a risk of collision, each should turn right and pass the other vessel on the port side. Therefore, the overtaking and bypass should be to the right, recorded as ; S6024, the result of the passage decision, indicates that the ship maintains its course and speed and sails normally in the original channel, recorded as .

[0031] S603: Build a behavior decision tree classification model, set corresponding decision results under different feature states to form training data, and train the behavior decision tree classification model: S6031, set the training data feature state set to whether it is an emergency collision avoidance scene True, whether to slow down to avoid the scene Is it true and whether to overtake the bypass scenario True, the corresponding decision result is static ; S6032: Set the training data feature state set to whether it is an emergency collision avoidance scenario True, whether to slow down to avoid the scene Is it true and whether to overtake the bypass scenario If it is false, the corresponding decision result is static ; S6033: Set the training data feature state set to whether it is an emergency collision avoidance scenario True, whether to slow down to avoid the scene False and whether to overtake and bypass the scene True, the corresponding decision result is static ; S6034: Set the training data feature state set to whether it is an emergency collision avoidance scenario True, whether to slow down to avoid the scene False and whether to overtake and bypass the scene If it is false, the corresponding decision result is static ; S6035: Set the training data feature state set to whether it is an emergency collision avoidance scenario False, whether to slow down to avoid the scene Is it true and whether to overtake the bypass scenario If it is false, the corresponding decision result is avoid ; S6036: Set the training data feature state set to whether it is an emergency collision avoidance scenario False, whether to slow down to avoid the scene Is it true and whether to overtake the bypass scenario True, the corresponding decision result is avoid ; S6037, set the training data feature state set to whether it is an emergency collision avoidance scene False, whether to slow down to avoid the scene False and whether to overtake and bypass the scene If it is false, the decision result is pass ; S6038: Set the training data feature state set to whether it is an emergency collision avoidance scenario False, whether to slow down to avoid the scene False and whether to overtake and bypass the scene True, the decision result is overtaking .

[0032] S604, set the current data feature state set of the ship: S6041, Is it an emergency collision avoidance scenario? Characteristic state decision-making: The collision avoidance risk obtained from the step-based construction of a multivariate composite collision avoidance risk assessment model ; When the collision avoidance risk When the emergency collision avoidance scene is set The feature state is true, otherwise it is false; S6042, whether to slow down to avoid the scene Feature state decision-making: Based on the detection results of the ship's navigation encounter scenario model obtained from the multi-element composite collision avoidance risk assessment model constructed in the steps, combined with the International Regulations for Preventing Collisions at Sea, a ship on the starboard side of the ship should give way to the other ship; When the encounter area in front of the starboard side of the ship and the encounter area behind the starboard side have an encounter scenario relationship with other obstacle ships, set whether to slow down to avoid the scenario The feature state is true; In addition, when the bow of the ship encounters other obstacle ships in the area where they are following or crossing each other, the ship can set whether to slow down and avoid the scene. The feature state is true; In addition, when there is an overtaking encounter scenario in the stern encounter area of ​​the ship, set whether to slow down to avoid the scenario The feature state is true; In actual navigation practice, in order to improve safety, when the encounter area on the port side of the ship is in a situation where there is a high-speed obstacle vessel (here the high-speed vessel can be judged by setting a certain speed threshold in practice, if the obstacle vessel is greater than this threshold, it is considered to be a high-speed obstacle vessel), the speed reduction or avoidance scenario is set. The feature state is true; If none of the above decision conditions are met, set whether to slow down and avoid the scene The feature status is false; S6043, Overtaking and detouring scenario Characteristic state decision-making: The collision avoidance risk obtained from the step-based construction of a multivariate composite collision avoidance risk assessment model and the detection results of the situation model of the scenarios that the ship may encounter during navigation; When the collision avoidance risk When the ship's various encounter areas have no dynamic obstacle encounter scene relationship, that is, only static obstacle avoidance scenes, or the ship's bow encounter area has a collision or overtaking scene relationship with other obstacle ships, then set whether to overtake and bypass the scene The feature state is true, otherwise it is false; S605. Based on the current data feature state set of the ship: , The behavior decision tree classification model predicts the current decision tree result of this ship, which should be: One of the formulas; In step S70, the trajectory planning module in step S70 performs trajectory planning based on the behavior decision tree model results, including the following steps: S701, if the decision result of the behavior decision tree model is , set the current maximum speed limit of the ship to 0, and plan the stop trajectory with the ship's maximum deceleration parameter through the trajectory planning module; S702, if the decision result of the behavior decision tree model is , based on the steps to construct the spatial closest collision avoidance risk distance obtained in the multivariate composite collision avoidance risk assessment model Minimum collision avoidance time , calculate the maximum speed limit of the ship to avoid, including the following steps: S7021, based on the closest collision avoidance distance in space Calculate the maximum speed limit to avoid : , where is the maximum deceleration parameter of the ship; S7022, based on the minimum collision avoidance time Calculate the maximum speed limit to avoid : , Where, The final speed is set to 0, is the maximum deceleration parameter of the ship; S7023, based on the maximum speed limit of spatial collision avoidance risk and the maximum speed limit of temporal collision avoidance risk, set the maximum speed limit under the avoidance decision result ; The trajectory planning module plans the avoidance trajectory using the ship's maximum deceleration parameter; S703, if the decision result of the behavior decision tree model is , restore the maximum speed parameters of the ship itself, and plan a right-circling and overtaking trajectory through the trajectory planning module; S704, if the decision result of the behavior decision tree model is , restore the maximum speed parameters of the ship itself, and maintain the original channel trajectory through the trajectory planning module; S705: Send the trajectory planning module results to the ship control module for execution.

[0033] In step S80, establishing a continuous monitoring decision-making operation mechanism includes the following steps: S801. Establish a continuous monitoring decision-making operation program, continuously run from step S20 to step S70 in a loop, and adjust the decision-making planning results in real time until the autonomous cruise global route navigation is completed.

[0034] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent decision-making method based on maritime collision avoidance regulations, characterized by: The following steps are involved: S10. Setting the global route of the ship's autonomous cruise mission; S20, obtaining ship positioning and environment perception prediction data; S30, obtaining a local reference trajectory of the ship through a local trajectory calculation module; S40, constructing a ship navigation encounter situation model; S50. Construct a multivariate composite collision avoidance risk assessment model; S60. Construct a behavioral decision tree model based on maritime collision avoidance rules; S70, the trajectory planning module performs trajectory planning based on the behavior decision tree model results; S80. Establish a continuous monitoring and decision-making operation mechanism.

2. The intelligent decision-making method based on maritime collision avoidance regulations according to claim 1, characterized in that: In "step S10", the step S10 setting the global route of the ship autonomous cruise mission includes the following steps: S101, marking the cruise mission route of the intelligent ship on the electronic map in sequence, and recording the positions on the marked cruise mission route to obtain a point queue of the cruise mission route; S102: Convert the cruise mission route point queue into a Cartesian coordinate system through a latitude, longitude and Cartesian coordinate conversion module; S103 , performing waypoint compensation on two adjacent points in the global route point queue in the Cartesian coordinate system using a sample point interpolation algorithm to obtain a desired point queue, and using the desired point queue as the global route.

3. The intelligent decision-making method based on maritime collision avoidance regulations according to claim 1, characterized in that: In "step S20", the step S20 of obtaining ship positioning and environment perception prediction data includes the following steps: S201. The vessel obtains the current GPS longitude and latitude in real time through its own global positioning system, and converts it into a Cartesian coordinate system through a latitude and longitude & Cartesian coordinate system conversion module; S202: The vessel obtains its current speed in real time through its own configured speed calculation module; S203, the vessel obtains the current heading angle in real time through its own heading calculation module; S204: The ship obtains obstacle boundary data through the perception prediction module.

4. The intelligent decision-making method based on maritime collision avoidance regulations according to claim 1, characterized in that: In "step S30", the step S30 obtains the local reference trajectory of the ship through the local trajectory calculation module, including the following steps: S301: Calling a trajectory planning module to generate a current local reference path based on the global route.

5. The intelligent decision-making method based on maritime collision avoidance regulations according to claim 1, characterized in that: In "step S40", the step S40 constructs a ship navigation encounter situation model including the following steps: S401. The International Regulations for Preventing Collisions at Sea provide three types of encounter scenarios: head-on encounter, crossing encounter, and overtaking. However, in actual navigation, ships have different rights of way for different encounter situations, so it is necessary to classify the encounter scenarios in detail. S402. Based on the International Regulations for Preventing Collisions at Sea, a model of ship navigation encounter areas is established. Six encounter areas are divided based on the ship's orientation: bow, forward of starboard beam, aft of starboard beam, stern, aft of port beam, and forward of port beam. The areas are divided clockwise, with the ship's hull position as the center and the bow heading as arc 0. The coordinates of the boundary points of each encounter area are calculated using the arc area boundary point calculation tool. S403: Based on the ship navigation encounter area model, in each encounter area, in combination with the International Regulations for Preventing Collisions at Sea, a ship navigation encounter scenario situation model is set.

6. The intelligent decision-making method based on maritime collision avoidance regulations according to claim 1, characterized in that: In "step S50", the step S50 constructs a multivariate composite collision avoidance risk assessment model, including the following steps: S501, constructing a local reference path 2DBox detection frame set based on the local reference path data; S502, constructing a DCPA spatial collision avoidance risk detection model; S503: Construct a TCPA time collision avoidance risk detection model.

7. The intelligent decision-making method based on maritime collision avoidance regulations according to claim 1, characterized in that: In "step S60", the step S60 constructs a behavior decision tree model based on maritime collision avoidance rules, including the following steps: S601, constructing a behavior decision tree feature state; S602, constructing a behavior decision tree classification result; S603: Build a behavior decision tree classification model, set corresponding decision results under different feature states to form training data, and train the behavior decision tree classification model; S604, setting the current data feature state set of the ship; S605: Based on the current data feature state set of the ship, predict the current decision tree result of the ship through the behavior decision tree classification model.

8. The intelligent decision-making method based on maritime collision avoidance regulations according to claim 1, characterized in that: In step S70, the trajectory planning module in step S70 performs trajectory planning based on the behavior decision tree model result, including the following steps: S701, if the decision result of the behavior decision tree model is , set the current maximum speed limit of the ship to 0, and plan the stop trajectory with the ship's maximum deceleration parameter through the trajectory planning module; S702, if the decision result of the behavior decision tree model is , based on the spatial closest collision risk distance and minimum collision avoidance time obtained in the multivariate composite collision avoidance risk assessment model constructed in the steps, the maximum speed boundary of the ship's avoidance is calculated; S703, if the decision result of the behavior decision tree model is , restore the maximum speed parameters of the ship itself, and plan a right-circling and overtaking trajectory through the trajectory planning module; S704, if the decision result of the behavior decision tree model is , restore the maximum speed parameters of the ship itself, and maintain the original channel trajectory through the trajectory planning module; S705: Send the trajectory planning module results to the ship control module for execution.

9. The intelligent decision-making method based on maritime collision avoidance regulations according to claim 1, characterized in that: In "step S80", the step S80 establishes a monitoring continuous decision-making operation mechanism including the following steps: S801. Establish a continuous monitoring decision-making operation program, continuously run from step S20 to step S70 in a loop, and adjust the decision-making planning results in real time until the autonomous cruise global route navigation is completed.

Images