A distributed autonomous collaborative collision avoidance method for multiple intelligent ships

By combining the optimal interactive collision avoidance algorithm with real ship navigation constraints, the distributed autonomous collaborative collision avoidance of intelligent ships is achieved, and the problems of real-time collaborative collision avoidance of large-scale ships and high communication costs are solved, reducing collision risks and communication costs. .

CN115909817BActive Publication Date: 2024-11-12DALIAN MARITIME UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202211305715.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-10-24
Publication Date
2024-11-12
Estimated Expiration
2042-10-24

AI Technical Summary

Technical Problem

The existing intelligent ship collision avoidance technology is not suitable for real-time collaborative collision avoidance scenarios for large-scale ships, and communication and coordination are required during the coordination process, resulting in high computing and communication costs.

Method used

The optimal interactive collision avoidance (ORCA) algorithm is used and the real ship navigation constraints to achieve independent and autonomous distributed collaborative collision avoidance based on limited perception information, and there is no need for communication and coordination between ships.

Benefits of technology

It reduces the risk of collision between ships, saves communication costs, and realizes real-time collaborative collision avoidance in large-scale intelligent ship scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115909817B_ABST
    Figure CN115909817B_ABST
Patent Text Reader

Abstract

The present invention provides a distributed autonomous collaborative collision avoidance method for multiple intelligent ships. It mainly includes: constructing a dangerous ship list of the ship based on a ship collision risk model; judging whether the dangerous ship list of the ship is empty; dividing the encounter situation according to the navigation information of the ship and the target ship; determining the collision avoidance responsibility coefficient of each ship according to the encounter situation; constructing an interactive collision avoidance space between the ship and all dangerous ships based on the ORCA algorithm; selecting the speed vector with the smallest deviation from the current heading as the optimal speed vector; clearing the dangerous ship list. The present invention can continuously perceive, make decisions and respond to the traffic situation around the ship, so that each intelligent ship can independently and autonomously perform distributed collaborative collision avoidance based on limited perception information, and there is no need for communication and coordination between ships, which not only reduces the risk of collision between ships, but also saves the communication cost between ships.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of smart shipping technology, and in particular to a distributed autonomous collaborative collision avoidance method for multiple smart ships. Background Art

[0002] With the continuous advancement of science and technology and the continuous growth of world maritime trade, intelligent operation methods are gradually showing great potential. In recent years, intelligent ship technology with "autonomous navigation" as the core element has developed rapidly. Intelligent ships use advanced perception, cognition, decision-making and control technologies. Compared with traditional ships, they have significant advantages in economy, safety, reliability, environmental protection and efficiency. The ability of intelligent ships to make autonomous collision avoidance decisions is the key to ensuring their safe navigation. How to coordinate the collision avoidance behaviors between intelligent ships to meet the real-time collaborative collision avoidance needs between intelligent ships in dynamic scenes is the focus of research on collision avoidance decision-making technology for intelligent ships today.

[0003] In order to solve the problem of intelligent ship collision avoidance, domestic and foreign scholars have proposed many methods, representative of which are particle swarm-based methods and fuzzy logic-based methods. These methods have shown good performance for the problem of intelligent ship collision avoidance in simple scenarios. However, due to the limitations of model complexity and computational efficiency, the collision avoidance decisions of each intelligent ship sometimes conflict with each other, making it difficult to apply them to highly dynamic and high-density unmanned ship cluster scenarios. In addition, the above methods assume that the motion state of the target ship remains unchanged. This assumption is often not true in actual ship navigation, because once the target ship perceives that other ships pose a collision risk to it, it will normally take corresponding collision avoidance measures.

[0004] In addition, most of the current intelligent ship collaborative collision avoidance technologies use a centralized approach to make unified decisions, that is, it is assumed that there is a coordination center to formulate corresponding collision avoidance routes for each ship, and only when all ships sail according to their own collision avoidance routes can collisions be avoided. However, even in a situation where multiple ships meet, this method only makes one decision to avoid collisions, ignoring the autonomy of the ship, the real-time nature of the collision avoidance decision, and the dynamic nature of the traffic situation around the ship. It is not suitable for large-scale intelligent ships to avoid collisions, such as narrow areas with dense traffic such as bays and ports. The actual situation of intelligent ships making collision avoidance decisions should be that each intelligent ship makes the best plan it thinks in real time according to the traffic situation of the surrounding ships, so as to respond quickly to the behavior of the target ship.

[0005] Currently, there have been some related studies on distributed autonomous collaborative collision avoidance decision-making technology for intelligent ships, but most of these studies require communication and coordination between ships to exchange each other's avoidance intentions, which will consume a lot of computing and communication costs.

[0006] For example, in the case of multi-ship encounters, the distributed collision avoidance decision method can handle the situation where all ships comply with the collision avoidance rules and the situation where the ship that gives way does not cooperate. However, one problem with this method is that the collision avoidance decisions between different ships may conflict, and the ships need to communicate and cooperate to avoid conflicts.

[0007] Another example is the multi-agent based ship collision avoidance method. With multi-agent as the architecture, a multi-agent based ship collision avoidance system is constructed by combining ship communication methods and ship collision avoidance knowledge. However, ships need to negotiate to avoid collisions.

[0008] Another example is the ship collision avoidance method based on distributed collaborative strategy, which is mainly aimed at collision avoidance planning between multiple ships in open waters. First, the ant lion algorithm is used to plan the best collision avoidance route for the ship relative to each encountering ship; secondly, the ship sends each best collision avoidance route to the encountering ship; then, the ship uses the improvement index formula to calculate the improvement index of the ship and the encountering ship respectively, and the ship with a large improvement is used as the avoidance ship. The avoidance ship sails according to its corresponding best collision avoidance route, and other ships sail according to the original route. However, in this method, communication and coordination are also required between the ship and the encountering ship to avoid collision.

[0009] In summary, the current intelligent ship collision avoidance technology is not suitable for scenarios where large-scale ships conduct real-time collaborative collision avoidance. At the same time, the existing collaborative collision avoidance process of ships requires communication and coordination, resulting in high computing and communication costs. Summary of the invention

[0010] The present invention proposes a distributed autonomous collaborative collision avoidance method for multiple intelligent ships, which combines the Optimal Reciprocal Collision Avoidance (ORCA) algorithm with real ship navigation constraints. It can continuously perceive, make decisions and respond to the traffic situation around the ship, so that each intelligent ship can independently and autonomously perform distributed collaborative collision avoidance based on limited perception information (heading, speed, position), and there is no need for communication and coordination between ships. It not only reduces the risk of collision between ships, but also saves the communication cost between ships.

[0011] The technical means adopted by the present invention are as follows:

[0012] A distributed autonomous collaborative collision avoidance method for multiple intelligent ships comprises the following steps:

[0013] S1. Obtain navigation information of the own ship and all target ships in the visible domain, wherein the navigation information includes position, heading and speed;

[0014] S2. Calculate the expected velocity vector of the ship based on the ship's position and the terminal position;

[0015] S3, calculating the minimum encounter distance, minimum encounter time, ship safety zone and relative position between the own ship and the target ship, and constructing a dangerous ship list for the own ship based on the ship collision risk model;

[0016] S4, judging whether the dangerous ship list of the own ship is empty, if the dangerous ship list of the own ship is empty, executing S5, otherwise executing S6;

[0017] S5, determine whether the ship has reached the destination. If the ship has reached the destination, end the voyage; otherwise, continue to sail with the current expected speed vector and execute S1 after a decision cycle.

[0018] S6, determining whether the number of ships in the dangerous ship list of the own ship is greater than 1, if the number of ships in the list is equal to 1, executing S7, otherwise executing S9;

[0019] S7. Based on the navigation information of own ship and target ship and referring to the International Regulations for Preventing Collisions at Sea, the encounter situation is divided;

[0020] S8. Determine the collision avoidance responsibility coefficient of each ship according to the encounter situation, and execute S10;

[0021] S9. Determine that each vessel bears the same avoidance responsibility coefficient;

[0022] S10, obtaining the decision cycle of each decision made by the intelligent ship;

[0023] S11. The ship safety domain, decision cycle and avoidance responsibility coefficient are used as inputs of the ORCA algorithm to construct the interactive collision avoidance space between the ship and all dangerous ships.

[0024] S12, the ship selects the speed vector with the smallest deviation from the current course as the optimal speed vector for navigation through linear programming;

[0025] S13. Clear the list of dangerous ships and execute S1 after one decision cycle.

[0026] Further, the expected velocity vector of the ship is calculated according to the position of the ship and the terminal position, including obtaining the expected velocity vector of the ship according to the following formula:

[0027]

[0028] Among them, speed OS is the ship's speed, is the current position of the ship, The terminal position of this ship.

[0029] Furthermore, the minimum encounter distance between the own ship and the target ship is calculated according to the following formula:

[0030] DCPA=D×sin(c OT -θ-Π)

[0031]

[0032]

[0033] Where D is the distance between the own ship and the target ship, θ is the true azimuth of the target ship, and v oT is the relative speed between own ship and target ship, v0 is own ship’s speed, c0 is own ship’s heading;

[0034] The minimum encounter time between own ship and target ship is calculated according to the following formula:

[0035]

[0036] Furthermore, a dangerous ship list of the own ship is constructed based on the ship collision risk model, including determining whether to add the target ship to the dangerous ship list according to the following methods:

[0037]

[0038] Among them, d s It is the ship safety zone.

[0039] Furthermore, according to the navigation information of the own ship and the target ship, the encounter situation is divided with reference to the International Regulations for Preventing Collisions at Sea, including dividing the encounter situation according to the following methods:

[0040] The encounter situation that satisfies the following formula is defined as an encounter:

[0041]

[0042] The encounter situation that satisfies the following formula is defined as a starboard crossing encounter:

[0043]

[0044] The encounter situation that satisfies the following formula is defined as overtaking:

[0045]

[0046] The encounter situation that satisfies the following formula is defined as a port crossing encounter:

[0047]

[0048] Where B is the relative position between own ship and target ship.

[0049] Furthermore, the ship safety domain, decision cycle, and avoidance responsibility coefficient are used as inputs of the ORCA algorithm to construct the interactive collision avoidance space between the ship and all dangerous ships, including:

[0050] The speed obstacle area formed by the target ship to the own ship within the time window τ is obtained according to the following formula:

[0051]

[0052] where p TS 、p OS represents the position of the target ship and own ship respectively, d s|TS ,d s|OS represents the safety zone between the target ship and own ship, and

[0053] D(p TS -p OS ,d s|OS +d s|TS )={q|||q-(p TS -p OS )||<(d s|OS +d s|TS )}

[0054] if It means that the ship and the target ship will collide within the time window τ. This ensures that the own ship and the target ship can travel safely without collision before the time window τ;

[0055] If the target ship's speed changes in space v TS ∈V TS , then the set of own ship velocity vectors satisfying the following formula is defined as the collision avoidance velocity vector space:

[0056]

[0057]

[0058] in Represents With V TS of Minkowski and;

[0059] For the target ship, its collision avoidance velocity vector space relative to the own ship is: where v OS ∈V OS , when the ship's velocity vector And the target ship's velocity vector When τ is equal to τ, the ship and the target ship can be guaranteed to pass safely within τ time;

[0060] For the ships that meet, the velocity vector space that satisfies the following formula is defined as the optimal interactive collision avoidance space:

[0061]

[0062] Assume that the own ship and the target ship have speeds navigation, and this will cause the own ship to collide with the target ship within the τ time window, that is, Define λ as The distance between the starting point and the end point The nearest vector to the region is defined as follows:

[0063]

[0064] The set composition in velocity space is shown in the following formula:

[0065]

[0066] Among them, ω os is the avoidance responsibility coefficient, n is the speed obstacle area Located at point The outward normal vector at , λ is the minimum change in speed between the own ship and the target ship to avoid collision within the time window τ.

[0067] Compared with the prior art, the present invention has the following advantages:

[0068] 1. The present invention proposes a distributed autonomous collaborative collision avoidance method for multiple intelligent ships to solve the problem that the current intelligent ship collision avoidance technology is not suitable for large-scale ships to perform real-time collaborative collision avoidance and that the communication and computation costs are consumed during the ship collaboration process due to the need for communication and coordination. Even under limited communication conditions, each intelligent ship can independently and autonomously perform distributed collaborative collision avoidance based on limited observable information.

[0069] 2. The present invention combines the optimal interactive collision avoidance algorithm with real ship navigation constraints, and can continuously perceive, make decisions and respond to the traffic situation around the ship, so that each intelligent ship can independently perform distributed collaborative collision avoidance based on limited perception information (heading, speed, position), and there is no need for communication and coordination between ships, which not only reduces the risk of collision between ships, but also saves communication costs between ships. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0071] Figure 1 It is a flow chart of the distributed autonomous collaborative collision avoidance method for multiple intelligent ships in the embodiment.

[0072] Figure 2 Schematic diagram of relevant parameters of the own ship and the target ship in the embodiment.

[0073] Figure 3 Schematic diagram of the encounter situation between two ships in the embodiment.

[0074] Figure 4 Schematic diagram of relevant parameters of the ORCA algorithm in the embodiment.

[0075] Figure 5 Schematic diagram of the collision avoidance route of two ships in the embodiment.

[0076] Figure 6 Graph showing the distance variation between the two ships in the embodiment.

[0077] Figure 7 It is the collision avoidance route map of four ships in the embodiment.

[0078] Figure 8 : is a distance variation curve diagram between Ship1 and other ships in the embodiment.

[0079] Fig. 9 : is a distance variation curve diagram between Ship2 and other ships in the embodiment.

[0080] Fig.10 : is a distance variation curve diagram between Ship3 and other ships in the embodiment.

[0081] Fig.11 4 is a distance variation curve diagram between Ship4 and other ships in the embodiment. DETAILED DESCRIPTION

[0082] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0083] Since most of the current ships use unmanned ship swarms to perform tasks, each unmanned ship can be regarded as an intelligent entity with the ability to make autonomous collision avoidance decisions, and a distributed approach can be used to solve the problem of collaborative collision avoidance of unmanned ship swarms. In addition, the Optimal Interactive Collision Avoidance (ORCA) algorithm is a classic dynamic local multi-agent obstacle avoidance algorithm with a time complexity of O(n). Its calculation method is not only distributed, but the calculation process is based on mathematical geometric models and has a faster execution speed. It is very suitable for scenarios where large-scale intelligent entities perform collaborative collision avoidance. Therefore, the present invention provides a distributed autonomous collaborative collision avoidance method for multiple intelligent ships. This method regards each ship as an intelligent ship, that is, each ship has the ability to make autonomous collaborative collision avoidance decisions, and each intelligent ship can be used as "this ship." The collision avoidance process of each intelligent ship is as follows: Figure 1 As shown, the basic steps are as follows:

[0084] Step 1: Obtain the navigation information of own ship and all target ships in the visible domain.

[0085] Through AIS ship automatic identification system, ARPA radar, GPS positioning, electronic compass and other professional equipment, the navigation information of the ship and all target ships in the visible domain, including position, heading and speed, is obtained. The range of the visible domain is 6 nautical miles from the ship.

[0086] Step 2: Calculate the desired velocity vector of own ship.

[0087] remember is the desired velocity vector of the ship, which is expressed as the unit vector from the current position to the terminal position multiplied by the current speed. The calculation formula is shown in formula (1).

[0088]

[0089] Where speed OS is the ship's speed, is the current position of the ship, The terminal position of this ship.

[0090] Step 3: Construct a list of dangerous ships for the ship based on the ship collision risk model.

[0091] The ship and the ships in the dangerous ship list are all intelligent ships with autonomous collaborative collision avoidance decision-making capabilities, which work together to avoid collisions in a distributed manner, and there is no communication between the ships. According to the position, heading, and speed information of the ship and the target ship obtained in step 1, the minimum encounter distance DCPA, the minimum encounter time TCPA, and the ship safety domain d s, relative position B. Assume that the coordinates of the own ship S0 are (x0, y0), the speed and heading are v0, c0 respectively; the target ship S T The coordinates of (x T ,y T ), the speed and heading are v T ,c T The relative speed between the own ship and the target ship is v oT , and its component calculation formula on the x and y axes is shown in formula (2).

[0092]

[0093] The relative velocity direction is c OT The calculation formula of is shown in formula (3).

[0094]

[0095] θ is the true azimuth of the target ship, D is the distance between the own ship and the target ship, and the calculation formula of DCPA is shown in formula (4).

[0096] DCPA=D×sin(c OT -θ-Π) (4)

[0097] The calculation formula of TCPA is shown in formula (5).

[0098]

[0099] If the DCPA and TCPA of the own ship and the target ship reach a certain threshold, it means that there is a risk of collision between the own ship and the target ship, and the target ship is added to the list of dangerous ships. The collision risk judgment formula is shown in formula (6).

[0100]

[0101] where d s is the ship safety zone, and the calculation formula is shown in formula (7).

[0102]

[0103] An intuitive schematic diagram of the above related parameters, such as Figure 2 shown.

[0104] Step 4: Determine whether the list of dangerous ships is empty.

[0105] Each smart ship has a dangerous ship list, which contains ships that pose a collision risk to the ship. If the list is empty, it means that there is no ship in the visible domain that poses a danger to the ship, and go to step 5; otherwise, go to step 6.

[0106] Step 5: Determine whether the ship has reached the destination.

[0107] If the ship reaches the destination, the voyage ends; otherwise, the ship continues to sail with the current expected velocity vector, and after a decision cycle, it turns to step 1. The formula for determining whether the ship has reached the destination is shown in formula (8).

[0108]

[0109] in is the ordinate of the end point, The vertical coordinate of the current position.

[0110] Step 6: Determine whether the number of ships in the list is greater than 1.

[0111] If the number of ships in the list is equal to 1, it means that there are two intelligent ships in the visible domain that are performing collaborative collision avoidance, and the process goes to step 7; otherwise, the process goes to step 9.

[0112] Step 7: Divide the encounter situation.

[0113] If the number of ships in the dangerous ship list is 1, it means that there are two smart ships in the visible domain to avoid collision in coordination. At this time, the encounter situation should be divided according to the navigation information of the own ship and the target ship, referring to the International Regulations for Preventing Collisions at Sea, so as to determine the collision avoidance responsibility coefficient of each ship. According to the provisions of the collision avoidance regulations, the encounter situation between the two ships is divided into: encounter (A), starboard crossing encounter (B), overtaking (C), port crossing encounter (D), such as Figure 3 shown.

[0114] The encounter situation that satisfies formula (9) is defined as head-on.

[0115]

[0116] The encounter situation that satisfies formula (10) is defined as a starboard crossing encounter (CROSS_RIGHT).

[0117]

[0118] The encounter situation that satisfies formula (11) is defined as overtaking (OVER_TAKEN).

[0119]

[0120] The encounter situation that satisfies formula (12) is defined as a port crossing encounter (CROSS_LEFT).

[0121]

[0122] Where B is the relative position between the own ship and the target ship.

[0123] Step 8: Determine the avoidance responsibility coefficients of the two ships with reference to the collision avoidance rules.

[0124] According to the International Regulations for Preventing Collisions at Sea, the principle for determining the avoidance responsibility when two ships meet is defined as the avoidance responsibility coefficient, and its specific values ​​are shown in Table 1.

[0125] Table 1 Responsibility coefficients of two ships participating in coordinated collision avoidance

[0126]

[0127] Step 9: Each vessel bears the same avoidance responsibility coefficient.

[0128] If the number of ships in the dangerous ship list is greater than 1, it means that at least 3 smart ships are coordinating to avoid collision in the visible domain. However, the International Regulations for Preventing Collisions at Sea do not provide specific guidance on the determination of the cooperative avoidance responsibility coefficient when multiple smart ships meet. The present invention proposes that when multiple smart ships meet, each ship bears the same avoidance responsibility coefficient, i.e. Smart ships participating in collaborative collision avoidance should take efficient avoidance actions as their first goal, that is, they have the responsibility and obligation to participate in the collaborative collision avoidance process in dangerous situations and resolve the collision danger as soon as possible. For any two smart ships, each party bears half of the avoidance responsibility, which ensures that in the absence of communication, each ship can independently make distributed collaborative collision avoidance decisions based on observation information.

[0129] Step 10: Determine the decision cycle.

[0130] The decision cycle refers to the time interval between each decision made by the intelligent ship. In order to be able to respond quickly to the traffic situation around the intelligent ship, taking into account the existence of decision errors and action errors, an appropriate amount of safety should be retained when taking avoidance actions to put the ship in an active position. Therefore, when conducting simulation experiments, the present invention sets the decision cycle τ to 30s, and its value can be set according to specific needs.

[0131] Step 11: Construct a safe and feasible speed vector space for the ship based on the ORCA algorithm.

[0132] The ORCA algorithm is a dynamic local multi-agent obstacle avoidance algorithm. Its core principle is that both agents on the encountering side have the ability to make collision avoidance decisions and have the same collision avoidance responsibility and collision avoidance decision-making mechanism when eliminating collision risks. This method makes collision avoidance decisions for other agents based on the position, speed and size information of other agents. It does not require communication between agents and can respond quickly to the behavior of other agents. The ship's safety domain, decision cycle and avoidance responsibility coefficient are used as inputs to the ORCA algorithm to construct an interactive collision avoidance space between the ship and all dangerous ships. is the speed obstacle area formed by the target ship to the own ship within the time window τ, and the calculation formula is shown in formula (13).

[0133]

[0134] where p TS 、p OS represents the position of the target ship and own ship respectively, d s|TS ,d s|OS represents the safety zone between the target ship and the own ship, D(p TS -p OS ,d s|OS +d s|TS ) is expressed as formula (14).

[0135] D(p TS -p OS ,d s|OS +d s|TS )={q|||q-(p TS -p OS )||<(d s|OS +d s|TS )} (14)

[0136] if It means that the ship and the target ship will collide within the time window τ; on the contrary, if This ensures that the ship and the target ship can travel safely without collision before the time window τ. TS ∈V TS , the set of own ship velocity vectors satisfying formula (15) is defined as the collision avoidance velocity vector space.

[0137]

[0138] in Represents With V TS The Minkowski sum of is defined as shown in formula (16).

[0139]

[0140] Similarly, for the target ship, its collision avoidance velocity vector space relative to the own ship is where v OS ∈V OS , when the ship's velocity vector And the target ship's velocity vector When , it can be ensured that the ship and the target ship pass safely within the time τ without collision.

[0141] For the encountering ships, the velocity vector space satisfying formula (17) is defined as the optimal interactive collision avoidance space.

[0142]

[0143] Assume that the own ship and the target ship have speeds navigation, and this will cause the own ship to collide with the target ship within the τ time window, that is, Define λ as The distance between the starting point and the end point The vector closest to the region is defined as shown in formula (18).

[0144]

[0145] The set construction method in the velocity space is shown in formula (19).

[0146]

[0147] Among them, ω os is the avoidance responsibility coefficient determined in step 2. When there are only two ships participating in the coordinated collision avoidance in the visible domain, its specific value is referenced to Table 1. When there are more than two ships participating in the coordinated collision avoidance in the visible domain, the present invention proposes that the intelligent ships participating in the coordinated collision avoidance bear the same avoidance responsibility, that is, all ships need to take actions to avoid collision. The avoidance responsibility coefficient ω of each intelligent ship participating in the coordinated collision avoidance os The value is 0.5. Among them, n is the speed obstacle area Located at point λ is the minimum change in speed between the own ship and the target ship to avoid collision within the time window τ. An intuitive schematic diagram of the relevant parameters of the ORCA algorithm is shown in Figure 4 shown.

[0148] Step 12: Select the optimal speed vector navigation.

[0149] According to the advocacy of steering maneuvers by the International Regulations for Preventing Collisions at Sea, the present invention does not adopt the method of reducing speed when selecting the optimal speed vector for collision avoidance decision-making, that is, only steering is allowed. When making collision avoidance decisions, the intelligent ship selects the speed vector with the smallest deviation from the current course through linear programming. The allowed speed set of the ship is the intersection of the half plane formed with all target ships, and the calculation formula is shown in formula (20).

[0150]

[0151] The calculation formula for the optimal collision avoidance velocity vector of the ship is shown in formula (21).

[0152]

[0153] Step 13: Clear the list of dangerous vessels.

[0154] Clear the list of dangerous ships in this decision cycle of the ship, and go to step 1 after one decision cycle.

[0155] From the above description, it can be seen that the present invention proposes a distributed autonomous collaborative collision avoidance method for multiple intelligent ships. Each intelligent ship can autonomously perform continuous perception, decision-making and response according to the decision-making cycle to form a closed loop, and collaboratively avoid the occurrence of collisions in a distributed manner, and there is no communication between the ships. At the same time, each intelligent ship takes the efficiency of avoidance actions as the goal, and assumes different avoidance responsibility coefficients in different encounter situations. The time complexity of this method is O(n), which is determined by the total number of dangerous ships at the current decision-making moment. It has a high response speed and is suitable for large-scale intelligent ships to make collaborative collision avoidance decisions.

[0156] In order to verify the effectiveness and reliability of the solution of the present invention, the following specific implementation cases are simulated and verified using MATLAB software. The first case is two intelligent ships cooperatively avoiding collisions, and the second case is four intelligent ships cooperatively avoiding collisions.

[0157] (I) Simulation verification and analysis of collaborative collision avoidance between two intelligent ships

[0158] According to the scheme of the present invention, the coordinated collision avoidance behavior between the two ships should be determined in accordance with the International Regulations for Preventing Collisions at Sea. The effectiveness of the scheme is verified by the right crossing situation below. The relevant parameter settings of the ships are shown in Table 2. Ship 1 and Ship 2 form a right crossing situation. According to the collision avoidance rules, it can be seen that Ship 1 is a straight-line ship, maintaining direction and speed; Ship 2 is a way-giving ship, taking evasive action. The specific collision avoidance route planned by the present invention is as follows Figure 5 As shown in the figure, the distance variation curve between the two ships is as follows: Figure 6 As shown in the figure, we can see that the two ships passed safely and their avoidance behavior complied with the provisions of the International Regulations for Preventing Collisions at Sea.

[0159] Table 2 Initial state parameters of the two ships

[0160]

[0161] (II) Simulation verification and analysis of four intelligent ships’ collaborative collision avoidance

[0162] According to the scheme of the present invention, when there are more than two smart ships participating in the collaborative collision avoidance, the avoidance responsibility coefficient of each ship is 0.5, that is, each ship needs to take action. The effectiveness of the scheme is verified by experiments below. The relevant parameter settings of the ship are shown in Table 3, and the specific collision avoidance route is shown in Figure 7 As shown in Figure 2, the distance variation curve between ships is as follows: Figure 8 to Figure 11 As shown in the figure, it can be seen that the ships pass each other safely, which verifies the effectiveness of this scheme.

[0163] Table 3 Initial state parameters of the four ships

[0164]

[0165] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A distributed autonomous collaborative collision avoidance method for multiple intelligent ships, characterized in that: The following steps are involved: S1. Obtain navigation information of the own ship and all target ships in the visible domain, wherein the navigation information includes position, heading and speed; S2. Calculate the expected velocity vector of the ship based on the ship's position and the terminal position; S3, calculating the minimum encounter distance, minimum encounter time, ship safety zone and relative position between the own ship and the target ship, and constructing a dangerous ship list for the own ship based on the ship collision risk model; S4, judging whether the dangerous ship list of the own ship is empty, if the dangerous ship list of the own ship is empty, executing S5, otherwise executing S6; S5, determine whether the ship has reached the destination. If the ship has reached the destination, end the voyage; otherwise, continue to sail with the current expected speed vector and execute S1 after a decision cycle. S6, judging whether the number of ships in the dangerous ship list of the own ship is greater than 1, if the number of ships in the list is equal to 1, executing S7, otherwise executing S9; S7. Based on the navigation information of own ship and target ship and referring to the International Regulations for Preventing Collisions at Sea, the encounter situation is divided; S8. Determine the collision avoidance responsibility coefficient of each ship according to the encounter situation, and execute S10; S9. Determine that each vessel bears the same avoidance responsibility coefficient; S10, obtaining the decision cycle of each decision made by the intelligent ship; S11. The ship safety domain, decision cycle, and avoidance responsibility coefficient are used as inputs of the ORCA algorithm to construct the interactive collision avoidance space between the ship and all dangerous ships, including: According to the following formula, the time window The speed obstacle zone formed by the target ship to the own ship: in , Represents the positions of the target ship and own ship respectively. , represents the safety zone between the target ship and own ship, and if , which means that the own ship and the target ship will be in the time window If a collision occurs within , it can be ensured that the ship and the target ship are in the time window Previously it was possible to travel safely without collision; If the target ship's speed changes , then the set of own ship velocity vectors satisfying the following formula is defined as the collision avoidance velocity vector space: in Represents and of Minkowski and; For the target ship, its collision avoidance velocity vector space relative to the own ship is: ,in , when the ship's velocity vector And the target ship's velocity vector When Pass safely within the time limit; For the ships that meet, the velocity vector space that satisfies the following formula is defined as the optimal interactive collision avoidance space: Assume that the own ship and the target ship have speeds , This will cause own ship to sail with the target ship A collision occurs within the time window, i.e. ,definition For The distance between the starting point and the end point The nearest vector to the region is defined as follows: The set composition in velocity space is shown in the following formula: in, To avoid the liability factor, Speed ​​barrier area Located at point The outward normal vector at For the time window The minimum change in speed between own ship and target ship to avoid collision; S12, the ship selects the speed vector with the smallest deviation from the current course as the optimal speed vector for navigation through linear programming; S13. Clear the list of dangerous ships and execute S1 after one decision cycle.

2. The distributed autonomous collaborative collision avoidance method for multiple intelligent ships according to claim 1 is characterized in that: The expected velocity vector of the ship is calculated according to the position of the ship and the terminal position, including obtaining the expected velocity vector of the ship according to the following formula: in, is the ship's speed, is the current position of the ship, The terminal position of this ship.

3. The distributed autonomous collaborative collision avoidance method for multiple intelligent ships according to claim 1 is characterized in that: The minimum encounter distance between own ship and target ship is calculated according to the following formula: Where D is the distance between the own ship and the target ship, is the true azimuth of the target ship, is the relative speed between own ship and target ship, is the ship's speed, is the ship’s heading; The minimum encounter time between own ship and target ship is calculated according to the following formula: 。 4. The distributed autonomous collaborative collision avoidance method for multiple intelligent ships according to claim 3 is characterized in that: The dangerous ship list of the ship is constructed based on the ship collision risk model, including judging whether to add the target ship to the dangerous ship list according to the following methods: in, It is the ship safety zone.

5. The distributed autonomous collaborative collision avoidance method for multiple intelligent ships according to claim 1 is characterized in that: According to the navigation information of own ship and target ship, refer to the International Regulations for Preventing Collisions at Sea, divide the encounter situation, including dividing the encounter situation according to the following methods: The encounter situation that satisfies the following formula is defined as an encounter: The encounter situation that satisfies the following formula is defined as a starboard crossing encounter: The encounter situation that satisfies the following formula is defined as overtaking: The encounter situation that satisfies the following formula is defined as a port crossing encounter: in It is the relative position between own ship and target ship.