Water surface unmanned ship dynamic potential energy field obstacle avoidance method

Abstract

The invention provides a dynamic potential energy field obstacle avoidance method for an unmanned surface ship, and belongs to the technical field of unmanned ship obstacle avoidance path planning. Designing a universal inertia extension dynamics model, and combining the model with a digital map; setting an unmanned ship, and importing the designed kinetic model; various attributes of the unmanned ship are set in combination with the digital map, and early-stage preparation of an experiment is completed and the experiment is carried out; detecting the advancing state of the unmanned ship in the digital map in real time, and judging whether the unmanned ship arrives at the terminal point or not; on the basis of a randomly generated digital map, calculating an adaptive ring flow situation energy field acting force and an adaptive boundary dynamic potential energy field acting force; calculating a final resultant force, and calculating a forward direction and a step length by using the final resultant force; and the experiment is carried out circularly until the number of iterations is used up or the moving path of the unmanned ship is output.

Description

Technical Field

[0001] The present invention belongs to the technical field of obstacle avoidance path planning for unmanned ships, and more specifically, relates to a method for obstacle avoidance of a surface unmanned ship using a dynamic potential field. Background Art

[0002] The obstacle avoidance technology of surface unmanned ships (USV) is a core part of the autonomous navigation of unmanned systems, and its development is driven by demands such as marine resource development, environmental monitoring, and military applications. With the progress of artificial intelligence and sensor technology, obstacle avoidance technology has evolved from early rule-based avoidance to an intelligent decision-making system with multi-modal perception fusion, but still faces many challenges in complex dynamic water environments.

[0003] Relying on RGB cameras and traditional image processing algorithms (such as Canny edge detection and Hough transform), it is vulnerable to environmental influences such as changes in lighting, water surface reflection, and haze, and has a low recognition rate for semi-submerged obstacles (such as floating ice and fishing nets).

[0004] In rainy, foggy weather or under the interference of water surface wave reflection, the noise of point cloud data is significant, and the high-precision LiDAR is costly (for example, the unit price of Velodyne HDL-64E exceeds $50,000), making it difficult to deploy on a large scale.

[0005] Based on assumptions of a static environment, methods such as A* and DWA (Dynamic Window Approach) have severe path oscillations in scenarios where dynamic obstacles move frequently (such as port channels).

[0006] Sampling algorithms such as RRT* (Rapidly-Exploring Random Trees) have insufficient real-time performance in high-dimensional state spaces (6-DOF motion models), resulting in a high delay in obstacle avoidance responses.

[0007] Traditional fusion methods such as Kalman filtering lack an adaptive arbitration mechanism for conflicting perception data (such as an obstacle detected by radar but not recognized visually). Summary of the Invention

[0008] In view of this, the present invention is proposed to solve the problem of severe path oscillations existing in current existing obstacle avoidance methods for surface unmanned ships; at the same time, the present invention can also autonomously optimize the path and can more efficiently determine the driving path of the unmanned ship.

[0009] The present invention is implemented as follows: The present invention provides a method for obstacle avoidance of a surface unmanned ship using a dynamic potential field, wherein the method includes the following steps: S1 Generate a digital map in a random manner; S2 Design a general inertial extended dynamics model and combine the model with the digital map in step S1; S3 Set up the unmanned ship and import the dynamic model designed in step S2; combine with the digital map in step S1, set the various attributes of the unmanned ship, complete the preliminary preparation for the experiment and conduct the experiment; S4 Detect the moving state of the unmanned ship in the digital map in real time and determine whether the unmanned ship reaches the end point; S5 Based on the two results obtained from the judgment in step S4: the unmanned ship reaches the end point, at this time, organize the result of the moving path of the unmanned ship and output it; the unmanned ship does not reach the end point, and proceed to the next step; S6 Based on the randomly generated digital map in step S1, calculate the adaptive circulation potential energy field force and the adaptive boundary dynamic potential energy field force ; S7 Combine the , in step S6, calculate the final resultant force, and use the final resultant force to calculate the forward direction and step size; S8 Loop through steps S4 to S7 until the iteration count is exhausted or step S5 outputs the moving path of the unmanned ship.

[0010] Furthermore, the inertial extended dynamic model mentioned in step S2 is specifically as follows: ; In the above formula, , , are the displacement, velocity, and acceleration of the unmanned ship respectively, , are the resultant force acting on the unmanned ship and its virtual mass respectively, represents the discrete operation entering the th iteration, represents the set time for a single iteration, represents the inertial extended mapping function, is order difference operator.

[0011] Beneficial effects of adopting the above technical solution: In the traditional artificial potential field motion model, when conducting experiments in practice, there will be special positions. When the unmanned ship is in this special position, environmental disturbances cause the applied resultant force , resulting in the model losing its ability to move; the inertial expansion dynamic model proposed by the present invention enables the ship to quickly refer to historical information when dealing with complex dynamic environments and get out of the unstable equilibrium trap outside the expected target.

[0012] Further, the various properties mentioned in step S3 include: the virtual mass of the unmanned ship, the acceleration generated by the resultant force effect; dead zone detection is added to the speed and steering angle, where the speed is obtained based on the previous moment and finite historical information; straight line detection is added at the end of the journey. When the distance between the current position of the unmanned ship and the target point is less than 25% of the distance between the initial position and the target point, and there is no obstruction on the straight path, the unmanned ship abandons the inertial motion displacement and adopts a straight running method, directly connecting the current point to the end point.

[0013] Further, the specific pre - preparation content mentioned in step S3 includes: S31 introduces a virtual mass for the unmanned ship , representing the magnitude of virtual inertia; modifying the displacement directly caused by the resultant force to the second - derivative of the displacement generated, and calculating the candidate speed for this round of iteration based on acceleration, historical speed, and the number of iteration steps ; ; In the above formula, , respectively represent the weight of the current iteration information and the weight of the historical speed information, is the number of reference historical information; the parameter selection criterion is: incorporating the historical speed information reference with a finite inverse - proportional decay weight to make the actual motion smoother; S32 performs amplitude limiting processing on the actual speed magnitude of the unmanned ship: ; In the above formula, is the set speed upper limit value; By setting the speed upper limit, when the unmanned ship system faces a ramp disturbance input , the amplitude - limiting nonlinearity will ensure the actual rate ; S33 performs amplitude limiting processing on the actual heading change of the unmanned ship: ; In the above formula, is the angle between the heading of the unmanned ship and the positive direction of the - axis of the map, is the set upper limit value of the heading change, , respectively represent the components of the unmanned ship's speed along the - axis, - axis, , respectively represent - axis, - axis unit vectors; By setting the maximum value of the angle of the unmanned ship's heading change, it is avoided that the overall path of the unmanned ship deviates due to excessive correction during the autonomous correction process.

[0014] S34 According to the actual speed and the set time for a single iteration calculate the actual displacement and perform a collision check: ; Before the unmanned ship actually moves forward, it will check according to the water-ship-obstacle map. If the displacement of the current step will cause a collision, it will be adjusted cyclically until no collision occurs, is an adjustable parameter; S35 When approaching the end position, perform an additional straight-line connection detection; when the current position is less than 25% of the distance between the initial position and the target point and there is no obstruction, discard the inertial motion displacement and directly connect the current point and the target point.

[0015] Furthermore, the specific process of calculating the adaptive circulation potential energy field force in step S6 includes: S611 Take each obstacle on the map as the circulation source of the system, and make the radial repulsive force modified to an approximate annular repulsive force ; S612 Design a normal repulsive force , and this force corresponds to a rotational source-free field; S613 Determine the rotation description matrix of the normal potential field according to the heading of the unmanned ship and the relative position of the obstacle ; S614 Adaptively superimpose the normal repulsive force field onto the radial repulsive force field to obtain the resultant force after arrangement; S615 If the resultant force vector enters the obstacle interior under a certain step extension, trigger a detachment adjustment to reduce the normal repulsive force field, so that the adaptive circulation as a whole surrounds the obstacle without entering it Furthermore, the specific process of calculating the adaptive boundary dynamic potential energy field force in step S6 includes: S621 Select boundary points. Extract the circumscribed circle of the obstacle area and make a tangent line. Select the tangent line with a smaller included angle with the speed direction of the unmanned ship as the boundary line, and the intersection of the boundary line and the circle is the boundary point; When the unmanned ship approaches the boundary point, it shows an attracting effect, and when the unmanned ship moves away from the boundary point, it shows a repulsive effect. Calculate the potential field force according to the relative position relationship : ; In the above formula, is the boundary action intensity coefficient, , are the unmanned ship and the boundary point coordinates respectively, , , , are the bell curve adjustment parameters; the adjustment parameters are used to optimize , making the position and the force magnitude non-linear, and avoiding the unmanned ship falling into the local optimal position during the model simulation process; In S623, the gravitational effect of the final target point on the unmanned ship is replaced by multiple obstacle boundary points, that is, multi-waypoints in stages are adaptively generated; The gravitational force of the final target point in S624 :

[0016] In the above formula , are the linear action coefficients, is the repulsive force cut-off threshold; when the repulsive force exceeds the threshold, it means that the obstacle is adjacent, and the movement of the unmanned ship is dominated by the obstacle potential field.

[0017] The beneficial effects of the present invention are as follows: For the obstacle avoidance path planning of the unmanned ship on the water surface, an inertial extended travel control algorithm is adopted, so that the ship can quickly refer to historical information when dealing with complex dynamic environments and get out of the unstable equilibrium trap outside the expected target. When approaching the end position, additional line segment direct connection detection is performed, so that the unmanned ship can stop at the target point more smoothly.

[0018] The present invention adopts two generalized adaptive dynamic potential fields, which can show a unified force form for obstacles of different shapes and sizes, increasing the degree of freedom of artificial potential field design.

[0019] The solution proposed by the present invention can adapt to the potential energy field of the circulating flow state. Through the condition that the divergence is zero, a rotational and source-free normal repulsive force is designed, and the potential field circulation effect is constructed by combining the centripetal repulsive force, making it easier for the unmanned ship to bypass obstacles and enhancing the robustness to environmental changes.

[0020] The solution proposed by the present invention can adapt to the dynamic potential field of the boundary. Through the nearest neighbor polygon corner points or the circular proximal tangent, an adaptive attraction / repulsion force field is constructed in combination with the velocity direction, further enhancing the obstacle avoidance ability.

[0021] The experimental results show that under the decision of the method of the present invention, the driving path of the unmanned ship reaches an optimal level both in terms of efficiency and the ability to cope with complex obstacle environments. Description of the Drawings

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and thus should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0023] Figure 1 It is the overall algorithm execution flowchart; Figure 2 It is a randomly generated water - ship - obstacle digital map; Figure 3 It is the comparison chart of the effect before the improvement of the artificial potential field; Figure 4 It is the comparison chart of the effect after the improvement of the artificial potential field; Figure 5 It is the comparison chart of the actual path - finding performance effects of each optimization algorithm; Figure 6 It is the comparison chart of the quantitative indicators of each optimization algorithm; In the drawings, the content represented by each English and label is as follows: PathPlanningDigitalMap: Path planning digital map; start: starting point; goal: ending point; Scenario: solution; A^*: Astart algorithm; RRT: Rapidly - exploring Random Tree algorithm; RRT^*: Improved Rapidly - exploring Random Tree algorithm; DPF: The algorithm proposed in this solution. Specific implementation manners

[0024] In order to make the objectives, technical solutions and advantages of the present invention more clear and understandable, the following further details the present invention in conjunction with the drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0025] As Figure 1 —6 shows, the present invention provides a method for obstacle avoidance of a surface unmanned ship using a dynamic potential field.

[0026] Among them, Figure 1 It is the planning flowchart of the present invention, which successively realizes the modeling of the unmanned ship, the conventional attractive - repulsive potential field, the adaptive circulation - flow potential field, the adaptive - boundary dynamic potential field, and the composite - field superposition drives the unmanned ship to reach the target position in stages.

[0027] Figure 2It is a randomly generated digital map of water - ship - obstacle. The circular shadows represent the obstacle areas, and the starting and ending points are marked with asterisks. The virtual movement of the unmanned ship is carried out on the digital map, and through this algorithm, it is ensured that the unmanned ship finds a feasible route and avoids obstacles on the way.

[0028] Figure 3 , Figure 4 It is a comparison chart of the planning before and after the improvement of the artificial potential field. Figure 3 For the situation before improvement. Figure 4 For the situation after improvement; the original method is prone to falling into local optimal positions and shows path oscillation on the map, while the improved method can more effectively avoid dynamic obstacles. Through Figure 3 it can be seen that the local optimal solution will cause misguidance in path selection, resulting in the unmanned ship circling among multiple dynamic obstacles.

[0029] Figure 4 It is a comparison of the planning effects of the method proposed in this invention with three commonly used planning algorithms, A*, RRT, and RRT*. It intuitively shows the advantages of this method. The path planned using this method is smoother and has less redundancy.

[0030] Figure 5 Then, from a quantitative perspective, it gives the iteration step lengths and potential energy decline processes of each algorithm from the starting point to the ending point, and compares the advantages and disadvantages of the algorithms with specific numerical values.

[0031] Specifically, this method includes the following processes: S1 Generate a digital map in a random manner and use this map as the basis for subsequent actual simulations; S2 Design a general inertial extended dynamics model and apply the model to the random digital map in step S1; S3 Make corresponding adjustments to the dynamics model designed in step S2 in combination with the movement of the unmanned ship, design the various attributes of the unmanned ship, complete the preparations before the experiment and conduct the experiment; S4 Detect the movement state of the unmanned ship in the digital map and judge whether it reaches the ending point; S5 Based on the results obtained in step S4, there are two situations: the unmanned ship reaches the ending point, at this time, organize and output the movement route of the unmanned ship, and the process is completed; the unmanned ship does not reach the ending point, at this time, proceed to the next step; S6 Based on the digital map generated in step S1, calculate the acting force of the adaptive circulation potential energy field; at the same time, based on the digital map generated in step S1, calculate the acting force of the adaptive boundary dynamic potential energy field ; S7 Combine the , in step S6, calculate the final resultant force, convert the resultant force into the second - order derivative and input it into the model, and calculate the forward direction and step length; Step S4 - S7 is iteratively repeated in S8 until the result that the unmanned ship reaches the end point is obtained in S5, and the experimental result is output, and the test ends.

[0032] Optionally, in the above technical solution, the dynamic model adopted by the conventional artificial potential field in S2 step is as follows: (1) Wherein, represents the discrete operation entering the th iteration, represents the displacement of the unmanned ship, represents the set forward distance for a single-step iteration, , , respectively represent the resultant force, repulsive force, and attractive force received by the unmanned ship.

[0033] The solution proposed by the present invention is applied to the movement of an unmanned ship in a dynamic and complex environment; when the above artificial potential field motion model is used for simulation, there are special positions, and at these positions, environmental disturbances cause the applied resultant force , resulting in the original model losing its ability to move, and at this time the model will lose its function.

[0034] Based on the above problems, the present invention proposes an inertia-expanded second-order dynamic model: (2) In formula (2), , , are respectively the displacement, velocity, and acceleration of the unmanned ship, , are respectively the resultant force received by the unmanned ship and its virtual mass, represents the discrete operation entering the th iteration, represents the set time for a single iteration, represents the inertia-expanded mapping function, is order difference operator.

[0035] Optionally, in the above technical solution, the various attributes mentioned in S3 step include: the virtual mass of the unmanned ship, the acceleration generated by the resultant force effect; the velocity and steering angle are added with dead zone detection, wherein the velocity is obtained according to the previous moment and finite historical information; a straight line detection is added near the end point.

[0036] Optionally, in the above technical solution, the specific content of the pre-test preparation in S3 step includes: S31 introduces a virtual mass for the unmanned ship , representing the magnitude of virtual inertia. Modify the displacement directly caused by the resultant force to the second derivative of the generated displacement, and calculate the candidate speed for this iteration based on acceleration, historical speed, and the number of iterations. : (3) In Equation (3), , respectively represent the weight of the current iteration information and the weight of the historical speed information, is the number of historical information references. The parameter selection criterion is: incorporate the historical speed information reference with a finite inverse proportional decay weight to make the actual motion smoother.

[0037] S32 performs a limit processing on the actual speed magnitude of the unmanned ship: (4) In Equation (4), is the set speed upper limit value. When the unmanned ship system faces a ramp disturbance input , the limit nonlinearity will ensure the actual rate .

[0038] S33 performs a limit processing on the actual heading change of the unmanned ship: (5) In Equation (5), is the angle between the heading of the unmanned ship and the positive x-axis of the map, is the set upper limit value of the heading change, , respectively represent the magnitudes of the components of the unmanned ship's speed along the x-axis and y-axis, , respectively represent the unit vectors in the x-axis and y-axis directions.

[0039] S34 calculates the actual displacement based on the actual speed and the set time for a single iteration and performs a collision check: (6) The unmanned ship will check according to the water-ship-obstacle map before actual advancement. If the displacement in the current step will cause a collision, will be adjusted cyclically until no collision occurs, is an adjustable parameter.

[0040] S35 performs an additional straight-line connection detection when approaching the end position. If the current position is less than a certain distance (25% of the total distance) from the target point and there is no obstruction, the inertial motion displacement will be discarded, and the current point and the target point will be directly connected.

[0041] Optionally, in the above technical solution, step S6 calculates the acting force of the adaptive circulation potential energy field. The specific process includes: S611 determines that the circulation source is each obstacle on the map, and the goal is to modify the radial repulsive force into an approximate annular repulsive force .

[0042] The radial repulsive force is defined as: (7) In formula (7), , respectively represent the coordinates of the unmanned ship and the obstacle on the map, is the effective action range of the potential field generated by the obstacle, is the intensity coefficient of the potential field action.

[0043] S612 designs the normal repulsive acting force , and this acting force should correspond to a rotational source-free field: (8) In formula (8), represents the normal repulsive potential energy field with zero divergence, is the intensity coefficient of the potential field action, is the rotation direction description matrix of the potential field.

[0044] S613 determines the rotation direction description matrix of the normal potential field according to the heading of the unmanned ship and the relative position of the obstacle .

[0045] The angle formed by the heading of the unmanned ship and the positive x-axis direction is , and the angle formed by the connection line between the unmanned ship and the obstacle and the positive x-axis direction is . The matrix is determined according to the difference between the two: (9) is a custom extended sign function, expressed as: .

[0046] S614 adaptively superimposes the normal repulsive force field on the radial repulsive force field: (10) In formula (10), is the adaptive adjustment coefficient. If the resultant force vector enters the obstacle under a certain step extension, the detachment adjustment is triggered to reduce the normal repulsive force field, so that the adaptive circulation as a whole surrounds the obstacle without entering the obstacle. It is related to the forward distance of the collision and the detection step size.

[0047] Optionally, in the above technical solution, the specific process of calculating the force of the adaptive boundary dynamic potential field in step S6 also includes: S621 Boundary point selection: Extract the circumcircle of the obstacle area and make a tangent line. Select the tangent line with a smaller included angle with the direction of the unmanned ship's speed as the boundary line. The intersection point of the boundary line and the circle is the boundary point. The included angle from the boundary line to the speed direction is the deflection angle , and its size affects the nature of the potential field action.

[0048] The boundary points generate an attractive / repulsive potential field with variable properties for the unmanned ship. Specifically, when the unmanned ship approaches the boundary point, it shows an attractive effect, and when the unmanned ship moves away from the boundary point, it shows a repulsive effect. The potential field action is reflected as force , and it is calculated according to the relative position relationship: (11) In formula (11), is the boundary action intensity coefficient, , are the coordinates of the unmanned ship and the boundary point respectively, is the adjustment parameter of the bell-shaped curve, which makes the position and the magnitude of the force show non-linearity and is less likely to fall into the local optimal position.

[0049] The gravitational action of the final target point on the unmanned ship is replaced by multiple obstacle boundary points, that is, adaptively generate multi-waypoints in stages. The gravitational force of the final target point: (12) In formula (12), is the linear action coefficient, is the repulsive force cut-off threshold. When the repulsive force exceeds the threshold, it means that the unmanned ship is adjacent to the obstacle, and the movement of the unmanned ship is dominated by the obstacle potential field.

[0050] The above is only the preferred embodiment of the present invention and is not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for obstacle avoidance of an unmanned surface vessel based on dynamic potential energy field, characterized in that This method includes the following steps: S1 Generate a digital map in a random manner; S2 Design a general inertial extended dynamics model and combine the model with the digital map in step S1; S3 Set up an unmanned ship, import the dynamics model designed in step S2; combine with the digital map in step S1, set various attributes of the unmanned ship, complete the pre-experiment preparation and conduct the experiment; S4 Real-time detect the traveling state of the unmanned ship in the digital map and judge whether the unmanned ship reaches the end point; S5 Based on the two results obtained from the judgment in step S4: the unmanned ship reaches the end point, at this time, organize the result of the moving path of the unmanned ship and output it; the unmanned ship does not reach the end point, perform the next operation; Based on the randomly generated digital map in the S1 step, S6 calculates the acting force of the adaptive loop flow potential energy field and the acting force of the adaptive boundary dynamic potential energy field ; S7 combines with those in step S6 and to calculate the final resultant force, and calculate the forward direction and step length using the final resultant force; S8 Loop through steps S4 to S7 until the iteration count is exhausted or the moving path of the unmanned ship is output in step S5.

2. The dynamic potential energy field obstacle avoidance method for an unmanned surface vessel according to claim 1, characterized in that The inertial extended dynamics model mentioned in step S2 is specifically as follows: ； In the above formula, , , are the displacement, velocity, and acceleration of the unmanned ship respectively, , are the resultant force and its virtual mass received by the unmanned ship respectively, represents that the discrete operation enters the th iteration, represents the set time for a single iteration, represents the inertial extended mapping function, is order difference operator.

3. A method for obstacle avoidance of a surface unmanned ship based on a dynamic potential energy field according to claim 1, characterized in that, The various attributes mentioned in step S3 include: the virtual mass of the unmanned ship, the acceleration generated by the resultant force effect; the speed and the steering angle are added with dead zone detection, where the speed is obtained according to the previous moment and limited historical information.

4. The dynamic potential energy field obstacle avoidance method for an unmanned surface vessel according to claim 3, wherein The specific content of the pre-experiment preparation mentioned in step S3 is as follows: S31 introduces a virtual mass for the unmanned ship to characterize the magnitude of virtual inertia; modifies the displacement directly caused by the resultant force to the second derivative of the displacement generated, and calculates the candidate speed for this round of iteration based on the acceleration, historical speed, and number of iteration steps. ; ； In the above formula, , respectively represent the weight of the current iteration information and the weight of the historical velocity information, is the number of historical information references; the parameter selection criterion is: incorporating the reference of the historical velocity information with a limited inverse proportional decay weight to make the actual movement smoother; S32 Perform amplitude limiting processing on the actual speed magnitude of the unmanned ship; S33 Perform amplitude limiting processing on the actual heading change of the unmanned ship; S34 According to the actual speed and the set time for a single iteration calculate the actual displacement and perform a collision check; before the unmanned boat actually moves forward, it will check according to the water-boat-obstacle map. If the displacement in the current step will cause a collision, it will be adjusted cyclically until no collision occurs is an adjustable parameter; S35 When approaching the end point, perform additional straight-line connection detection; when the current position is less than 25% of the distance between the initial position and the target point and there is no obstruction, discard the inertial motion displacement and directly connect the current point and the target point.

5. A method for obstacle avoidance of an unmanned surface vessel based on a dynamic potential energy field according to claim 1, characterized in that, The specific process of calculating the acting force of the adaptive loop flow potential energy field in step S6 is as follows: S611 uses each obstacle on the map as a circulation source of the system, making the radial repulsive force modified to an approximately annular repulsive force ; S612 Design normal repulsive force , and this force corresponds to a rotational source-free field; S613 Determine the rotation description matrix of the normal potential field according to the heading of the unmanned ship and the relative position of the obstacle ; The normal repulsive force field is adaptively superimposed on the radial repulsive force field , and the resultant force is obtained by sorting. S615 If the resultant force vector enters the obstacle interior under a certain step extension, trigger the disengagement adjustment, reduce the normal repulsive force field, so that the adaptive circulation as a whole surrounds the obstacle without entering the obstacle.

6. The dynamic potential energy field obstacle avoidance method for a surface unmanned ship according to claim 1, characterized in that, The specific process of calculating the force of the adaptive boundary dynamic potential energy field in step S6 is as follows: S621 Select boundary points, extract the circumscribed circle of the obstacle area and make a tangent line, select the tangent line with a smaller included angle with the speed direction of the unmanned ship as the boundary line, and the intersection point of the boundary line and the circle is the boundary point; When the unmanned ship approaches the boundary point, it shows an attracting effect, and when it moves away from the boundary point, it shows a repulsive effect. The potential field force is calculated according to the relative position relationship. : ； In the above formula, is the boundary action intensity coefficient, , are the coordinates of the unmanned ship and the boundary point respectively, , , , are the bell curve adjustment parameters; the adjustment parameters are used to optimize to make the position and the force magnitude show non-linearity, and to prevent the unmanned ship from falling into a locally optimal position during the model simulation process; S623 The gravitational effect of the final target point on the unmanned ship is replaced by multiple obstacle boundary points, that is, adaptively generate multi-waypoints in stages; Gravitational force of the S624 final target point : ； The above formula , are linear action coefficients, is the repulsive force cut-off threshold; when the repulsive force exceeds the threshold, it indicates that the unmanned ship is in close proximity to an obstacle, and the movement of the unmanned ship is dominated by the obstacle potential field.

Images