Inland river ship autonomous collision avoidance method based on improved speed obstacle method

Abstract

The invention discloses an inland ship autonomous collision avoidance method based on an improved speed obstacle method. The inland ship autonomous collision avoidance method comprises the steps of collecting and obtaining real-time ship data of a ship and a target ship; the method comprises the following steps: determining a ship in an inland river water area, obtaining a collision risk index of a target ship and the ship, establishing an encounter situation research and judgment strategy, judging whether the ship is a way-giving ship or a straight ship according to a ship head direction and a ship flow direction, obtaining relative motion parameters of the target ship and the ship in real time, and obtaining a ship collision risk degree according to the relative motion parameters; according to the ship collision risk degree, when it is confirmed that the target ship forms urgent danger to the ship, collision avoidance decision making is carried out on the ship based on an improved speed obstacle algorithm, so that autonomous collision avoidance of the ship in the inland river water area is achieved; the problem that the speed suddenly changes in the collision avoidance process of a traditional speed obstacle method is solved, and in addition, based on the collision avoidance characteristics of inland ships, the problem that the collision avoidance efficiency is low due to the fact that the inland water area is limited cannot be effectively solved through the traditional speed obstacle method.

Description

Technical Field

[0001] The present invention relates to the technical field of autonomous collision avoidance of ships, and in particular to an autonomous collision avoidance method for inland ships based on an improved speed barrier method. Background Art

[0002] With the rapid development of inland waterway shipping, collision avoidance has become increasingly crucial for ensuring navigation safety. The limited waterways, complex traffic flows, and unique ship maneuvers in inland waterways increase the risk of collisions. This is especially true in busy waterways, where frequent encounters, often head-on or overtaking, can easily lead to collisions. Therefore, the research and development of efficient and intelligent collision avoidance algorithms to ensure safe navigation has become a key area of inland waterway technology research.

[0003] At present, there are many algorithms for ship collision avoidance, which can be roughly divided into several categories. One category is based on expert systems [1][2][3], which is the earliest method used for ship collision avoidance; the other category is based on virtual vectors / virtual fields. The more commonly used methods include artificial potential field method [4][5][6][7] and speed barrier method; in addition, with the development of intelligent technology, intelligent algorithms such as artificial neural networks [8][9]

[10] and deep reinforcement learning

[11]

[12]

[13] have also begun to be gradually applied to the field of ship collision avoidance. The speed barrier method is the main research object and is applied to inland ship collision avoidance. The main reason is that the calculation process of the speed barrier method is simple, the algorithm is concise and clear, and the method studies the speed and changes of the intelligent body, which is very suitable for the characteristics of inland waters. The width of inland waters is limited and the navigation environment is complex, which requires the algorithm to have good real-time performance; inland ships have good maneuverability and will frequently change the size and direction of speed when sailing. This is very different from marine collision avoidance. Marine collision avoidance mainly relies on turning to achieve "passing and giving way", while the speed barrier method is more flexible in speed changes, which just meets this characteristic.

[0004] In recent years, the speed barrier method and various extended algorithms have been gradually applied to the field of ship collision avoidance, achieving promising results. Initially, this method was primarily used for intelligent collision avoidance of unmanned boats. It later migrated to collision avoidance decision-making for large vessels, and research in the field of inland waterway collision avoidance has also begun to emerge. A review of recent related research has revealed several key features of the speed barrier method in the field of ship collision avoidance. First, it quantifies collision avoidance rules, enabling ships to act in accordance with these rules during avoidance maneuvers. Second, it considers the motion characteristics of ships and incorporates a maneuvering motion model into the algorithm to enhance simulation realism. The speed barrier method requires geometric manipulation of the intelligent agent, a feature that resonates with the field of nautical vessels. Therefore, selecting the appropriate vessel domain to incorporate into the algorithm is essential. However, most studies using the speed barrier method for collision avoidance typically treat the intelligent agent as a uniform circular shape. At present, most related research is based on the judgment mechanism of the speed barrier method itself to determine whether there is a collision risk. This is obviously different from the way DCPA and TCPA are usually used on ships to judge the collision risk of ships. Therefore, this topic needs to introduce the concept of collision risk to divide the collision priority of the current navigation environment in order to determine the collision priority of the ship and each ship.

[0005] Because the speed barrier method was originally developed to solve the problem of robot collision avoidance, improvements are needed for ship collision avoidance. First, in the field of ship research, ships cannot be simply treated as circular. In inland waterways, due to the limited width of the channel, an elliptical ship field that is closer to the ship type should be adopted. Currently, ship field models are mainly divided into three categories: ship fields based on statistical methods, ship fields based on analytical expressions, and ship fields based on intelligent technology. Statistical methods mainly obtain the boundaries of the ship field through probabilistic analysis, but they have certain subjectivity and limitations. Analytical expression methods quantify the ship's maneuverability and better adapt to changes in the ship field. Intelligent technology analysis rules are applied to the training of ship parameters and empirical data. Although they show certain advantages in handling complex situations, they still face problems such as poor model generalization ability and limited scope of application.

[0006] The collision risk model is a crucial component of ship collision avoidance algorithms, used to quantify the risk of ship collisions. It effectively compensates for the fact that the speed barrier method itself is not consistent with the identification of collision risk in actual navigation. Currently, weighted methods based on DCPA (distance to closest approach) and TCPA (time to closest approach) are widely used for collision risk assessment. These methods are simple and intuitive, but they typically only consider the impact of DCPA and TCPA, ignoring other factors such as the environment and ship characteristics, and are unable to comprehensively assess collision risk. Fuzzy set methods can combine multiple factors through membership functions to assess collision risk, making them suitable for application in complex environments. However, they are highly subjective and difficult to standardize. Intelligent methods such as artificial neural networks can automatically extract complex relationships through big data training, but they lack interpretability in practical applications and place high demands on data quality and quantity.

[0007] In summary, there are some urgent problems to be solved in the current research on ship collision avoidance, mainly including the following two points:

[0008] 1) The problem of sudden speed change in traditional speed barrier method

[0009] The speed barrier method is a real-time collision avoidance algorithm based on geometric solution. It avoids collision by solving the collision-free speed at a certain moment. It lacks a slow transition from the original speed to the collision-free speed, so there is a problem of sudden speed change.

[0010] 2) Problems in combining the speed barrier method with the characteristics of inland waterway vessel collision avoidance

[0011] The speed barrier method was originally used for collision avoidance between intelligent agents. When applied to the field of inland ship collision avoidance, how to make the speed barrier method conform to the ship dynamic constraints, whether the judgment of collision risk is consistent with the actual navigation, what kind of collision avoidance decision can solve the problem of limited inland waters, and what kind of ship field is suitable for narrow inland waterways, resulting in low collision avoidance efficiency. Summary of the Invention

[0012] The present invention provides an autonomous collision avoidance method for inland river vessels based on an improved speed barrier method, so as to overcome the above technical problems.

[0013] In order to achieve the above object, the technical solution of the present invention is:

[0014] An autonomous collision avoidance method for inland waterways vessels based on an improved speed barrier method specifically comprises the following steps:

[0015] S1: Collect and obtain real-time ship data of own ship and target ship;

[0016] The real-time ship data shall at least include ship AIS data, Beidou data and speed log data;

[0017] S2: Confirm the inland waters where the ship is located based on the longitude and latitude information of the ship in the AIS data;

[0018] The inland waters where the vessel is located include river waters, waters where main and tributary rivers meet, and waters at the mouth of a river;

[0019] S3: Based on BeiDou data and speed log data, obtain the collision risk index between the target ship and the own ship, and sort the collision risk indexes in descending order to obtain a ship collision risk table;

[0020] Identify the target ship with the highest collision risk index in the ship collision risk table;

[0021] A strategy for assessing encounter situations is established based on the rules for avoiding collisions in inland waterways and at sea. This strategy determines whether the own ship is a give-way ship or a straight-ahead ship based on the ship's heading and current direction. The relative motion parameters between the target ship and the own ship are acquired in real time, and the collision risk of the ships is determined based on the relative motion parameters.

[0022] The relative motion parameters include DCPA membership, TCPA membership, distance between two ships membership, minimum approach factor membership, ship speed ratio membership and minimum distance from shore membership;

[0023] If it is a give-way vessel, proceed to step S4;

[0024] If it is a straight-moving ship, continue to step S5;

[0025] S4: When the target ship is confirmed to pose an imminent danger to the own ship based on the collision risk, the own ship will execute a collision avoidance decision based on the improved speed barrier algorithm to achieve autonomous collision avoidance in inland waters.

[0026] S5: Confirm whether the target ship poses an imminent danger to the own ship;

[0027] If so, collision avoidance decision-making is performed on the own ship based on the improved speed barrier algorithm to achieve autonomous collision avoidance of ships in inland waters;

[0028] Otherwise, keep sailing straight;

[0029] S6: After the ship performs autonomous collision avoidance, steps S1 to S3 are repeated.

[0030] Furthermore, the method for enabling the own ship to perform collision avoidance decision-making based on the improved speed barrier algorithm in S4 specifically includes the following steps:

[0031] S001: Get the current data information of the ship;

[0032] And the current data information includes position data and speed data;

[0033] S002: Based on the elliptical ship field model, obtain the speed barrier cone according to the current data information;

[0034] S003: Obtain the current combined speed of the own ship relative to the target ship based on the speed data, and confirm whether the current combined speed falls within the speed obstacle cone;

[0035] If yes, it is confirmed that there is a risk of collision between the own ship and the target ship, and the reachable obstacle avoidance speed set is obtained according to the current data information based on the speed obstacle algorithm, and step S004 is continued;

[0036] Otherwise, no action is taken;

[0037] S004: According to the improved simulated annealing algorithm, search and obtain the optimal speed in the reachable obstacle avoidance speed set, and make collision avoidance decisions for the ship based on the optimal speed.

[0038] Furthermore, the improved simulated annealing algorithm in S004 specifically includes the following steps:

[0039] S100: Obtain a feasible speed from the reachable obstacle avoidance speed set, and confirm whether the feasible speed satisfies kinematic constraints, namely, speed and acceleration constraints;

[0040] If it is satisfied, then only the speed of the optimal speed in the set of achievable obstacle avoidance speeds is searched;

[0041] If not, search for the speed magnitude and speed direction of the optimal speed in the achievable obstacle avoidance speed set;

[0042] S101: Initialize algorithm parameters of the simulated annealing algorithm;

[0043] The algorithm parameters include at least the initial temperature T0, the initial solution, the objective function J ij And the termination condition temperature; and the initial solution includes the initial collision risk CRI0 and the initial velocity solution v o ;

[0044] And the expression of the objective function is

[0045]

[0046] Where: W CRI With W v Represents the weight coefficient of the objective function;

[0047] S102: Through kinematic constraints, the set of achievable obstacle avoidance velocities is randomly perturbed at the initial temperature to generate a new solution, namely a new collision risk index (CRI). new With the new speed solution v new ;

[0048] And the kinematic constraints: θnew ∈(v 0θ -θt,v 0θ +θt),v new ∈(|V0|-at,|V0|+at)

[0049] Where: v 0θ Indicates the direction of the ship's initial velocity; θ indicates the velocity turning rate; V0 indicates the magnitude of the ship's initial velocity; a indicates the ship's acceleration;

[0050] S103: Determine whether the new speed solution falls within the speed barrier cone;

[0051] If yes, execute step S104;

[0052] If not, proceed to step S105;

[0053] S104: confirm whether the maximum number of iterations has been reached;

[0054] If so, the current new speed solution is used as the output optimal speed;

[0055] If not, the initial velocity solution is updated based on the velocity update formula, and step S105 is continued; and the velocity update formula is expressed as

[0056] v new =v current -a max ×t,θ new =θ current -θ max ×t

[0057] Where: v new Indicates the updated speed; v current Indicates the speed of the last iteration; a max represents the maximum value of the ship's acceleration; t represents the time parameter; θ new Indicates the updated velocity direction; θ current Indicates the velocity direction of the previous iteration; θ max Indicates the maximum value of the speed turning rate;

[0058] S105: Obtain the current target increment ΔE between the initial speed solution and the new speed solution according to the objective function; and the expression of the target increment ΔE is

[0059] ΔE=J new -J current

[0060] Where: J new Represents the objective function J under the new velocity solution ij Function value of J currentRepresents the objective function J under the current velocity solution ij The function value of ;

[0061] S106: Confirm the size of the current target increment ΔE;

[0062] If ΔE≤0, then accept the new velocity solution; if ΔE>0, then with probability e -(ΔE / T) >rand accepts the new velocity solution; where T represents the current temperature; rand represents a random number between 0 and 1;

[0063] S107: Confirm whether the current condition temperature reaches the termination condition temperature;

[0064] If so, the current new speed solution is used as the output optimal solution, that is, the ship speed corresponding to the minimum risk of ship collision;

[0065] Otherwise, update the current condition temperature and repeat steps S102 to S106;

[0066] The updating formula of the current condition temperature is: T′=T·cooling_rate; wherein T′ represents the updated condition temperature; cooling_rate represents the temperature decay rate.

[0067] Furthermore, the encounter situation assessment strategy established in S3 based on the rules for avoiding collisions in inland waters and at sea is specifically:

[0068] Define A as the relative position of the target ship relative to the own ship, and the range is [0, 2π);

[0069] The angle between the own ship's sailing direction and the water flow direction is Alpha, ranging from [0, 2π); the angle between the target ship's sailing direction and the water flow direction is Beta, ranging from [0, 2π);

[0070] Define and obtain the encounter situation area of the target ship relative to the own ship in the inland waterway,

[0071] The encounter situation area includes the overtaking area, the encounter area, the intersection and give way area, the intersection and direct navigation area, the crossing area, the river fork area and the water area where the main / tributary rivers meet;

[0072] Based on the relative bearing A, angle Alpha, and angle Beta in the encounter situation area, determine whether own vessel is a give-way vessel or a straight-ahead vessel. Specifically:

[0073] If the target ship is in the overtaking zone: when A∈(5π / 8,11π / 8], the own ship is defined as the give-way ship, otherwise it is the straight-ahead ship;

[0074] If the target ship is in the encounter zone: when A∈(0,π / 36]∪(71π / 36,2π], and Alpha∈(π / 2,3π / 2], then the own ship is defined as the give-way ship; otherwise, the own ship is defined as the straight-ahead ship;

[0075] If the target ship is in the intersection give-way zone: when A∈(π / 36,4π / 9]∪(5π / 9,5π / 8], then the own ship is defined as the give-way ship, otherwise it is the straight-ahead ship;

[0076] If the target ship is in the crossing and meeting direct navigation area: when A∈(14π / 9,71π / 36]∪(11π / 8,13π / 9], the ship is defined as a straight-going ship; otherwise, it is a give-way ship;

[0077] If the target ship is in the crossing area: when A∈(4π / 9,5π / 9]∪(13π / 9,14π / 9], then the ship is defined as a straight-moving ship; otherwise, it is a give-way ship;

[0078] If the target vessel is in the fork in the river:

[0079] When Alpha∈(π / 2,3π / 2] and Beta∈(π / 2,3π / 2], or and

[0080] If A∈(0,π], then the ship is defined as a give-way ship; if A∈(π,2π], then the ship is defined as a straight-moving ship;

[0081] when and This defines this vessel as a give-way vessel;

[0082] when And Beta∈(π / 2,3π / 2], then the ship is defined as a straight-moving ship;

[0083] If the target ship is in the confluence of the main stream and tributary, when A∈(0,π], the ship is defined as a give-way ship; when A∈(π,2π], the ship is defined as a straight-moving ship.

[0084] Furthermore, S3 obtains the ship collision risk based on the relative motion parameters, specifically including:

[0085] S31: Obtain membership functions for solving DCPA membership, TCPA membership, two-ship distance membership, minimum approach factor membership, ship speed ratio membership, and minimum offshore distance membership respectively;

[0086] The expression of the DCPA membership function is:

[0087]

[0088] d2=2d1

[0089] Where: U(DCPA) represents the DCPA membership function; D s represents the radius of the ship's area; d1 and d2 represent intermediate parameters; k1 represents the design parameter determined by the visibility factor; k2 represents the design parameter determined by the current water conditions; f_min represents the minimum proximity factor between the target ship and the own ship;

[0090] The expression of the TCPA membership function is:

[0091]

[0092] Where: U(TCPA) represents the TCPA membership function; t1 and t2 represent intermediate parameters; D1 represents the latest avoidance distance; D2 represents the safe distance for taking avoidance measures; v r Indicates the relative speed between the target ship and own ship;

[0093] The expression of the membership function of the distance between the two ships is:

[0094]

[0095] D1=k1·k2·k3·DLA

[0096] D2=k1·k2·k3·r

[0097] Where: U(D) represents the membership function of the distance between the two ships; k3 represents the design parameter determined by human factors; DLA represents the latest steering distance of the ship; r represents the design parameter proposed based on the wide sea area;

[0098] The expression of the minimum proximity factor membership function is:

[0099]

[0100] Where: U(f_min) represents the minimum proximity factor membership function; f_min represents the minimum proximity factor between the target ship and the own ship;

[0101] The expression of the ship speed ratio membership function is:

[0102]

[0103] Where: U(K) represents the ship speed ratio membership function; K and W represent design parameters; C represents the ship collision angle and 0°≤C<180°;

[0104] The expression of the offshore minimum distance membership function is:

[0105]

[0106] Where: U(Q) represents the minimum distance from shore membership function; Q represents the minimum distance from shore; d warn Indicates the latest avoidance distance between the ship and the shore; d safe Indicates the safe avoidance distance between the ship and the shore;

[0107] S32 constructs a ship collision risk model based on the obtained DCPA membership function, TCPA membership function, two-ship distance membership function, minimum approach factor membership function, ship speed ratio membership function, and offshore minimum distance membership function;

[0108] And the expression of ship collision risk model is

[0109] e=a DCPA U(DCPA)+a TCPA U(TCPA)+a D U(D)+a f_min U(f_min)+a K U(K)+a Q U(Q)

[0110] Where: a DCPA ,a TCPA ,a D ,a f_min ,a K ,a Q They represent the weight coefficients of DCPA membership, TCPA membership, distance between two ships membership, minimum approach factor membership, ship speed ratio membership and minimum offshore distance membership respectively; e represents the risk of ship collision;

[0111] Obtaining ship collision risk according to a ship collision risk model;

[0112] The system also divides the collision risk into different levels based on the set collision risk, and determines the type of situation between the own ship and the target ship according to the collision risk level;

[0113] And the situation type represents imminent danger, dangerous situation and observation situation.

[0114] Furthermore, in S002, based on the elliptical ship domain model, a speed barrier cone is obtained according to current data information, specifically including:

[0115] S0021: Set the geographic coordinates of this ship to Pos A =(x A ,y A ), ship speed The length of the ship is l, and the major axis of the ship is l A =3l, semi-minor axis w A =0.8l;

[0116] Assume the geographic coordinates of the target ship is Pos B =(x B ,y B ), ship speed The length of the ship is lo, and the major axis of the ship is l B =3lo, semi-minor axis w B =0.8lo;

[0117] S0022: Obtain the elliptical ship domain model of the own ship and the elliptical ship domain model of the target ship; and the expression of the elliptical ship domain model of the own ship is

[0118] A1x 2 +B1xy)+C1y 2 +D1x+E1y+F1=0

[0119] A1=(l A sinθ A ) 2 +(w A cosθ A ) 2

[0120]

[0121] C1=(l A cosθ A ) 2 +(w A sinθ A ) 2

[0122] D1=-2A1x A -B1y A

[0123] E1=-B1x A -2C1y A

[0124]

[0125] Where: A1, B1, C1, D1, E1, F1 represent intermediate variables; Indicates the ellipse direction angle θ corresponding to the ship A Speed; x, y represent the equation parameters of the elliptical ship field model;

[0126] The expression of the elliptical ship domain model of the target ship is:

[0127] A2x 2 +B2xy+C2y 2 +D2x+E2y+F2=0

[0128] A2=(l B sinθ B ) 2 +(w B cosθ B ) 2

[0129]

[0130] C2=(l B cosθ B ) 2 +(w B sinθ B ) 2

[0131] D2=-2A2x B -B2y B

[0132] E2=-B2x B -2C2y B

[0133]

[0134] Where: A2, B2, C2, D2, E2, F2 represent intermediate variables; Indicates the ellipse direction angle θ corresponding to the ship B speed;

[0135] S0023: Assume that the tangent equation of the elliptical ship domain model is y=kx+b, where k and b represent the parameters to be solved in the tangent equation;

[0136] And bring it into the elliptical ship domain model of the own ship and the elliptical ship domain model of the target ship respectively, and solve k and b simultaneously to obtain several tangent equations to be solved;

[0137] S0024: Obtain the angle α corresponding to each tangent line based on the slope of each solved tangent line equation i And α i =arctan(k i ); where α i represents the angle corresponding to the equation of the i-th tangent line; k i represents the slope of the i-th tangent line equation;

[0138] Obtain the direction angle φ of the line connecting the center of the own ship and the target ship, and φ=arctan2(y B -y A ,x B -x A ), and according to the direction angle φ and angle αi Determine the boundary of the speed barrier cone to obtain the speed barrier cone;

[0139] And the boundary expression of the speed barrier cone is

[0140] Alpha=max(α i -φ),Beta=min(α i -φ)

[0141] Where: Alpha and Beta represent the two boundaries of the speed barrier cone.

[0142] Beneficial effects: The present invention discloses an autonomous collision avoidance method for inland vessels based on an improved speed barrier method, which aims to solve the problem of ship collision avoidance in complex inland water environments. By improving the traditional speed barrier method and combining it with a simulated annealing algorithm, variable speed collision avoidance is combined with variable speed plus steering collision avoidance. Kinematic constraints are introduced based on the kinematic characteristics of the ship to ensure that the speed change of the ship is smoother, avoiding the problem of sudden speed changes in traditional algorithms. At the same time, based on the existing "Inland Waterway Collision Prevention Regulations" and "Sea Collision Prevention Regulations", a meeting situation analysis strategy is constructed to accurately judge the meeting situation of ships and quantify the collision risk. In addition, the present invention combines the elliptical ship field model with the collision risk model to obtain a multi-model fusion intelligent collision avoidance decision-making method, which can effectively improve the collision avoidance capability of inland vessels in narrow waterways. BRIEF DESCRIPTION OF THE DRAWINGS

[0143] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. 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 any creative labor.

[0144] Figure 1 This is a flow chart of the inland river vessel autonomous collision avoidance method based on the improved speed barrier method of the present invention;

[0145] Figure 2 This is a technical flowchart of the autonomous collision avoidance method for inland waterways vessels in this embodiment;

[0146] Figure 3 This is a collision avoidance flow chart in this embodiment;

[0147] Figure 4 This is a schematic diagram of the speed barrier method in this embodiment;

[0148] Figure 5 Schematic diagram of the absolute speed barrier cone under speed and acceleration constraints in this embodiment;

[0149] Figure 6 This is a flow chart of the speed barrier method in this embodiment;

[0150] Figure 7 This is a flow chart of the improved speed barrier method in this embodiment;

[0151] Figure 8 This is the flow chart of the improved simulated annealing algorithm in this embodiment. DETAILED DESCRIPTION

[0152] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of 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. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0153] This embodiment provides an autonomous collision avoidance method for inland vessels based on an improved speed barrier method, which is characterized in that: Figures 1 to 2 As shown, the specific steps include:

[0154] S1: Collect and obtain real-time ship data of own ship and target ship;

[0155] The real-time ship data shall at least include ship AIS data, Beidou data and speed log data;

[0156] Specifically, this embodiment collects AIS data, Beidou data, and speed log data of the target ship and other ships through preset sensors, and these data at least include the current position, heading, speed, and status information of the ship during navigation;

[0157] S2: Confirm the inland waters where the ship is located based on the longitude and latitude information of the ship in the AIS data;

[0158] The inland waters where the vessel is located include river waters, waters where main and tributary rivers meet, and waters at the mouth of a river;

[0159] Specifically, if Figure 3 As shown, this embodiment identifies the vessel's route segment based on the vessel's latitude and longitude information: that is, using the vessel's latitude and longitude information, the vessel's specific route segment is determined based on the vessel's location and trajectory, such as whether it is in a waterway, a confluence of main and tributary rivers, or a port.

[0160] S3: Based on BeiDou data and speed log data, obtain the collision risk index between the target ship and the own ship, and sort the collision risk indexes in descending order to obtain a ship collision risk table;

[0161] Identify the target ship with the highest collision risk index in the ship collision risk table;

[0162] A strategy for assessing encounter situations is established based on existing inland waterway collision avoidance rules and maritime collision avoidance rules, so as to determine whether the vessel is a give-way vessel or a straight-ahead vessel based on the vessel's heading and current direction.

[0163] In a specific embodiment, the encounter situation analysis strategy established in S3 based on the inland river collision avoidance rules and the maritime collision avoidance rules is specifically:

[0164] Define A as the relative position of the target ship relative to the own ship, and the range is [0, 2π);

[0165] The angle between the own ship's sailing direction and the water flow direction is Alpha, ranging from [0, 2π); the angle between the target ship's sailing direction and the water flow direction is Beta, ranging from [0, 2π);

[0166] Define and obtain the encounter situation area of the target ship relative to the own ship in the inland waterway,

[0167] The encounter situation area includes the overtaking area, the encounter area, the intersection and give way area, the intersection and direct navigation area, the crossing area, the river fork area and the water area where the main / tributary rivers meet;

[0168] Based on the relative bearing A, angle Alpha, and angle Beta in the encounter situation area, determine whether own vessel is a give-way vessel or a straight-ahead vessel. Specifically:

[0169] If the target ship is in the overtaking zone: when A∈(5π / 8,11π / 8], the own ship is defined as the give-way ship, otherwise it is the straight-ahead ship;

[0170] If the target ship is in the encounter zone: when A∈(0,π / 36]∪(71π / 36,2π], and Alpha∈(π / 2,3π / 2], then the own ship is defined as the give-way ship; otherwise, the own ship is defined as the straight-ahead ship;

[0171] If the target ship is in the intersection give-way zone: when A∈(π / 36,4π / 9]∪(5π / 9,5π / 8], then the own ship is defined as the give-way ship, otherwise it is the straight-moving ship;

[0172] If the target ship is in the crossing and meeting direct navigation area: when A∈(14π / 9,71π / 36]∪(11π / 8,13π / 9], the ship is defined as a straight-going ship; otherwise, it is a give-way ship;

[0173] If the target ship is in the crossing area: when A∈(4π / 9,5π / 9]∪(13π / 9,14π / 9], then the ship is defined as a straight-moving ship; otherwise, it is a give-way ship;

[0174] If the target vessel is in the fork in the river:

[0175] When Alpha∈(π / 2,3π / 2] and Beta∈(π / 2,3π / 2], or and

[0176] If A∈(0,π], then the ship is defined as a give-way ship; if A∈(π,2π], then the ship is defined as a straight-moving ship;

[0177] When Alpha∈(π / 2,3π / 2] and This defines this vessel as a give-way vessel;

[0178] when And Beta∈(π / 2,3π / 2], then the ship is defined as a straight-moving ship;

[0179] If the target ship is in the confluence of the main stream and tributary; when A∈(0,π], the ship is defined as a give-way ship; when A∈(π,2π], the ship is defined as a straight-moving ship;

[0180] The relative motion parameters between the target ship and the own ship are acquired in real time, and the collision risk of the ships is obtained based on the relative motion parameters; the relative motion parameters include DCPA membership, TCPA membership, distance between the two ships membership, minimum approach factor membership, ship speed ratio membership, and minimum distance from shore membership;

[0181] In a specific embodiment, the method for obtaining the ship collision risk in S3 specifically includes:

[0182] S31: Obtain membership functions for solving DCPA membership, TCPA membership, two-ship distance membership, minimum approach factor membership, ship speed ratio membership, and minimum offshore distance membership respectively;

[0183] The expression of the DCPA membership function is:

[0184]

[0185] d2=2d1

[0186] Where: U(DCPA) represents the DCPA membership function; D s represents the radius of the ship area; d1 and d2 represent intermediate parameters; k1 represents the design parameter determined by the visibility factor, and the smaller the visibility, the larger the k1 value; k2 represents the design parameter determined by the current water conditions, and when the visibility is poor in the current water conditions, the d1 value can be expanded to increase the warning range; in complex waters, such as coastal ships, the d1 value should be increased due to the limited navigation waters; f_min represents the minimum proximity factor between the target ship and the own ship, and f _minThe smaller the value, the deeper the target ship invades the ship's area and the greater the risk of collision. _min When <1, the radius of the ship area should be expanded. At this time, the DCPA membership function increases and the collision risk increases;

[0187] The expression of the TCPA membership function is:

[0188]

[0189] Where: U(TCPA) represents the TCPA membership function; t1 and t2 represent intermediate parameters; D1 represents the latest avoidance distance; D2 represents the safe distance for taking avoidance measures; v r Indicates the relative speed between the target ship and own ship;

[0190] The expression of the membership function of the distance between the two ships is:

[0191]

[0192] D1=k1·k2·k3·DLA

[0193] D2=k1·k2·k3·r

[0194] Where: U(D) represents the membership function of the distance between the two ships; k3 represents the design parameter determined by human factors; r represents the design parameter proposed based on wide sea areas; the smaller the membership of the distance between the two ships, the closer the distance between the two ships, the greater the degree of danger; D1 is the latest avoidance distance, and D2 is the distance at which avoidance measures can be taken, and their values are affected by the conditions of the navigation area, visibility conditions and human factors; k3 is determined by human factors, including the experience, skills, reaction ability and psychological quality of the operator; when the operator's quality is good, k3 is small; when the operator's quality is poor, k3 is large; DLA represents the latest steering distance of the ship, which is taken as 5 times the ship length in this paper; since the dynamic range model is proposed based on wide sea areas, the dynamic range should be narrowed in this application, so r = 0.5R;

[0195] The method for obtaining the minimum proximity factor membership function is:

[0196] Get the close factor membership mathematical model, its expression is

[0197]

[0198] in

[0199] x o =x1+V o t sinC o

[0200] y o=y1+V o t cosC o

[0201] x t =x2+V t t sinC o

[0202] y t =y2+V t t cosC o

[0203] After simplifying it, we can get

[0204] f(t) 2 =At 2 +Bt+C

[0205] Differentiating f(t) with respect to t yields

[0206]

[0207] Let the above formula be equal to 0, then we can get

[0208]

[0209] T min Substitute f(t) 2 =At 2 +Bt+C, we can get

[0210]

[0211] Therefore, when the minimum proximity factor f min ≤1 indicates that the target ship will invade the ship's area at some point in the future, and f min The deeper the penetration, the greater the risk of collision. min ≤0.5 means the two ships will pose an imminent danger, at this time U(f_min)=1, when 0.5<f min ≤1 trigonometric function construct f min function, when the minimum close factor f min >1 indicates that the target ship will not invade the ship's area, U(f_min)=0;

[0212] Then the minimum proximity factor membership function can be expressed as:

[0213]

[0214] Where: U(f_min) represents the minimum proximity factor membership function; f_min represents the minimum proximity factor between the target ship and the own ship;

[0215] The expression of the ship speed ratio membership function is:

[0216]

[0217] Where: U(K) represents the ship speed ratio membership function; K and W represent design parameters; C represents the ship collision angle and 0°≤C<180°; and the greater the ship speed ratio, the greater the collision risk;

[0218] The expression of the offshore minimum distance membership function is:

[0219]

[0220] Where: U(Q) represents the minimum distance from shore membership function; Q represents the minimum distance from shore; d warn Indicates the latest avoidance distance between the ship and the shore; d safe Indicates the safe avoidance distance between the ship and the shore. In this embodiment, the degree of danger posed by the offshore distance to the ship is very obvious. The smaller the offshore distance, the greater the collision risk.

[0221] S32 constructs a ship collision risk model based on the obtained DCPA membership function, TCPA membership function, two-ship distance membership function, minimum approach factor membership function, ship speed ratio membership function, and offshore minimum distance membership function;

[0222] And the expression of ship collision risk model is

[0223] e=a DCPA U(DCPA)+a TCPA U(TCPA)+a D U(D)+a f_min U(f_min)+a K U(K)+a Q U(Q)

[0224] Where: a TCPA ,a TCPA ,a D ,a f_min ,a K ,a Q They represent the weight coefficients of DCPA membership, TCPA membership, distance between two ships membership, minimum approach factor membership, ship speed ratio membership and minimum offshore distance membership respectively; e represents the risk of ship collision;

[0225] Obtaining ship collision risk according to a ship collision risk model;

[0226] The system also divides the collision risk into different levels based on the set collision risk, and determines the type of situation between the own ship and the target ship according to the collision risk level;

[0227] and the situation type represents an imminent danger, a dangerous situation, and an observation situation;

[0228] According to the empirical value, the importance of factors affecting the collision risk in this embodiment is ranked and the weight values are determined. Since the waterway is the key factor to be considered in inland navigation, the weight distribution of each factor is as follows: a DCPA =0.29,a TCPA =0.26,a D =0.11,a f_min =0.08,a K =0.06,a Q =0.2;

[0229] According to the size of the e value, the collision risk is graded as shown in Table 1;

[0230] Table 1. Collision risk classification table

[0231] I 1.0～0.61 Urgent situation II 0.6～0.41 dangerous situation III 0.4～0 Observation situation

[0232] In this embodiment, when the collision risk is level 1, it means that an urgent situation has formed. Regardless of whether the target ship takes evasive action, the own ship should take evasive action. When the collision risk is level II, if the own ship is a give-way ship, it should take immediate action to avoid the formation of an urgent situation. Otherwise, if the own ship is a straight-line ship, it is necessary to observe whether the other ship takes evasive action. When the collision risk is level III, it is necessary to pay attention to observing the ships in the waters near the own ship. When TCPA is less than 0, it is generally considered that the collision risk of the own ship has been resolved and the target ship has passed and given way. Therefore, in this paper, when TCPA is less than 0, the collision risk is equal to 0.

[0233] If it is a give-way vessel, proceed to step S4;

[0234] If it is a straight-moving ship, continue to step S5;

[0235] S4: When the target ship is confirmed to pose an imminent danger to the own ship based on the collision risk, the own ship will execute a collision avoidance decision based on the improved speed barrier algorithm to achieve autonomous collision avoidance in inland waters.

[0236] In a specific embodiment, Figures 4 to 7 As shown, the method for enabling the own ship to perform collision avoidance decision-making based on the improved speed barrier algorithm in S4 specifically includes the following steps:

[0237] S001: Get the current data information of the ship;

[0238] And the current data information includes position data and speed data;

[0239] S002: Based on the elliptical ship field model, obtain the speed barrier cone according to the current data information;

[0240] The specific steps include:

[0241] S0021: Set the geographic coordinates of this ship to Pos A =(x A ,y A ), ship speed The length of the ship is l, and the major axis of the ship is l A =3l, semi-minor axis w A =0.8l;

[0242] Assume the geographic coordinates of the target ship is Pos B =(x B ,y B ), ship speed The length of the ship is lo, and the major axis of the ship is l B =3lo, semi-minor axis w B =0.8lo;

[0243] S0022: Obtain the elliptical ship domain model of the own ship and the elliptical ship domain model of the target ship; and the expression of the elliptical ship domain model of the own ship is

[0244] A1x 2 +B1xy+C1y 2 +D1x+E1y+F1=0

[0245] A1=(l A sinθ A ) 2 +(w A cosθ A ) 2

[0246]

[0247] C1=(l A cosθ A ) 2 +(w A sinθ A ) 2

[0248] D1=-2A1x A -B1y A

[0249] E1=-B1x A -2C1y A

[0250]

[0251] Where: A1, B1, C1, D1, E1, F1 represent intermediate variables; Indicates the ellipse direction angle θ corresponding to the ship A Speed; x, y represent the equation parameters of the elliptical ship field model;

[0252] The expression of the elliptical ship domain model of the target ship is:

[0253] A2x 2 +B2xy+C2y 2 +D2x+E2y+F2=0

[0254] A2=(l B sinθ B ) 2 +(w B cosθ B ) 2

[0255]

[0256] C2=(l B cosθ B ) 2 +(w B sinθ B ) 2

[0257] D2=-2A2x B -B2y B

[0258] E2=-B2x B -2C2y B

[0259]

[0260]

[0261] Where: A2, B2, C2, D2, E2, F2 represent intermediate variables; Indicates the ellipse direction angle θ corresponding to the ship B speed;

[0262] S0023: Assume that the tangent equation of the elliptical ship domain model is y=kx+b, where k and b represent the parameters to be solved in the tangent equation;

[0263] And bring it into the elliptical ship domain model of the own ship and the elliptical ship domain model of the target ship respectively, and solve k and b simultaneously to obtain several tangent equations to be solved;

[0264] S0024: Obtain the angle α corresponding to each tangent line according to the slopes of the solved tangent equations. i And α i = arctan(k i ); where α i represents the angle corresponding to the i-th tangent equation; k i represents the slope of the i-th tangent equation.

[0265] Obtain the direction angle φ of the line connecting the center of the own ship and the target ship, and φ = arctan2(y B - y A , x B - x A ), and determine the boundary of the velocity obstacle cone based on the direction angle φ and the angle α i to obtain the velocity obstacle cone.

[0266] And the expression for the boundary of the velocity obstacle cone is

[0267] Alpha = max(α i - φ), Beta = min(α i - φ)

[0268] In the formula: Alpha and Beta represent the two boundaries of the velocity obstacle cone.

[0269] This embodiment further includes a method for selecting a feasible velocity based on the velocity obstacle cone:

[0270] Obtain the relative velocity direction v of the own ship and the target ship R as

[0271]

[0272] According to the relative velocity direction v R obtain the ship velocity angle, and its expression is

[0273]

[0274] In this embodiment, if Beta < Γ < Alpha, it is confirmed that there is a collision risk for the ship, and then it is necessary to search for a new angle of the own ship's velocity V new whose magnitude ∈ (|v A | - at, |v A | + at) and the angle of the velocity

[0275] S003: Obtain the current combined velocity of the own ship relative to the target ship according to the velocity data, and confirm whether the current combined velocity falls within the velocity obstacle cone;

[0276] If yes, it is confirmed that there is a risk of collision between the own ship and the target ship, and the reachable obstacle avoidance speed set is obtained according to the current data information based on the speed obstacle algorithm, and step S004 is continued;

[0277] Otherwise, no processing is done;

[0278] S004: Based on the improved simulated annealing algorithm, search and obtain the optimal speed in the set of achievable obstacle avoidance speeds, and make collision avoidance decisions for the own ship based on the optimal speed;

[0279] In a specific embodiment, Figure 8 As shown, the improved simulated annealing algorithm includes the following steps:

[0280] S100: Obtain a feasible speed from the reachable obstacle avoidance speed set, and confirm whether the feasible speed satisfies kinematic constraints, namely, speed and acceleration constraints;

[0281] If it is satisfied, then only the speed of the optimal speed in the set of achievable obstacle avoidance speeds is searched;

[0282] If it is not satisfied, search for the speed magnitude and speed direction of the optimal speed in the achievable obstacle avoidance speed set;

[0283] S101: Initialize algorithm parameters of the simulated annealing algorithm;

[0284] The algorithm parameters include at least the initial temperature T0, the initial solution, the objective function J ij and the termination condition temperature; and the initial solution includes the initial collision risk CRI0 and the initial velocity solution v0;

[0285] And the expression of the objective function is

[0286]

[0287] Where: W CRI With W v Represents the weight coefficient of the objective function;

[0288] S102: Through kinematic constraints, the set of achievable obstacle avoidance velocities is randomly perturbed at the initial temperature to generate a new solution, namely a new collision risk index (CRI). new With the new speed solution v new ;

[0289] And the kinematic constraints: θ new ∈(v 0θ -θt,v 0θ +θt),v new ∈(|V0|-at,|V0|+at)

[0290] Where: v 0θIndicates the direction of the ship's initial velocity; θ indicates the velocity turning rate; V0 indicates the magnitude of the ship's initial velocity; a indicates the ship's acceleration;

[0291] S103: Determine whether the new speed solution falls within the speed barrier cone;

[0292] If yes, execute step S104;

[0293] If not, proceed to step S105;

[0294] S104: confirm whether the maximum number of iterations has been reached;

[0295] If so, the current new speed solution is used as the output optimal speed;

[0296] If not, the initial velocity solution is updated based on the velocity update formula, and step S105 is continued; and the velocity update formula is expressed as

[0297] v new -v current -a max ×t,O new -O current -O max ×t

[0298] Where: v new Indicates the updated speed; v current Indicates the speed of the last iteration; a max represents the maximum value of the ship's acceleration; t represents the time parameter; θ new Indicates the updated velocity direction; θ current Indicates the velocity direction of the previous iteration; θ max Indicates the maximum value of the speed turning rate;

[0299] S105: Obtain the current target increment ΔE between the initial speed solution and the new speed solution according to the objective function; and the expression of the target increment ΔE is

[0300] ΔE=J new -J current

[0301] Where: J new Represents the objective function J under the new velocity solution ij Function value of J current Represents the objective function J under the current velocity solution ij The function value of ;

[0302] S106: Confirm the size of the current target increment ΔE;

[0303] If ΔE≤0, then accept the new velocity solution; if ΔE>0, then with probability e-(ΔE / T) >rand accepts the new velocity solution; where T represents the current temperature; rand represents a random number between 0 and 1;

[0304] S107: Confirm whether the current condition temperature reaches the termination condition temperature;

[0305] If so, the current new speed solution is used as the output optimal solution, that is, the ship speed corresponding to the minimum risk of ship collision;

[0306] Otherwise, update the current condition temperature and repeat steps S102 to S106;

[0307] The update formula of the current condition temperature is: T′=T·cooling_rate; where T′ represents the updated condition temperature; cooling_rate represents the temperature decay rate;

[0308] S5: Confirm whether the target ship poses an imminent danger to the own ship;

[0309] If so, collision avoidance decision-making is performed on the own ship based on the improved speed barrier algorithm to achieve autonomous collision avoidance of ships in inland waters;

[0310] Otherwise, keep sailing straight;

[0311] S6: After the ship performs autonomous collision avoidance, steps S1 to S3 are repeated.

[0312] This embodiment also includes a method for confirming whether autonomous collision avoidance is completed:

[0313] In navigation calculation, the coordinate system oy points to the north and ox points to the east. Assume that the geographic coordinates of the ship are (x o ,y o ), own ship’s speed v o , heading C0, when the target ship is found, the relevant navigation data of the ship is obtained by using the existing ship-borne AIS, GPS, RADAR / ARPA and other equipment. The geographic coordinates of the target ship are (x t ,y t ), ship speed v t 、Course C t The distance between the two ships is D, and the relative bearing of the target ship to the own ship is T r ; The following data can be obtained

[0314] (1) The distance between the two ships is:

[0315]

[0316] (2) Change the own ship's speed to v o With the target ship speed v tBy performing vector decomposition on the x and y axes, we can obtain

[0317]

[0318] (3) Calculate the relative speed of the two ships

[0319] ① Set the relative speed v r Vector decomposition on the x,y axis

[0320]

[0321] ②Calculate the relative velocity v r Size

[0322]

[0323] ③Calculate the relative velocity v r Direction

[0324]

[0325] Where α1 is

[0326]

[0327] (4) Calculate the true bearing of the target ship

[0328]

[0329] Where α2 is

[0330]

[0331] (5) Target ship true bearing T and relative bearing T r and own ship's course C o The relationship between:

[0332]

[0333] (6) Target ship position (x t ,y t ) and the relationship between the true bearing T and the distance D between the two ships:

[0334]

[0335] (7) The DCPA between the two ships is DCPA = Dsin(rT-π)

[0336] (8) The TCPA between the two ships is

[0337] By obtaining the values of DCPA and TCPA in real time, we can confirm whether autonomous collision avoidance has been completed:

[0338] If DCPA is negative, it means the target ship has passed by the stern of own ship. Otherwise, it means the target ship has passed by the bow of own ship. If TCPA is negative, it means the own ship and target ship have already passed by and given way. Otherwise, it means the own ship and target ship have not yet passed by and given way.

[0339] Compared with the prior art, this embodiment has the following beneficial effects:

[0340] 1. Innovation of autonomous collision avoidance algorithm for ships based on improved speed barrier method

[0341] Introducing kinematic constraints: By introducing velocity and acceleration constraints, the ship can adjust its speed more smoothly during collision avoidance, preventing sudden speed changes and drastic turns. This innovation better meets the collision avoidance requirements of narrow inland waterways and complex traffic conditions, and enhances the algorithm's practical feasibility.

[0342] Improvements to the motion characteristics of inland vessels: The traditional speed barrier method is designed to resolve collisions between intelligent agents. When directly transplanted to inland vessels, it cannot fully consider the dynamic constraints of the ship and the hydrological characteristics of the inland river. By combining the simulated annealing algorithm, the traditional speed barrier algorithm is improved to make it better adapt to the maneuvering characteristics of ships in inland waters, further improving the effectiveness of collision avoidance decisions.

[0343] In inland collision avoidance scenarios, due to the narrow waterways and limited space for ship operation, the collision avoidance decision based on speed and direction in the traditional VO algorithm needs to be redesigned. Inland collision avoidance usually adopts a collision avoidance decision of speed change or speed change plus steering, with the priority being speed change > speed change plus steering. Speed change operation and speed change plus steering cannot be directly obtained through geometric operations. Therefore, by introducing a heuristic algorithm - simulated annealing algorithm, a speed change plus steering collision avoidance decision is generated that can effectively reduce the risk of collision and ensure the minimum steering amplitude. The improved speed barrier algorithm can avoid the violent and large-scale steering that may occur in the traditional algorithm, which not only reduces fuel consumption but also improves the ship's maneuverability and stability.

[0344] 2. Innovation in the construction of an autonomous collision avoidance model for inland waterways vessels based on multi-model fusion

[0345] Introduction of ship maneuvering motion model: Traditional speed barrier methods are mostly based on geometric algorithms, ignoring the dynamic constraints of the ship. This embodiment takes into account the actual physical motion characteristics of the ship by introducing an elliptical ship domain model, avoiding the limitation of the algorithm based only on geometric constraints for collision avoidance. This innovation increases the physical accuracy of the collision avoidance algorithm and can more accurately reflect the maneuverability of ships in complex waters. In addition, the elliptical ship domain model adds an additional safety margin to the collision avoidance algorithm, especially for narrow channels and multi-ship intersection scenarios. By defining the safe space around the ship, it effectively prevents excessive approach between ships, which not only improves the collision avoidance safety of the ship, but also combines with the speed barrier method to make up for the lack of adaptability of the VO algorithm in complex environments.

[0346] 3. Quantification of the Collision Risk Model: By introducing a ship collision risk model, the potential collision risk in complex inland waterway traffic flows is quantified. This model prioritizes collisions based on the relative motion of different ships, helping the collision avoidance algorithm make more reasonable collision avoidance decisions in multi-vessel environments. This quantification method makes the collision avoidance decision-making process more transparent and controllable, improving the algorithm's reliability in complex traffic environments.

[0347] This embodiment integrates a vessel domain model with a collision risk model and an improved speed barrier method. This allows the collision avoidance algorithm to not only consider the vessel's motion and speed, but also incorporate multiple factors such as safety, real-time performance, and complex water environments. This innovative multi-model fusion ensures the algorithm can address the diverse needs of inland vessel collision avoidance and improves the performance and reliability of the overall collision avoidance system.

[0348] The references involved in this embodiment are as follows:

[0349] [1] Hayabusa Imazu, Akio Sugisaki, Saburo Tsuruta, et al. Basic research on ship navigation expert system [J]. China Navigation, 1989(02): 106-109. DOI:CNKI:SUN:ZGHH.0.1989-02-013.

[0350] [2] Zhang Xuankui. Research on Ship Collision Avoidance Expert System Based on AIS[D]. Shanghai Maritime University, 2006. DOI: 10.7666 / d.y1236359.

[0351] [3] Lv Hongguang, Yin Yong, Yin Jianchuan, et al. Ship automatic collision avoidance decision algorithm based on artificial intelligence and soft computing [J]. China Navigation, 2016, 39(3):7. DOI: 10.3969 / j.issn.1000-4653.2016.03.009.

[0352] [4] Lee SM, Kwon KY, Joh JA fuzzy logic for autonomous navigation of marine vehicles satisfying COLREG guidelines [J]. International journal of control, automation, and systems, 2004, 2(2): 171-181.

[0353] [5]Lyu H,Yin Y.Ship's trajectory planning for collision avoidance atsea based on modified artificial potential field[C] / / 20172nd International conference on robotics and automation engineering(ICRAE).IEEE,2017:351-357.

[0354] [6]Bosanquet C B.Autonomous Vessel Policy Work Requires ProactiveMeasures and Measurements[J].Coast Guard Journal of Safety&Security at Sea,Proceedings of the Marine Safety&Security Council,2022,79(1).

[0355] [7] Lv Hongguang. Research on multi-ship collision avoidance decision and path planning based on electronic charts [J]. Dalian Maritime University [2024-09-06].

[0356] [8] Ni Haozhe. Application of neural network algorithm in artificial intelligence recognition [J]. Information Communication, 2021(034-012).

[0357] [9] Yang Guobin. Research on the application of neural network based on genetic algorithm to determine the risk of ship collision[D]. Hebei University of Technology, 2006.

[0358]

[10] Chen Xiaofeng. Research on ship collision risk based on neural network in complex environment[D]. Dalian Maritime University, 2022.DOI:10.26989 / d.cnki.gdlhu.2022.000771.

[0359]

[11] Wang Chengbo. Research on key technologies of intelligent collision avoidance decision-making for autonomous ships based on deep reinforcement learning[D]. Dalian Maritime University, 2023. DOI: 10.26989 / d.cnki.gdlhu.2023.000087.

[0360]

[12] Guan Wei, Luo Wenzhe, Cui Zhewen. Decision-making method for collision avoidance behavior of unmanned ships based on deep reinforcement learning[J]. Journal of Dalian Maritime University, 2024, 50(01):11-19. DOI:10.16411 / j.cnki.issn1006-7736.2024.01.002.

[0361]

[13] Liu Jiao. Collision avoidance decision-making of autonomous surface ships based on deep reinforcement learning[D]. Dalian Maritime University, 2023.DOI:10.26989 / d.cnki.gdlhu.2023.000094.

[0362] 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 above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above 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. An autonomous collision avoidance method for inland vessels based on an improved speed barrier method, characterized in that: The specific steps include: S1: Collect and obtain real-time ship data of own ship and target ship; The real-time ship data shall at least include ship AIS data, Beidou data and speed log data; S2: Confirm the inland waters where the ship is located based on the longitude and latitude information of the ship in the AIS data; The inland waters where the vessel is located include river waters, waters where main and tributary rivers meet, and waters at the mouth of a river; S3: Based on BeiDou data and speed log data, obtain the collision risk index between the target ship and the own ship, and sort the collision risk indexes in descending order to obtain a ship collision risk table; Identify the target ship with the highest collision risk index in the ship collision risk table; A strategy for assessing encounter situations is established based on the rules for avoiding collisions in inland waterways and at sea. This strategy determines whether the own ship is a give-way ship or a straight-ahead ship based on the ship's heading and current direction. The relative motion parameters between the target ship and the own ship are acquired in real time, and the collision risk of the ships is determined based on the relative motion parameters. The relative motion parameters include DCPA membership, TCPA membership, distance between two ships membership, minimum approach factor membership, ship speed ratio membership and minimum distance from shore membership; If it is a give-way vessel, proceed to step S4; If it is a straight-moving ship, continue to step S5; S4: When the target ship is confirmed to pose an imminent danger to the own ship based on the collision risk, the own ship will execute a collision avoidance decision based on the improved speed barrier algorithm to achieve autonomous collision avoidance in inland waters. S5: Confirm whether the target ship poses an imminent danger to the own ship; If so, collision avoidance decision-making is performed on the own ship based on the improved speed barrier algorithm to achieve autonomous collision avoidance of ships in inland waters; Otherwise, keep sailing straight; S6: After the ship performs autonomous collision avoidance, steps S1 to S3 are repeated.

2. The method for autonomous collision avoidance of inland waterway vessels based on the improved speed barrier method according to claim 1, characterized in that: The method for enabling the own ship to perform collision avoidance decision-making based on the improved speed barrier algorithm in S4 specifically includes the following steps: S001: Get the current data information of the ship; And the current data information includes position data and speed data; S002: Based on the elliptical ship field model, obtain the speed barrier cone according to the current data information; S003: Obtain the current combined speed of the own ship relative to the target ship based on the speed data, and confirm whether the current combined speed falls within the speed obstacle cone; If yes, it is confirmed that there is a risk of collision between the own ship and the target ship, and the reachable obstacle avoidance speed set is obtained according to the current data information based on the speed obstacle algorithm, and step S004 is continued; Otherwise, no action is taken; S004: According to the improved simulated annealing algorithm, search and obtain the optimal speed in the reachable obstacle avoidance speed set, and make collision avoidance decisions for the ship based on the optimal speed.

3. The method for autonomous collision avoidance of inland waterway vessels based on the improved speed barrier method according to claim 2, characterized in that: The improved simulated annealing algorithm in S004 specifically includes the following steps: S100: Obtain a feasible speed from the reachable obstacle avoidance speed set, and confirm whether the feasible speed satisfies kinematic constraints, namely, speed and acceleration constraints; If it is satisfied, then only the speed of the optimal speed in the set of achievable obstacle avoidance speeds is searched; If not, search for the speed magnitude and speed direction of the optimal speed in the achievable obstacle avoidance speed set; S101: Initialize algorithm parameters of the simulated annealing algorithm; The algorithm parameters include at least the initial temperature T0, the initial solution, the objective function J ij and the termination condition temperature; and the initial solution includes the initial collision risk CRI0 and the initial velocity solution v0; And the expression of the objective function is Where: W CRI With W v Represents the weight coefficient of the objective function; S102: Through kinematic constraints, the set of achievable obstacle avoidance velocities is randomly perturbed at the initial temperature to generate a new solution, namely a new collision risk index (CRI). new With the new speed solution v new ; And the kinematic constraints: θ new ∈(v 0θ -θt,v 0θ +θt),v new ∈(|V0|-at,|V0|+at) Where: v 0θ Indicates the direction of the ship's initial velocity; θ indicates the velocity turning rate; V0 indicates the magnitude of the ship's initial velocity; a indicates the ship's acceleration; S103: Determine whether the new speed solution falls within the speed barrier cone; If yes, proceed to step S104; If not, proceed to step S105; S104: confirm whether the maximum number of iterations has been reached; If so, the current new speed solution is used as the output optimal speed; If not, the initial velocity solution is updated based on the velocity update formula, and step S105 is continued; And the expression of the velocity update formula is v new =v current -a max ×t,θ new =θ current -θ max ×t Where: v new Indicates the updated speed; v current Indicates the speed of the last iteration; a max represents the maximum value of the ship's acceleration; t represents the time parameter; θ new Indicates the updated velocity direction; θ current Indicates the velocity direction of the previous iteration; θ max Indicates the maximum value of the speed turning rate; S105: Obtain the current target increment ΔE between the initial speed solution and the new speed solution according to the target function; and the expression of the target increment ΔE is ΔE=J new -J current Where: J new Represents the objective function J under the new velocity solution ij Function value of J current Represents the objective function J under the current velocity solution ij The function value of ; S106: Confirm the size of the current target increment ΔE; If ΔE≤0, then accept the new velocity solution; if ΔE>0, then with probability e -(ΔE / T) >rand accepts the new velocity solution; where T represents the current temperature; rand represents a random number between 0 and 1; S107: Confirm whether the current condition temperature reaches the termination condition temperature; If so, the current new speed solution is used as the output optimal solution, that is, the ship speed corresponding to the minimum risk of ship collision; Otherwise, update the current condition temperature and repeat steps S102 to S106; The updating formula of the current condition temperature is: T'=T·cooling_rate; wherein T' represents the updated condition temperature; cooling_rate represents the temperature decay rate.

4. The method for autonomous collision avoidance of inland waterway vessels based on the improved speed barrier method according to claim 1, characterized in that: The encounter situation assessment strategy established in S3 based on the rules for avoiding collisions in inland waters and at sea is as follows: Define A as the relative position of the target ship relative to the own ship, and the range is [0, 2π); The angle between the own ship's sailing direction and the water flow direction is Alpha, ranging from [0, 2π); the angle between the target ship's sailing direction and the water flow direction is Beta, ranging from [0, 2π); Define and obtain the encounter situation area of the target ship relative to the own ship in the inland waterway, The encounter situation area includes the overtaking area, the encounter area, the intersection and give way area, the intersection and direct navigation area, the crossing area, the river fork area and the water area where the main / tributary rivers meet; Based on the relative bearing A, angle Alpha, and angle Beta in the encounter situation area, determine whether own vessel is a give-way vessel or a straight-ahead vessel. Specifically: If the target ship is in the overtaking zone: when A∈(5π / 8,11π / 8], the own ship is defined as the give-way ship, otherwise it is the straight-ahead ship; If the target ship is in the encounter zone: when A∈(0,π / 36]∪(71π / 36,2π], and Alpha∈(π / 2,3π / 2], then the own ship is defined as the give-way ship; otherwise, the own ship is defined as the straight-ahead ship; If the target ship is in the intersection give-way zone: when A∈(π / 36,4π / 9]∪(5π / 9,5π / 8], then the own ship is defined as the give-way ship, otherwise it is the straight-ahead ship; If the target ship is in the crossing and meeting direct navigation area: when A∈(14π / 9,71π / 36]∪(11π / 8,13π / 9], the ship is defined as a straight-going ship; otherwise, it is a give-way ship; If the target ship is in the crossing area: when A∈(4π / 9,5π / 9]∪(13π / 9,14π / 9], then the ship is defined as a straight-moving ship; otherwise, it is a give-way ship; If the target vessel is in the fork in the river: When Alpha∈(π / 2,3π / 2] and Beta∈(π / 2,3π / 2], or and If A∈(0,π], then the ship is defined as a give-way ship; if A∈(π,2π], then the ship is defined as a straight-moving ship; When Alpha∈(π / 2,3π / 2] and This defines this vessel as a give-way vessel; when And Beta∈(π / 2,3π / 2], then the ship is defined as a straight-moving ship; If the target ship is in the confluence of the main stream and tributary, when A∈(0,π], the ship is defined as a give-way ship; when A∈(π,2π], the ship is defined as a straight-moving ship.

5. The method for autonomous collision avoidance of inland waterway vessels based on the improved speed barrier method according to claim 1, characterized in that: The method for obtaining the ship collision risk according to the relative motion parameters in S3 specifically includes the following steps: S31: Obtain membership functions for solving DCPA membership, TCPA membership, two-ship distance membership, minimum approach factor membership, ship speed ratio membership, and minimum offshore distance membership respectively; The expression of the DCPA membership function is: Where: U(DCPA) represents the DCPA membership function; D s represents the radius of the ship's area; d1 and d2 represent intermediate parameters; k1 represents the design parameter determined by the visibility factor; k2 represents the design parameter determined by the current water conditions; f_min represents the minimum proximity factor between the target ship and the own ship; The expression of the TCPA membership function is: Where: U(TCPA) represents the TCPA membership function; t1 and t2 represent intermediate parameters; D1 represents the latest avoidance distance; D2 represents the safe distance for taking avoidance measures; v r Indicates the relative speed between the target ship and own ship; The expression of the membership function of the distance between the two ships is: Where: U(D) represents the membership function of the distance between the two ships; k3 represents the design parameter determined by human factors; DLA represents the latest steering distance of the ship; r represents the design parameter proposed based on the wide sea area; The expression of the minimum proximity factor membership function is: Where: U(f_min) represents the minimum proximity factor membership function; f_min represents the minimum proximity factor between the target ship and the own ship; The expression of the ship speed ratio membership function is: Where: U(K) represents the ship speed ratio membership function; K and W represent design parameters; C represents the ship collision angle and 0°≤C<180°; The expression of the offshore minimum distance membership function is: Where: U(Q) represents the minimum distance from shore membership function; Q represents the minimum distance from shore; d warn Indicates the latest avoidance distance between the ship and the shore; d safe Indicates the safe avoidance distance between the ship and the shore; S32 constructs a ship collision risk model based on the obtained DCPA membership function, TCPA membership function, two-ship distance membership function, minimum approach factor membership function, ship speed ratio membership function, and offshore minimum distance membership function; And the expression of the ship collision risk model is e=a DCPA U(DCPA)+a TCPA U(TCPA)+a D U(D)+a f_min U(f_min)+a K U(K)+a Q U(Q) Where: a DCPA ,a TCPA ,a D ,a f_min ,a K ,a Q They represent the weight coefficients of DCPA membership, TCPA membership, distance between two ships membership, minimum approach factor membership, ship speed ratio membership and minimum offshore distance membership respectively; e represents the risk of ship collision; Obtaining the ship collision risk according to the ship collision risk model; The system also divides the collision risk into different levels based on the set collision risk, and determines the type of situation between the own ship and the target ship according to the collision risk level; And the situation type represents imminent danger, dangerous situation and observation situation.

6. The method for autonomous collision avoidance of inland waterway vessels based on the improved speed barrier method according to claim 2, characterized in that: In S002, based on the elliptical ship domain model, the method for obtaining the speed barrier cone according to the current data information specifically includes: S0021: Set the geographic coordinates of this ship to Pos A =(x A ,y A ), ship speed The length of the ship is l, and the major axis of the ship is l A =3l, semi-minor axis w A =0.8l; Assume the geographic coordinates of the target ship is Pos B =(x B ,y B ), ship speed The length of the ship is lo, and the major axis of the ship is l B =3lo, semi-minor axis w B =0.8lo; S0022: Obtain the elliptical ship domain model of the own ship and the elliptical ship domain model of the target ship; And the expression of the elliptical ship domain model of the ship is Where: A1, B1, C1, D1, E1, F1 represent intermediate variables; Indicates the ellipse direction angle θ corresponding to the ship A Speed; x, y represent the equation parameters of the elliptical ship field model; The expression of the elliptical ship domain model of the target ship is: Where: A2, B2, C2, D2, E2, F2 represent intermediate variables; Indicates the ellipse direction angle θ corresponding to the ship B speed; S0023: Assume that the tangent equation of the elliptical ship domain model is y=kx+b, where k and b represent the parameters to be solved in the tangent equation; And bring it into the elliptical ship domain model of the own ship and the elliptical ship domain model of the target ship respectively, and solve k and b simultaneously to obtain several tangent equations to be solved; S0024: Obtain the angle α corresponding to each tangent line based on the slope of each solved tangent line equation i And α i =arctan(k i ); where α i represents the angle corresponding to the equation of the i-th tangent line; k i represents the slope of the i-th tangent line equation; Obtain the direction angle φ of the line connecting the center of the own ship and the target ship, and φ=arctan2(y B -y A ,x B -x A ), and according to the direction angle φ and angle α i Determine the boundary of the speed barrier cone to obtain the speed barrier cone; And the boundary expression of the speed barrier cone is Alpha=max(α i -φ),Beta=min(α i -f) Where: Alpha and Beta represent the two boundaries of the speed barrier cone.

Images