A multi-scene autonomous obstacle avoidance method for a water unmanned surface vehicle based on perception information

Abstract

The application relates to the field of autonomous obstacle avoidance of a water unmanned surface vehicle, in particular to a multi-scene autonomous obstacle avoidance method for a water unmanned surface vehicle based on sensing information. The method comprises the following steps: obtaining unmanned vehicle navigation information; using the sensing device of the unmanned vehicle to sense the target to be avoided; respectively optimizing the sensing information of each target to be avoided; obtaining the type of each target to be avoided according to the sensing information of each target to be avoided after optimization; constructing a threat information list based on the sensing information and the type of all targets to be avoided; traversing the threat target information list, if the distance between the current threat target and the unmanned vehicle reaches a preset threat distance threshold, then dividing the obstacle avoidance scene according to the type of the current threat target and the number of threat targets in the threat target information list, and controlling the unmanned vehicle to autonomously avoid the threat target based on the obstacle avoidance scene. In this way, the safety guarantee and the task completion degree of the unmanned vehicle in the navigation task can be improved.

Description

Technical Field

[0001] This invention generally relates to the field of autonomous obstacle avoidance for unmanned surface vessels, and more specifically, to a multi-scenario autonomous obstacle avoidance method for unmanned surface vessels based on perception information. Background Technology

[0002] Unmanned surface vessels (USVs), as unmanned surface platforms, possess advantages such as rapid deployment, integrated reconnaissance and strike capabilities, strong payload capacity, and optimal relay platform status. In recent years, USV technology has received significant attention, and with increased funding and effort, various aspects of USV technology have improved. As a relay intelligent control center, the control and decision-making capabilities of USVs are crucial for executing more complex tasks and should therefore be a key research focus.

[0003] However, unmanned surface vessels (USVs) often face complex environments during missions, potentially limited by blind spots and communication delays. Under these constraints, shore-based / ship-based command centers cannot issue instructions in a timely manner in case of emergencies. Given that the safety of the USV itself is a necessary and sufficient condition for mission success, autonomous obstacle avoidance capability has become an essential basic function and a crucial indicator for evaluating the intelligence level of USVs.

[0004] Current autonomous obstacle avoidance technologies for unmanned surface vessels (USVs) are mostly limited to application verification for single static targets, single dynamic targets, and single large targets. This lack of targeted autonomous decision-making design in a global scenario fails to intuitively display the environmental situation of the USV and cannot perform timely obstacle avoidance based on different target types within the threat time domain. As a result, USVs still have certain uncertainties and dangers in real-world applications.

[0005] In conclusion, simple single-obstacle avoidance methods are insufficient to cope with the complex situations faced by unmanned surface vessels (USVs) during mission execution, and the autonomous obstacle avoidance function of USVs should not be limited to simple scenarios. Summary of the Invention

[0006] According to the present invention, a multi-scenario autonomous obstacle avoidance method for unmanned surface vessels (USVs) based on perception information is provided. This method completes relevant tasks while ensuring the safety of the USV; by improving the comprehensiveness and stability of the USV's autonomous obstacle avoidance function, the safety of the USV is no longer a limiting factor for its task completion.

[0007] The method includes: Obtain navigation information of unmanned surface vessels; During the mission, the unmanned surface vessel (USV) uses its sensing equipment to perceive the target to be avoided. The perception information of each target to be avoided is optimized; then, based on the optimized perception information of each target to be avoided, the type of each target to be avoided is obtained. Based on the perceived information and type of all targets to be avoided, a threat information list is constructed; The system iterates through the threat target information list. If the distance between the current threat target and the unmanned surface vessel (USV) reaches a preset threat distance threshold, then it divides the obstacle avoidance scenario based on the information of the current threat target and the number of threat targets in the threat target information list. Based on the obstacle avoidance scenario, it controls the USV to autonomously avoid the threat target.

[0008] Furthermore, the type of the target to be avoided includes: target type and target threat type; The target type is obtained by classifying the type based on the optimized perception information of each target; The target threat type is obtained by identifying the threat type from the optimized perception information of each target.

[0009] Furthermore, based on the obstacle avoidance scenario, the unmanned surface vessel (USV) is controlled to autonomously avoid threatening targets, including: When the obstacle avoidance scenario is a single static target autonomous obstacle avoidance scenario, the target heading control quantity of the unmanned surface vessel is solved using a single static target obstacle avoidance algorithm; When the obstacle avoidance scenario is a single large target autonomous obstacle avoidance scenario, the target heading control quantity of the unmanned surface vessel is solved by using the single large target obstacle avoidance algorithm; When the obstacle avoidance scenario is a single dynamic target autonomous obstacle avoidance scenario, the target heading control quantity of the unmanned surface vessel is solved by using a single dynamic target obstacle avoidance algorithm; When the obstacle avoidance scenario is an autonomous obstacle avoidance scenario with or without large static targets, the target heading control quantity of the unmanned surface vessel is solved by using the obstacle avoidance algorithm with or without large targets; When the obstacle avoidance scenario is a multi-dynamic target autonomous obstacle avoidance scenario, the multi-dynamic target obstacle avoidance algorithm is used to solve the target heading control quantity of the unmanned surface vessel; When the obstacle avoidance scenario is an autonomous obstacle avoidance scenario with mixed static and dynamic targets, the target heading control quantity of the unmanned surface vessel is solved by using the static and dynamic target obstacle avoidance algorithm. Construct the obstacle avoidance termination condition; The unmanned surface vessel (USV) is controlled to perform autonomous obstacle avoidance based on the target heading control parameters and obstacle avoidance termination conditions.

[0010] Furthermore, the single large target obstacle avoidance algorithm includes: All threat targets with the same target ID as the current threat target in the threat target information list are considered as large targets; The large target is subjected to an expansion process; The results of the puffing process of continuous targets are calculated to obtain the tangent angle of the continuous target contour, the distance between the unmanned surface vessel and the target contour, and the distance between the unmanned surface vessel and the continuous targets. The target heading control quantity of the unmanned surface vessel (USV) is estimated based on the USV's heading, the USV's preset navigation route, the tangent angle of the continuous target contour, the distance between the USV and the target contour, and the distance between the USV and the continuous targets. Among them, the unmanned surface vessel's heading and preset navigation route are considered unmanned surface vessel navigation information.

[0011] Furthermore, the single dynamic target obstacle avoidance algorithm includes: The motion state of the current threatening target is estimated in real time based on the least squares method to obtain the target's speed. Set a threat distance threshold based on the unmanned surface vessel's speed and the speed of the threat target; If the distance between the current threat target and the unmanned surface vessel (USV) reaches the threat distance threshold, the USV target heading control quantity is calculated based on the single static target obstacle avoidance algorithm. Among them, the unmanned surface vessel's (USV) speed is part of its navigation information.

[0012] Furthermore, the single static target obstacle avoidance algorithm includes: Exaggerate the current threat targets; The results of the puffing process are used to model the target environment using the smallest circumscribed circle to obtain the target environment; The target environment is analyzed using the tangent method; The target environment analysis results are dynamically calculated to obtain the target heading control quantity of the unmanned surface vessel.

[0013] Furthermore, the obstacle avoidance algorithm with / without large targets includes: Exaggerate the current threat targets; The expanded results of the current threat targets are rasterized to obtain a global raster map of multiple static targets; A safe endpoint is set based on the unmanned surface vessel's preset route and the current location of the threat target; Based on the global grid map of the multiple static targets and the safe endpoint, autonomous obstacle avoidance path planning for the multiple static targets is performed to obtain the paths for the multiple static targets; The paths of the multiple static targets are smoothed to obtain the target heading control quantity of the unmanned surface vessel; Among them, the preset route of the unmanned surface vessel (USV) belongs to the USV navigation information, and the current threat target location belongs to the optimized perception information of the current threat target.

[0014] Furthermore, the multi-dynamic target obstacle avoidance algorithm includes: Extract the motion state of the current threat target from the optimized perception information of the current threat target; The least squares method is used to estimate the motion state of the current threatening target; Set the delay time based on the motion estimation results; The threat target area is obtained based on the delay time, and the threat target area is materialized into an obstacle area. The obstacle area is rasterized to obtain a global raster map of multiple dynamic targets; Based on the global grid map of the multi-dynamic targets, autonomous obstacle avoidance path planning for the multi-dynamic targets is performed to obtain the multi-dynamic target paths; The paths of the multiple dynamic targets are smoothed to obtain the target heading control quantity of the unmanned surface vessel.

[0015] Furthermore, the static / dynamic target obstacle avoidance algorithm includes: Based on the target type of the current threat target, the current threat target is expanded; The expanded result of the current threat target is rasterized to obtain a global raster map of static and dynamic targets; Based on the global grid map of static and dynamic targets, autonomous obstacle avoidance path planning for static and dynamic targets is performed to obtain the paths for static and dynamic targets. The static and dynamic target paths are smoothed to obtain the unmanned surface vessel target heading control quantity.

[0016] Furthermore, the obstacle avoidance termination condition includes: When the target is a single static target or a single dynamic target, the obstacle avoidance termination condition is that the Y-axis coordinate of the unmanned surface vessel in the Cartesian judgment coordinate system is greater than a preset coordinate threshold, the minimum turning radius path does not intersect with the puffed target, and the distance between the target and the unmanned surface vessel is greater than a preset safe distance threshold. When the target is a single large target, the obstacle avoidance termination condition is that the path tangent intersects with the preset route of the unmanned surface vessel, the minimum turning radius path does not intersect with the puffed target, and the minimum distance between the unmanned surface vessel and the edge of the target is greater than the preset safe distance threshold. The rectangular coordinate system is constructed with the target center as the origin and the direction from the start to the end of the unmanned surface vessel's preset route as the positive Y-axis. The expanded target is obtained by expanding the target. The minimum turning radius path is obtained based on the unmanned surface vessel's speed. The path tangent is obtained by drawing a tangent from the current position of the unmanned surface vessel towards the expanded target in the forward direction. The preset route and speed of the unmanned surface vessel are part of the unmanned surface vessel's navigation information.

[0017] Compared with the prior art, the present invention has the following beneficial technical effects: The method of this invention overcomes the limitations of existing technologies that are only applicable to obstacle avoidance of single static, dynamic, and single large targets. It can cope with complex mixed scenarios of different targets and solves the problem of lack of targeted autonomous decision-making in global scenarios. At the same time, continuous autonomous obstacle avoidance control ensures the safety of the unmanned surface vessel in communication-restricted or emergency situations, reduces the uncertainty and danger in practical applications, and eliminates the constraint of safety on mission completion while improving the comprehensiveness and stability of the obstacle avoidance function of the unmanned surface vessel.

[0018] It should be understood that the description in the Summary of the Invention is not intended to limit the key or essential features of the embodiments of the present invention, nor is it intended to restrict the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0019] The above and other features, advantages, and aspects of the various embodiments of the present invention will become more apparent from the accompanying drawings and the following detailed description. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein: Figure 1 A flowchart of a multi-scenario autonomous obstacle avoidance method for unmanned surface vessels based on perception information according to an embodiment of the present invention is shown; Figure 2 A flowchart illustrating the autonomous obstacle avoidance control of an unmanned surface vessel according to the present invention is shown; Figure 3 A flowchart of the single static target obstacle avoidance algorithm according to the present invention is shown; Figure 4 A flowchart of the single large target obstacle avoidance algorithm according to the present invention is shown; Figure 5 A flowchart of the single dynamic target obstacle avoidance algorithm according to the present invention is shown; Figure 6 A flowchart of the obstacle avoidance algorithm with / without large targets according to the present invention is shown; Figure 7 A flowchart of the multi-dynamic target obstacle avoidance algorithm according to the present invention is shown; Figure 8 A flowchart of the static and dynamic target obstacle avoidance algorithm according to the present invention is shown. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0022] In this invention, firstly, an autonomous obstacle avoidance function thread is established, which runs continuously during the unmanned surface vessel's (USV) mission. Within this thread, the input target information is first processed to remove outliers and optimize data quality. Then, the input targets are categorized by type and threat type. A threat target information list is constructed and updated. When the distance between a threat target and the USV reaches a threat distance threshold, an obstacle avoidance scenario is defined based on the current threat target situation. The corresponding obstacle avoidance algorithm for that scenario is then invoked to perform autonomous obstacle avoidance actions. The system continuously monitors whether the obstacle avoidance termination criterion is met. Upon completion of obstacle avoidance, the threat target information list is updated, and the USV autonomously returns to its preset route to continue other tasks. This approach improves the safety and mission completion rate of the USV during navigation missions.

[0023] Figure 1 A flowchart of a multi-scenario autonomous obstacle avoidance method for unmanned surface vessels based on perception information, according to an embodiment of the present invention, is shown.

[0024] The method includes S101~S105: S101. Obtain navigation information of unmanned surface vessels.

[0025] In this embodiment, the unmanned surface vessel (USV) navigation information includes: USV heading, preset USV route, USV speed, current USV position, and target point position. Target perception information consists of radar, optoelectronic equipment, and other perception information or data fusion results under stable USV platform conditions, including: target center position, target ID, target heading (moving target), target speed, target perception area, distance and angle between the USV and the target's left and right edge tangent points (for large targets), etc.

[0026] S102. During the unmanned surface vessel's mission, the vessel's sensing equipment is used to perceive the target to be avoided.

[0027] Specifically, the sensing equipment of unmanned surface vessels includes radar, optoelectronic devices, and all other equipment capable of detecting obstacle targets (i.e., targets to be avoided), and recording the sensing information of the obstacle targets.

[0028] S103. Optimize the perception information of each target separately, and then obtain the type of each target to be avoided based on the optimized perception information of each target to be avoided.

[0029] Specifically, data optimization includes: relative position calculation, coordinate transformation, outlier removal, and filtering.

[0030] In this embodiment, the type of the target to be avoided includes: target type and target threat type. The target type is obtained by classifying the type of each target based on its optimized perception information; the target threat type is obtained by labeling the threat type of each target based on its optimized perception information.

[0031] Specifically, the type classification refers to dividing targets into two types—discrete targets and continuous targets—by setting a speed threshold (which can be flexibly set according to the actual situation of the unmanned surface vessel). Continuous targets refer to large, non-moving targets such as shorelines, islands, reefs, and mountains; discrete targets refer to small surface targets such as ships. When the target speed is below the set speed threshold, the target is determined to be a continuous target or a discrete target in a stationary state; when the target speed is not lower than the set speed threshold, the target is determined to be a small, discrete target moving on the surface.

[0032] Specifically, threat type identification refers to classifying the target to be avoided as a static threat target or a dynamic threat target: when the target speed is less than 0.1 m / s, it is classified as a static threat target, and when the target speed is not less than 0.1 m / s, it is classified as a dynamic threat target.

[0033] S104. Based on the perception information and type of all targets to be avoided, construct a threat information list.

[0034] In this embodiment, the process of constructing the threat target information list is as follows: (1) When the target type of the target is a continuous target, the continuous threat judgment is made based on the current target's optimized perception information and the unmanned surface vessel's navigation information; if the continuous threat judgment result is that a collision will occur, the current target's optimized perception information, target type and target threat type are used as continuous threat target information.

[0035] Specifically, continuous threat assessment includes: 1) Construct an unmanned surface vessel coordinate system based on the unmanned surface vessel navigation information.

[0036] 2) Based on the optimized perception information of the current target, the navigation information of the unmanned surface vessel (USV), and the coordinate system of the USV, calculate the target's azimuth and distance. The specific process is as follows: 1. Based on the position information in the unmanned surface vessel's navigation information and the target position information in the optimized perception information of the current target, the straight-line distance between the two is calculated to obtain the target distance.

[0037] 2. Using the current position of the unmanned surface vessel as the origin of the coordinate system, calculate the target's orientation in the geodetic coordinate system.

[0038] 3. Convert the target orientation in the geodetic coordinate system to the target orientation in the unmanned surface vessel coordinate system.

[0039] 3) Based on the unmanned surface vessel's navigation information, the motion trend of the unmanned surface vessel is obtained. The specific process is as follows: 1. Based on the current mission of the unmanned surface vessel (USV), range constraints are imposed on the estimated USV speed and heading.

[0040] 2. By taking the average speed and heading of the unmanned surface vessel (USV) over a short period, a simple prediction of the USV's motion trend over a future time period is made. The USV's speed and heading are calculated after range constraints.

[0041] 4) Perform continuous target collision detection based on the target orientation, target distance, and unmanned surface vessel movement trend; and use the results of continuous target collision detection as the results of continuous threat assessment.

[0042] Specifically, continuous target collision detection includes: 1. Check the relative bearing and distance of the target to the unmanned surface vessel in the coordinate system.

[0043] 2. Calculate the straight-line distance between the unmanned surface vessel's position and the target position at each time point.

[0044] 3. Set a distance threshold. If there is a straight-line distance value lower than the distance threshold, it is determined that the unmanned surface vessel will collide with the target; otherwise, it is determined that the unmanned surface vessel will not collide with the target.

[0045] (2) When the target type is a discrete target, the discrete threat is judged based on the current target's optimized perception information and the unmanned surface vessel's navigation information; if the discrete threat judgment result is that a collision will occur, the current target's optimized perception information, target type and target threat type are used as discrete threat target information.

[0046] Specifically, discrete threat assessment includes: storing the position, heading, and velocity values ​​of a single target; deriving the future motion pose of the current target using the least squares fitting method; and determining whether there is a collision risk by combining the current heading and velocity of the unmanned surface vessel with the current navigating conditions.

[0047] (3) Generate a threat target information list, which includes continuous threat target information and discrete threat target information, as well as their corresponding target IDs. Among them, all continuous threat target information belonging to the same continuous target has the same target ID.

[0048] In this embodiment, if a target ID in the threat target information list is lost for a certain period of time, it is deleted from the threat target information list; if the target type of a certain ID changes, its previously stored information is transferred to the corresponding list of the current target type, thus completing the construction and updating of the threat target information list.

[0049] S105. Traverse the threat target information list. If the distance between the current threat target and the unmanned surface vessel (USV) reaches a preset threat distance threshold, then divide the obstacle avoidance scenario based on the information of the current threat target and the number of threat targets in the threat target information list. Based on the obstacle avoidance scenario, control the USV to autonomously avoid the threat target. The information of the current threat target includes target ID, target type, target threat category, target center position, target heading, target speed, target sensing area, distance and angle between the USV and the left and right edge tangent points of the target (for large targets), etc.

[0050] In this embodiment, the obstacle avoidance scenarios include: autonomous obstacle avoidance scenario for a single static target, autonomous obstacle avoidance scenario for a single large target, autonomous obstacle avoidance scenario for a single dynamic target, autonomous obstacle avoidance scenario for multiple static targets with / without large targets, autonomous obstacle avoidance scenario for multiple dynamic targets, and autonomous obstacle avoidance scenario for a mixture of static and dynamic targets.

[0051] Specifically, the single static target autonomous obstacle avoidance scenario refers to the presence of only one fixed obstacle in the waters where the unmanned surface vessel (USV) is navigating; the single large target autonomous obstacle avoidance scenario refers to the presence of a stationary or slowly changing obstacle in the waters where the USV's geometric dimensions are significantly larger than the USV itself; the single dynamic target autonomous obstacle avoidance scenario refers to the presence of an obstacle in the waters where the USV has the ability to move (its position changes over time); the multi-static target autonomous obstacle avoidance scenario with / without large targets refers to the presence of multiple fixed obstacles in the waters where the USV is navigating, which may or may not include large obstacles, forming a complex static obstacle environment; the multi-dynamic target autonomous obstacle avoidance scenario refers to the presence of multiple obstacles moving simultaneously in the waters where the USV is navigating; and the mixed static and dynamic target autonomous obstacle avoidance scenario refers to the presence of both stationary and moving obstacles in the waters where the USV is navigating, which is a combination of all the above scenarios and is the scenario closest to the real marine environment.

[0052] In this embodiment, as Figure 2 As shown, the step of controlling the unmanned surface vessel to autonomously avoid threatening targets according to the obstacle avoidance scenario includes steps S201~S207: S201. When the obstacle avoidance scenario is a single static target autonomous obstacle avoidance scenario, the target heading control quantity of the unmanned surface vessel is solved using the single static target obstacle avoidance algorithm.

[0053] In this embodiment, as Figure 3As shown, the single static target obstacle avoidance algorithm includes S301~S304: S301. Exaggerate the current threat target.

[0054] In this embodiment, the radar-sensing image is a binary matrix, assuming 1 represents a target and 0 represents no target, with a spacing of 5 meters between adjacent elements. The specific steps for expanding the threat target are as follows: Setting the expansion distance to 20 meters, each "1" of the threat target needs to be expanded outward by 20 meters to form a circular area. Combined with the parameter of a 5-meter spacing between adjacent elements, this area corresponds to covering the original "1" element and its surrounding "1" elements (i.e., a distribution of a total of 4 "1" elements) in the matrix, thereby achieving the expansion of the threat target.

[0055] S302. Using the result of the puffing process, model the target environment using the minimum circumscribed circle method to obtain the target environment. The specific process is as follows: (1) Construct convex polygons for the expanded threat targets, that is, connect each convex binary "1" value.

[0056] (2) Solve for the minimum circumcircle of the convex polygon, and then model the environment of the target to obtain the target environment.

[0057] S303. Analyze the target environment using the tangent method.

[0058] In this embodiment, two tangents (left tangent AC and right tangent AD) are drawn from the current unmanned surface vessel position to the circle (the circle formed by environmental modeling obtained in step S302). The line segment AB is formed by connecting the current unmanned surface vessel position with the center of the circle. Information such as the AC distance, AD distance, ∠CAB value and ∠DAB value can be obtained and used as the analysis result of the target environment.

[0059] S304. Dynamically calculate the target environment analysis results to obtain the target heading control quantity of the unmanned surface vessel.

[0060] In this embodiment, given the current unmanned surface vessel's heading value, target bearing, ∠CAB value, and ∠DAB value, the heading difference cost generated by the unmanned surface vessel turning left and right can be calculated. The heading difference with the smallest value is taken as the heading control quantity for the unmanned surface vessel to turn left or right, and this is used as the basis for the left and right turn decision.

[0061] The single static target obstacle avoidance algorithm in this embodiment is based on the input of perception information and the real-time prediction results. It can calculate a reasonable heading control amount in real time to ensure that the unmanned surface vessel always keeps the threatening target outside the threat distance, thereby realizing the autonomous obstacle avoidance function of the unmanned surface vessel for single static targets.

[0062] S202. When the obstacle avoidance scenario is a single large target autonomous obstacle avoidance scenario, the target heading control quantity of the unmanned surface vessel is solved using the single large target obstacle avoidance algorithm.

[0063] In this embodiment, as Figure 4 As shown, the single large target obstacle avoidance algorithm includes S401~S404: S401. All threat targets with the same target ID as the current threat target in the threat target information list are considered as large targets.

[0064] S402, The large target is expanded.

[0065] The specific process is the same as step S301, and will not be repeated here.

[0066] S403. Solve the puffing results of the continuous targets to obtain the tangent angle of the continuous target contour, the distance between the unmanned surface vessel and the target contour, and the distance between the unmanned surface vessel and the continuous targets.

[0067] Specifically, the tangent angles of the continuous target contours include: the tangent angles of the left and right contours of the continuous targets. The distances between the unmanned surface vessel (USV) and the target contours include: the distance between the USV and the tangent point of the left contour, and the distance between the USV and the tangent point of the right contour. The distances between the USV and the continuous targets include: the closest distance between the USV and the continuous targets.

[0068] In this embodiment, the calculation of the puffing result of continuous targets includes: (1) First, construct a convex polygon for the expanded threat target, that is, connect each convex binary "1" value, then solve for the minimum circumcircle of the convex polygon, and use it as the target circumcircle.

[0069] (2) Draw two tangents (left tangent AC and right tangent AD) from the current unmanned surface vessel position to the outer circle of the target. Connect the current unmanned surface vessel position with the center of the circle to form line segment AB. Then you can obtain information such as AC distance (distance between the unmanned surface vessel and the left contour tangent point), AD distance (distance between the unmanned surface vessel and the right contour tangent point), ∠CAB value (angle of the left contour tangent) and ∠DAB value (angle of the right contour tangent).

[0070] (3) Obtain the closest distance between the unmanned surface vessel and the continuous targets by traversing the distance values ​​between the current position and each binary "1" value of the threat target, and use it as the distance between the unmanned surface vessel and the continuous targets.

[0071] S404. Based on the unmanned surface vessel's (USV) heading, its preset navigation route, the tangent angles of continuous target contours, the distance between the USV and the target contours, and the distance between the USV and the continuous targets, the USV's target heading control parameters are estimated. The USV's heading and preset navigation route are considered part of its navigation information.

[0072] Specifically, the estimation process of the unmanned surface vessel (USV) target heading control quantity is as follows: based on the current USV preset navigation route, target bearing and continuous target contour tangent angle, and the distance between the USV and the target contour, the heading difference cost generated by the USV turning left and right is calculated. The smaller heading difference is taken as the heading control quantity for the USV to turn left or right, and this is used as the basis for left and right turn decisions.

[0073] In this embodiment, if it is determined that the obstacle avoidance action cannot be completed on the target based on factors such as distance, navigation direction, and maximum turning radius, the unmanned surface vessel enters a reverse navigation mode until it reaches a safe position to wait for the next mission.

[0074] The single large target obstacle avoidance algorithm in this embodiment is based on perception information input and real-time prediction results. It can calculate a reasonable heading control amount in real time to ensure that the unmanned surface vessel always keeps a single, single-line threat target outside the threat distance, thereby realizing the autonomous obstacle avoidance function of the unmanned surface vessel for single large targets. In addition, when it is determined that the obstacle avoidance action cannot be completed for the target, it can enter the reverse navigation mode to ensure the safe navigation of the unmanned surface vessel and the reachability of the mission.

[0075] S203. When the obstacle avoidance scenario is a single dynamic target autonomous obstacle avoidance scenario, the target heading control quantity of the unmanned surface vessel is solved using a single dynamic target obstacle avoidance algorithm.

[0076] In this embodiment, as Figure 5 As shown, the single dynamic target obstacle avoidance algorithm includes S501~S503: S501. Based on the least squares method, the motion state of the current threatening target is estimated in real time to obtain the speed of the threatening target.

[0077] Specifically, the least squares method is used to estimate the location of the threat target in the future based on its location information and movement distance over a certain period. The target's speed is then estimated by comparing its current location with the estimated future location.

[0078] S502. Set a threat distance threshold based on the unmanned surface vessel's (USV) speed and the speed of the threat target. The USV's speed is considered part of its navigation information.

[0079] Specifically, the threat distance threshold is proportional to both the speed of the unmanned surface vessel (USV) and the speed of the threat target; its minimum value is generally no less than 30 meters, and its maximum value is adapted to the maximum speed of the USV.

[0080] S503. If the distance between the current threat target and the unmanned surface vessel reaches the threat distance threshold, calculate the target heading control quantity of the unmanned surface vessel based on the single static target obstacle avoidance algorithm.

[0081] Specifically, the single static target obstacle avoidance algorithm solution is as shown in steps S301~S304, which will not be repeated here. The above processing follows the maritime collision avoidance rules.

[0082] The single dynamic target obstacle avoidance algorithm in this embodiment is based on the input of perception information and the real-time prediction results. It can calculate a reasonable heading control amount in real time to ensure that the unmanned surface vessel always keeps a single dynamic threat target outside the threat distance, thereby realizing the autonomous obstacle avoidance function of the unmanned surface vessel for a single dynamic target.

[0083] S204. When the obstacle avoidance scenario is an autonomous obstacle avoidance scenario with or without large targets, the target heading control quantity of the unmanned surface vessel is solved using the obstacle avoidance algorithm with or without large targets.

[0084] In this embodiment, as Figure 6 As shown, the obstacle avoidance algorithm with / without large targets includes S601~S605: S601. Exaggerate the current threat target.

[0085] The specific process is the same as step S301, and will not be repeated here.

[0086] S602. The expanded result of the current threat target is rasterized to obtain a global raster map of multiple static targets. The location of the current threat target belongs to the optimized perception information of the current threat target.

[0087] Specifically, the rasterization process is as follows: (1) Design the size of the grid cell, which is generally larger than the spacing between members after the perception information is binarized.

[0088] (2) Set a threshold for the area ratio of "1" members in each grid cell; when the ratio of "1" members in a grid cell is greater than the threshold, the grid cell is set to have a target, otherwise it is set to have no target, and finally a global grid map is formed.

[0089] S603. Set a safe endpoint based on the unmanned surface vessel's (USV) preset route and the current location of the threat target. The USV's preset route is considered part of its navigation information.

[0090] Specifically, a perpendicular line is drawn from the center point of the threat target to the preset route. Let the radius of the outer circle of the threat target corresponding to the farthest perpendicular intersection point be R, and the minimum turning radius of the unmanned surface vessel at its current speed be r. The point at the farthest perpendicular intersection point from the current position of the unmanned surface vessel, extended forward by R+r, is taken as the safe endpoint.

[0091] S604. Based on the global grid map of the multiple static targets and the safe endpoint, perform autonomous obstacle avoidance path planning for the multiple static targets to obtain the paths for the multiple static targets.

[0092] Specifically, in a global grid map with multiple static targets, the A* algorithm is used to search for paths based on the start and end points of the unmanned surface vessel (USV), completing the autonomous obstacle avoidance path planning for static targets and obtaining paths for multiple static targets. The A* algorithm is existing technology and will not be elaborated upon here.

[0093] As some optional implementation methods of this embodiment, after obtaining the multi-static target path, trajectory tracking control can be performed on the multi-static target path.

[0094] S605. Smooth the paths of the multiple static targets to obtain the unmanned surface vessel (USV) target heading control quantity. The specific process is as follows: Traverse all point sets on the multi-static target path; perform connection tests on each point. If the connected line segment does not intersect with the target, continue the connection test on the next point; otherwise, take the current point as the starting point and continue searching for the next endpoint of the line segment; until all point sets on the multi-static target path have been traversed, thereby achieving smoothing of the multi-static target path.

[0095] The obstacle avoidance algorithm for large / non-large targets in this embodiment is based on the input of perception information and the real-time prediction results. It can calculate a reasonable heading control amount in real time, ensuring that the unmanned surface vessel (USV) is always kept outside the threat distance for multiple large / non-large targets. This enables the USV to achieve autonomous obstacle avoidance for large / non-large targets and ensures the rationality of the USV's route planning, navigation safety, and mission reachability.

[0096] S205. When the obstacle avoidance scenario is a multi-dynamic target autonomous obstacle avoidance scenario, the multi-dynamic target obstacle avoidance algorithm is used to solve the target heading control quantity of the unmanned surface vessel.

[0097] In this embodiment, as Figure 7 As shown, the multi-dynamic target obstacle avoidance algorithm includes S701~S707: S701. Extract the motion state of the current threat target from the optimized perception information of the current threat target.

[0098] Specifically, the current motion status of the threat target includes information such as speed, heading, and position.

[0099] S702. The least squares method is used to estimate the motion state of the current threatening target.

[0100] Specifically, the least squares method is used to estimate the position of the threat target in the future based on the target's location information and movement distance over a certain period. Then, by comparing the future target position with the current target position, the target's speed is estimated. The target's location information and movement distance both relate to the target's current motion state.

[0101] S703. Set the delay time based on the motion estimation results.

[0102] Specifically, the delay time is generally positively correlated with the distance threshold when the unmanned surface vessel begins obstacle avoidance maneuvers at that speed. The minimum delay time is generally no less than 10 seconds, and the maximum delay time is generally no more than 1 minute.

[0103] S704. Obtain the threat target area based on the delay time, and perform obstacle materialization processing on the threat target area to obtain the obstacle area.

[0104] Specifically, obstacle materialization refers to transforming the binary "0" value within the driving area of ​​the threatening target into "1" within a time delay, thereby achieving obstacle materialization.

[0105] In this embodiment, obtaining the threat target area based on the delay time refers to the area where the target is located during the delay time.

[0106] S705. The obstacle area is rasterized to obtain a global raster map of multiple dynamic targets.

[0107] The specific process is the same as step S602, and will not be repeated here.

[0108] S706. Perform autonomous obstacle avoidance path planning for multiple dynamic targets based on the global grid map of the multiple dynamic targets to obtain the paths for the multiple dynamic targets.

[0109] The specific process is the same as step S604, and will not be repeated here.

[0110] S707. Smooth the paths of the multiple dynamic targets to obtain the target heading control quantity of the unmanned surface vessel.

[0111] The specific process is the same as step S605, and will not be repeated here.

[0112] The multi-dynamic target obstacle avoidance algorithm in this embodiment can calculate a reasonable heading control amount in real time based on the input of perception information and the real-time prediction results, so as to ensure that the unmanned surface vessel always keeps the multi-dynamic target outside the threat distance, thereby realizing the autonomous obstacle avoidance function of the unmanned surface vessel for multi-dynamic targets.

[0113] S206. When the obstacle avoidance scenario is an autonomous obstacle avoidance scenario with mixed static and dynamic targets, the target heading control quantity of the unmanned surface vessel is solved by using the static and dynamic target obstacle avoidance algorithm.

[0114] In this embodiment, as Figure 8 As shown, the static and dynamic target obstacle avoidance algorithm includes S801~S804: S801. Based on the target type of the current threat target, the current threat target is expanded.

[0115] The specific process is the same as step S301, and will not be repeated here.

[0116] S802. The puffing result of the current threat target is rasterized to obtain a global raster map of static and dynamic targets.

[0117] The specific process is the same as step S602, and will not be repeated here.

[0118] S803. Based on the global grid map of static and dynamic targets, perform autonomous obstacle avoidance path planning for static and dynamic targets to obtain the paths for static and dynamic targets.

[0119] The specific process is the same as step S604, and will not be repeated here.

[0120] S804. Smooth the static and dynamic target paths to obtain the unmanned surface vessel target heading control quantity.

[0121] The specific process is the same as step S605, and will not be repeated here.

[0122] The static and dynamic target obstacle avoidance algorithm in this embodiment can calculate a reasonable heading control amount in real time based on the input of perception information and the real-time prediction results, so as to ensure that the unmanned surface vessel always keeps static and dynamic targets outside the threat distance, thereby realizing the autonomous obstacle avoidance function of the unmanned surface vessel for static and dynamic targets.

[0123] In this embodiment, after obstacle avoidance of a certain threat target (any of the situations referred to in S202~S207) is completed, its corresponding information is removed from the threat target information list, the obstacle avoidance scenario is adjusted, and autonomous obstacle avoidance actions are continuously performed until the unmanned surface vessel reaches a safe area.

[0124] S207. Construct obstacle avoidance termination conditions, and control the unmanned surface vessel to perform autonomous obstacle avoidance based on the target heading control quantity and obstacle avoidance termination conditions.

[0125] Specifically, the process of controlling the unmanned surface vessel (USV) to perform autonomous obstacle avoidance based on the target heading control quantity and obstacle avoidance termination condition is as follows: the USV's heading is adjusted based on the target heading control quantity, and during the adjustment process, it is detected in real time whether the obstacle avoidance termination condition is met. If the condition is met, obstacle avoidance ends, and the USV is controlled to return to the preset route autonomously to continue performing the task; otherwise, autonomous obstacle avoidance continues.

[0126] In this embodiment, the obstacle avoidance termination condition includes: (1) When the target is a single static target or a single dynamic target, the obstacle avoidance termination condition is that the Y-axis coordinate of the unmanned surface vessel in the rectangular judgment coordinate system is greater than the preset coordinate threshold, the minimum turning radius path does not intersect with the puffed target, and the distance between the target and the unmanned surface vessel is greater than the preset safe distance threshold.

[0127] Specifically, the preset coordinate threshold is the absolute value of the y-coordinate in the unmanned surface vessel coordinate system, while the x-coordinate value is unlimited. The preset safe distance threshold is proportional to both the speed of the unmanned surface vessel and the speed of the threatening target. The minimum value is generally not less than 30 meters, and the maximum value is adapted according to the maximum speed of the unmanned surface vessel.

[0128] Specifically, the rectangular coordinate system is constructed with the target center as the origin and the direction from the start to the end of the unmanned surface vessel's preset route as the positive Y-axis; the expanded target is obtained by expanding the target; the minimum turning radius path is obtained based on the unmanned surface vessel's speed; the unmanned surface vessel's preset route and speed belong to the unmanned surface vessel's navigation information.

[0129] (2) When the target is a single large target, the obstacle avoidance termination condition is that the path tangent intersects with the preset route of the unmanned vessel, the minimum turning radius path does not intersect with the puffed target, and the minimum distance between the unmanned vessel and the edge of the target is greater than the preset safe distance threshold.

[0130] Specifically, the path tangent is obtained by drawing a tangent from the current position of the unmanned surface vessel (USV) towards the expanding target in the direction forward of the USV.

[0131] This embodiment determines the end of the obstacle avoidance action based on the obstacle avoidance end condition, and standardizes the end condition of the collision avoidance action for different targets. It focuses on the safety of the unmanned surface vessel and minimizes the movement range of the unmanned surface vessel and reduces energy consumption while ensuring that the obstacle avoidance action can be completed.

[0132] This embodiment achieves autonomous obstacle avoidance for unmanned surface vessels (USVs) in any scenario through classification, prediction, and global processing. By visualizing the obstacle avoidance scenario, operators can intuitively understand and control the USV's situational information, which plays a crucial role in subsequent mission planning and decision-making.

[0133] As some optional implementations of this embodiment, the unmanned surface vessel (USV) also includes the following during mission execution: (1) Real-time detection of whether the unmanned surface vessel and the target meet the obstacle avoidance termination conditions.

[0134] (2) When the obstacle avoidance termination condition is met, control the unmanned surface vessel to return to the preset route and continue to perform the mission; wherein, the preset route is the navigation information of the unmanned surface vessel.

[0135] As some optional implementations of this embodiment, S102~S105 can be used as an autonomous obstacle avoidance thread, and this thread can be kept running continuously during the unmanned surface vessel's mission.

[0136] This embodiment realizes the visualization of target types and the scenario-based representation of obstacle avoidance situations: through classification, prediction, and global processing, it enables autonomous obstacle avoidance of unmanned surface vessels (USVs) in any scenario, allowing for stable application of autonomous obstacle avoidance in various complex mission scenarios. Simultaneously, it clearly displays each threat target in the autonomous obstacle avoidance scenario, enabling operators to intuitively understand and control the USV's situational information, playing a crucial role in subsequent mission planning and decision-making.

[0137] According to embodiments of the present invention, the following advantages are available compared to the prior art: (1) The method of the present invention enables unmanned surface vessels (USVs) to establish a target type list and classify obstacle avoidance scenarios based on the continuous input of perception information from the USV's onboard perception devices when performing complex tasks involving navigation control. This allows for clear display and interpretation of the USV's autonomous obstacle avoidance situation information. The autonomous obstacle avoidance algorithm performs optimal self-course and speed control based on the current course status and target distribution to avoid obstacles and ensure its own safety. This method has no requirements on the USV's structure or actuators and has the advantages of strong universality, low safety risk, and strong stability.

[0138] (2) The method of the present invention can effectively predict the target movement trend in the threat time domain according to different target types, thereby significantly reducing the uncertainty and danger of unmanned surface vessels in actual environmental applications and improving operational safety. In addition, the present invention overcomes the limitations of the prior art which can only avoid obstacles to single static, dynamic targets and single large targets. It can cope with complex mixed scenarios containing single targets and multiple targets, and solves the problems of lack of targeted autonomous decision-making and inability to predict target movement trends in global scenarios.

[0139] (3) The method of the present invention has a fast calculation speed and low resource consumption. It can output control allocation values ​​in real time at a high frequency in actual tasks. Through the rapid response of the relevant actuators of the unmanned surface vessel, the control accuracy can be improved. It has good real-time performance and robustness.

[0140] (4) The method of the present invention completes the relevant tasks while ensuring the safety of the unmanned surface vessel (USV). By improving the comprehensiveness and stability of the USV's autonomous obstacle avoidance function, the safety of the USV is no longer a limiting factor for its task completion, thereby promoting the technological development of the USV-related field.

[0141] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0142] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0143] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A multi-scenario autonomous obstacle avoidance method for unmanned surface vessels based on perception information, characterized in that, include: Obtain navigation information of unmanned surface vessels; During the mission, the unmanned surface vessel (USV) uses its sensing equipment to perceive the target to be avoided. The perception information of each target to be avoided is optimized; then, based on the optimized perception information of each target to be avoided, the type of each target to be avoided is obtained. Based on the perceived information and type of all targets to be avoided, a threat information list is constructed; The system iterates through the threat target information list. If the distance between the current threat target and the unmanned surface vessel (USV) reaches a preset threat distance threshold, then it divides the obstacle avoidance scenario based on the information of the current threat target and the number of threat targets in the threat target information list. Based on the obstacle avoidance scenario, it controls the USV to autonomously avoid the threat target.

2. The method according to claim 1, characterized in that, The types of targets to be avoided include: target type and target threat type; The target type is obtained by classifying the type based on the optimized perception information of each target; The target threat type is obtained by identifying the threat type from the optimized perception information of each target.

3. The method according to claim 1, characterized in that, The method of controlling the unmanned surface vessel to autonomously avoid threatening targets based on the obstacle avoidance scenario includes: When the obstacle avoidance scenario is a single static target autonomous obstacle avoidance scenario, the target heading control quantity of the unmanned surface vessel is solved using a single static target obstacle avoidance algorithm; When the obstacle avoidance scenario is a single large target autonomous obstacle avoidance scenario, the target heading control quantity of the unmanned surface vessel is solved by using the single large target obstacle avoidance algorithm; When the obstacle avoidance scenario is a single dynamic target autonomous obstacle avoidance scenario, the target heading control quantity of the unmanned surface vessel is solved by using a single dynamic target obstacle avoidance algorithm; When the obstacle avoidance scenario is an autonomous obstacle avoidance scenario with or without large static targets, the target heading control quantity of the unmanned surface vessel is solved by using the obstacle avoidance algorithm with or without large targets; When the obstacle avoidance scenario is a multi-dynamic target autonomous obstacle avoidance scenario, the multi-dynamic target obstacle avoidance algorithm is used to solve the target heading control quantity of the unmanned surface vessel; When the obstacle avoidance scenario is an autonomous obstacle avoidance scenario with mixed static and dynamic targets, the target heading control quantity of the unmanned surface vessel is solved by using the static and dynamic target obstacle avoidance algorithm. Construct obstacle avoidance termination conditions; control the unmanned surface vessel to perform autonomous obstacle avoidance based on the target heading control quantity and obstacle avoidance termination conditions.

4. The method according to claim 3, characterized in that, The single large target obstacle avoidance algorithm includes: All threat targets with the same target ID as the current threat target in the threat target information list are considered as large targets; The large target is subjected to an expansion process; The results of the puffing process of continuous targets are calculated to obtain the tangent angle of the continuous target contour, the distance between the unmanned surface vessel and the target contour, and the distance between the unmanned surface vessel and the continuous targets. The target heading control quantity of the unmanned surface vessel (USV) is estimated based on the USV's heading, the USV's preset navigation route, the tangent angle of the continuous target contour, the distance between the USV and the target contour, and the distance between the USV and the continuous targets. Among them, the unmanned surface vessel's heading and preset navigation route are considered unmanned surface vessel navigation information.

5. The method according to claim 3, characterized in that, The single dynamic target obstacle avoidance algorithm includes: The motion state of the current threatening target is estimated in real time based on the least squares method to obtain the target's speed. Set a threat distance threshold based on the unmanned surface vessel's speed and the speed of the threat target; If the distance between the current threat target and the unmanned surface vessel (USV) reaches the threat distance threshold, the USV target heading control quantity is calculated based on the single static target obstacle avoidance algorithm. Among them, the unmanned surface vessel's (USV) speed is part of its navigation information.

6. The method according to claim 3 or 5, characterized in that, The single static target obstacle avoidance algorithm includes: Exaggerate the current threat targets; The results of the puffing process are used to model the target environment using the smallest circumscribed circle to obtain the target environment; The target environment is analyzed using the tangent method; The target environment analysis results are dynamically calculated to obtain the target heading control quantity of the unmanned surface vessel.

7. The method according to claim 3, characterized in that, The algorithm for obstacle avoidance with / without large targets includes: Exaggerate the current threat targets; The expanded results of the current threat targets are rasterized to obtain a global raster map of multiple static targets; A safe endpoint is set based on the unmanned surface vessel's preset route and the current location of the threat target; Based on the global grid map of the multiple static targets and the safe endpoint, autonomous obstacle avoidance path planning for the multiple static targets is performed to obtain the paths for the multiple static targets; The paths of the multiple static targets are smoothed to obtain the target heading control quantity of the unmanned surface vessel; Among them, the preset route of the unmanned surface vessel (USV) belongs to the USV navigation information, and the current threat target location belongs to the optimized perception information of the current threat target.

8. The method according to claim 3, characterized in that, The multi-dynamic target obstacle avoidance algorithm includes: Extract the motion state of the current threat target from the optimized perception information of the current threat target; The least squares method is used to estimate the motion state of the current threatening target; Set the delay time based on the motion estimation results; The threat target area is obtained based on the delay time, and the threat target area is materialized into an obstacle area. The obstacle area is rasterized to obtain a global raster map of multiple dynamic targets; Based on the global grid map of the multi-dynamic targets, autonomous obstacle avoidance path planning for the multi-dynamic targets is performed to obtain the multi-dynamic target paths; The paths of the multiple dynamic targets are smoothed to obtain the target heading control quantity of the unmanned surface vessel.

9. The method according to claim 3, characterized in that, The static and dynamic target obstacle avoidance algorithm includes: Based on the target type of the current threat target, the current threat target is expanded; The expanded result of the current threat target is rasterized to obtain a global raster map of static and dynamic targets; Based on the global grid map of static and dynamic targets, autonomous obstacle avoidance path planning for static and dynamic targets is performed to obtain the paths for static and dynamic targets. The static and dynamic target paths are smoothed to obtain the unmanned surface vessel target heading control quantity.

10. The method according to claim 3, characterized in that, The conditions for ending the obstacle avoidance construction include: When the target is a single static target or a single dynamic target, the obstacle avoidance termination condition is that the Y-axis coordinate of the unmanned surface vessel in the Cartesian judgment coordinate system is greater than a preset coordinate threshold, the minimum turning radius path does not intersect with the puffed target, and the distance between the target and the unmanned surface vessel is greater than a preset safe distance threshold. When the target is a single large target, the obstacle avoidance termination condition is that the path tangent intersects with the preset route of the unmanned surface vessel, the minimum turning radius path does not intersect with the puffed target, and the minimum distance between the unmanned surface vessel and the edge of the target is greater than the preset safe distance threshold. The rectangular coordinate system is constructed with the target center as the origin and the direction from the start to the end of the unmanned surface vessel's preset route as the positive Y-axis. The expanded target is obtained by expanding the target. The minimum turning radius path is obtained based on the unmanned surface vessel's speed. The path tangent is obtained by drawing a tangent from the current position of the unmanned surface vessel towards the expanded target in the forward direction. The preset route and speed of the unmanned surface vessel are part of the unmanned surface vessel's navigation information.

Images