Unmanned hovercraft cross-domain trajectory planning and dynamic collision avoidance method and device and storage medium

Abstract

The invention discloses an unmanned hovercraft cross-domain trajectory planning and dynamic collision avoidance method and device and a storage medium, and belongs to the field of land and water cross-domain path planning. Comprising the following steps: fusing terrain geographic information, three-dimensional point cloud information and unmanned hovercraft obstacle crossing ability of a cross-domain task area to construct a map; carrying out path search by using an improved mixed A * algorithm, carrying out node expansion by taking a water surface kinematics equation and a land kinematics equation as constraints to obtain a cross-domain discrete path, and carrying out path optimization in combination with the limitation of the speed and the acceleration of the unmanned hovercraft; when trajectory tracking is carried out and an unknown obstacle is found to conflict with a trajectory, a destination of path re-planning is firstly determined, then trajectory re-planning search is carried out by adopting a topological consistency kinematics RRT * algorithm, and local planning is carried out in combination with limitation of the speed and the acceleration of the unmanned hovercraft until the unmanned hovercraft reaches the destination. According to the method, searching is more oriented, the path searching speed is higher, and the node growth quality is higher.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of amphibious path planning, and particularly relates to an unmanned hovercraft cross-domain trajectory planning and dynamic collision avoidance method, equipment and a storage medium. BACKGROUND

[0002] With the continuous development of unmanned systems, as a kind of efficient amphibious operation load equipment, the necessity of unmanned autonomy of full-cushion hovercraft is increasing in various fields. The main operation content of the unmanned hovercraft is amphibious transportation, and according to the landing operation task requirement, planning a safe and collision-free trajectory that meets the operating limit is the core technology for realizing autonomous operation.

[0003] The full-cushion hovercraft has great load, high speed and harsh working environment, which makes it difficult to be unmanned and autonomous, especially in the autonomous planning technology, there are two key problems: when operating in the amphibious area, the selected landing area slope should not exceed the obstacle crossing ability of the hovercraft cushion state navigation; the unmanned hovercraft operating at high speed on the sea has a safety operating limit, and cannot flexibly change the forward speed and direction when performing collision avoidance.

[0004] In the field of path planning, the current method mainly improves or fuses the algorithm based on the classic algorithm for specific problems, including graph search-based algorithm, sampling-based search algorithm, bionics-based intelligent algorithm, artificial potential field-based algorithm, and dynamic constraint-based algorithm. In view of the key problems existing in the above-mentioned cross-domain trajectory planning of the unmanned hovercraft, the method provides the following planning framework: first, the feasible region of the sea-land cross-domain environment is analyzed, and a 2.5D digital elevation model (2.5D DEM) is established as the basis for trajectory planning; then, a hybrid A* algorithm is used to plan a path that meets the kinematics of the unmanned hovercraft, and according to the special dynamics model of the hovercraft, an executable reference trajectory corresponding to different speeds is optimized; after dynamically perceiving the obstacles conflicting with the reference trajectory, the topologically consistent kinematics RRT* algorithm improved by the method can quickly re-plan the trajectory based on the original reference trajectory, and ensure that the re-planned trajectory meets the operating requirements, thereby effectively solving the problem. SUMMARY

[0005] The application provides a trajectory planning and obstacle avoidance method based on a 2.5D digital elevation model to realize the accuracy of cross-domain trajectory planning and obstacle avoidance of the unmanned hovercraft.

[0006] The application provides an unmanned hovercraft cross-domain trajectory planning and dynamic collision avoidance method, which comprises the following steps:

[0007] A three-layer map is constructed by fusing a terrain segmentation layer of a map, a feasible region layer of the map and a cost parameter layer of the motion of the unmanned hovercraft;

[0008] The improved hybrid A* algorithm is used for global path planning, and the node expansion is performed with the water surface kinematic equation and the land kinematic equation as constraints to obtain a cross-domain discrete path from the starting point to the target point; the trajectory that meets the kinematic and dynamic characteristics of the unmanned air-cushion vehicle is obtained by optimization in combination with the speed and acceleration limitations of the unmanned air-cushion vehicle;

[0009] When the trajectory tracking of the unmanned air-cushion vehicle finds that an unknown obstacle conflicts with the trajectory, the endpoint of path re-planning is determined, and then the topologically consistent kinematic RRT* algorithm is used for trajectory re-planning search, and the local path planning is performed in combination with the speed and acceleration limitations of the unmanned air-cushion vehicle until the unmanned air-cushion vehicle reaches the endpoint.

[0010] Further, the terrain geographic information is contained in two-dimensional electronic chart data, the two-dimensional electronic chart data is subjected to binaryzation rendering processing by using an open source geographic library Shapely to obtain a terrain segmentation layer of the map; a landing feasible region and a land feasible region map layer, i.e., a feasible region layer of the map, are obtained from the terrain geographic information according to the obstacle crossing capability of the unmanned air-cushion vehicle, and further, a cost parameter layer of air-cushion vehicle movement is constructed.

[0011] The obstacle crossing capability of the unmanned air-cushion vehicle is:

[0012]

[0013] wherein, θ max , are the maximum pitch angle and the maximum roll angle of the dynamic obstacle crossing of the air-cushion vehicle respectively.

[0014] Further, in the global path planning and the local path planning process, the obstacle collision avoidance hard constraint is as follows:

[0015]

[0016] wherein, υ is the geometric shape of the unmanned air-cushion vehicle; P is a feasible region obtained from the feasible region layer; o d is a dynamic obstacle; and o i is a static obstacle.

[0017] Further, the water surface movement equation and the land movement equation of the unmanned air-cushion vehicle are as follows:

[0018]

[0019] st.ψ<ψ max

[0020] wherein, u is a constant longitudinal speed, ψ is a heading angle, a is a longitudinal acceleration, ψ max is the maximum turning angle of the air-cushion vehicle in high-speed navigation.

[0021] Further, the cost function f of the node expansion is:

[0022] f(n) = g(n) + h(n)

[0023] where g is the path cost, the length of the path from the start point to the current node multiplied by the penalty term; h is the heuristic cost, the length of the Dubins curve from the current node to the goal node.

[0024] Further, the path replanning path endpoint is:

[0025] S = a x l + S0

[0026] where S is the length of the trajectory from the initial start point to the replanning termination point; S0 is the length of the trajectory from the initial start point to the replanning start point; a is the limit distance of sensor perception; l is the forward exploration parameter.

[0027] Further, the cost function of the topologically consistent kinematic RRT* algorithm is:

[0028]

[0029] where f cost is the cost value of the expansion to the current node, given by the penalty function P; x near is the four-degree-of-freedom state of the neighboring node; a is the heading angle of the air cushion ship, i.e. y; a near is the orientation of the neighboring node to the current node; a goal is the orientation of the current node to the target node; m and n are distance cost parameters and steering cost parameters respectively which change with the switching of water and land environments.

[0030] Further, the discrete path kinematic constraint combining the speed and acceleration limits of the unmanned air cushion ship in the hybrid A* algorithm and the topologically consistent kinematic RRT* algorithm is:

[0031] x k+1 -L(x k ,u k ) = 0, k = 0, …, N-1,

[0032] u min ≤ u k ≤ u max , k = 0, …, N,

[0033] a min ≤ a k ≤ a max , k = 0, …, N-1.

[0034] where x is represented as {x, y, z, y} Tis the air cushion ship four-degree-of-freedom state, u is represented as {u, v, w, r} T is the air cushion ship four-degree-of-freedom velocity, L is the kinematics mapping, a is the air cushion ship four-degree-of-freedom acceleration, N is the total number of discrete path points, u max and u min are the maximum and minimum values of the velocity respectively, a max and a min are the maximum and minimum values of the acceleration respectively;

[0035]

[0036] where, T f is the time of the entire path tracking control, i.e., the solution item time item cost J t , J input (u k ) is the control change cost.

[0037] The application further provides a computer device / equipment / system, comprising a memory, a processor and a computer program stored in the memory, wherein the processor implements the steps of the air cushion ship cross-domain trajectory planning and dynamic collision avoidance method according to any one of the above embodiments when executing the computer program.

[0038] The application further provides a computer readable storage medium, which stores a computer program / instruction, wherein the computer program / instruction implements the steps of the air cushion ship cross-domain trajectory planning and dynamic collision avoidance method according to any one of the above embodiments when executed by a processor.

[0039] The application further provides a computer program product, which comprises a computer program / instruction, wherein the computer program / instruction implements the steps of the air cushion ship cross-domain trajectory planning and dynamic collision avoidance method according to any one of the above embodiments when executed by a processor.

[0040] The application has the following beneficial effects:

[0041] The proposed method for cross-domain trajectory planning and dynamic collision avoidance of unmanned hovercraft creatively integrates the algorithm's variable parameters with terrain information and stores it in map form. Combined with an improved hybrid A* algorithm, it achieves smooth switching of land and water kinematic parameters during cross-domain operations while meeting planning requirements. Compared to the original RRT algorithm, the improved RRT replanning method of this invention provides a more directional search, yielding collision avoidance paths that conform to the kinematic characteristics of the unmanned hovercraft, facilitating the fitting of trajectories that meet control needs. Furthermore, the constrained search process suppresses the growth of irrelevant nodes, improving the efficiency of effective search. When using the improved hybrid A* algorithm and the improved RRT algorithm for global and local planning, this invention introduces obstacle collision constraints. These constraints combine the feasible area of ​​the unmanned hovercraft, dynamic obstacle constraints, and static obstacle constraints, enabling the unmanned hovercraft to navigate within the feasible area while achieving precise obstacle avoidance. Attached Figure Description

[0042] Figure 1 This is a flowchart of the cross-domain trajectory planning and dynamic collision avoidance method for unmanned hovercraft of the present invention;

[0043] Figure 2 This is a schematic diagram of the geodetic coordinate system and the hull coordinate system of the unmanned hovercraft of the present invention.

[0044] Figure 3 This is a map of the electronic nautical chart processed by the method of the present invention after binarization;

[0045] Figure 4 This is a schematic diagram of a 2.5D DEM multi-layer two-dimensional map according to the method of the present invention;

[0046] Figure 5 This is a 2.5D DEM map after two-dimensional rasterization by the method of the present invention;

[0047] Figure 6 This is a schematic diagram illustrating the dynamic selection of the replanning termination point in the method of the present invention;

[0048] Figure 7 This is a schematic diagram of the topology consistency RRT* algorithm of the present invention. Detailed Implementation

[0049] The present invention will now be further described with reference to the accompanying drawings.

[0050] This invention discloses a method for cross-domain trajectory planning and dynamic collision avoidance of an unmanned hovercraft, comprising:

[0051] Step 1: Use the terrain geographic information of the cross-domain task area, three-dimensional point cloud information, and analyze the feasible area according to the obstacle crossing ability of the unmanned air cushion vehicle to perform fusion mapping. Obtain the sea-land boundary according to the terrain geographic information to construct the terrain segmentation layer of the map; obtain the distribution of known obstacles on the water surface according to the three-dimensional point cloud information, analyze the terrain geographic information according to the obstacle crossing ability of the unmanned air cushion vehicle to obtain the landing feasible area and land feasible area information, thereby constructing the feasible area layer of the map; comprehensively construct the cost parameter layer of the air cushion boat movement by integrating the terrain segmentation layer and the feasible area layer of the above-mentioned map; finally, store the information as a 2.5D DEM map using the GridMap library;

[0052] Step 2: Select the trajectory planning starting point and target point as needed, and determine whether the starting point and target point in the 2.5D DEM are in the reachable area of the unmanned air cushion vehicle; set the path planning obstacle avoidance constraint condition, and introduce the feasible area of the unmanned air cushion vehicle, dynamic obstacle constraint and static obstacle constraint;

[0053] Step 3: Use the improved hybrid A* algorithm for path search, and respectively use the water surface kinematics equation and the land kinematics equation as constraints for node expansion according to different driving environments of the unmanned air cushion vehicle to obtain a cross-domain discrete path from the starting point to the target point;

[0054] Step 4: Optimize the discrete path point information in combination with the speed and acceleration limitations of the unmanned air cushion vehicle, and fit a trajectory that meets the kinematics and dynamics characteristics of the unmanned air cushion vehicle, and simultaneously parameterize the time to obtain a control target;

[0055] Step 5: Control the unmanned air cushion vehicle to perform trajectory tracking and perceive the position information of the surrounding unknown obstacles, and if the unknown obstacles conflict with the initial trajectory, perform local path re-planning, and use the current position and attitude information of the unmanned air cushion vehicle as the re-planning starting point, and dynamically select the position and attitude information of the re-planning termination point on the initial trajectory at a set length;

[0056] Step 6: Use the starting and ending point information obtained in step 5 to perform trajectory re-planning search using the topologically consistent RRT* algorithm to ensure that the re-planned trajectory maintains topological consistency with the initial trajectory within a specified time t, thereby eliminating the speed jump when switching trajectories, and completing the trajectory re-planning. Repeat steps 4, 5 and 6 until the unmanned air cushion vehicle reaches the target point.

[0057] Embodiment 1

[0058] An unmanned air cushion vehicle cross-domain trajectory planning and dynamic collision avoidance method, as shown in Figure 1 , includes:

[0059] Step 1: Fusion mapping and feasible region analysis using terrain geographic information, three-dimensional point cloud information and hovercraft obstacle crossing ability in cross-domain task area.

[0060] For the analysis of the hovercraft, its inertial coordinate system, ship coordinate system and related symbols are shown in FIG. 1. Figure 2 For the two-dimensional electronic chart data containing terrain geographic information, binaryzation rendering processing is performed using the open source geographic library Shapely to construct the terrain segmentation layer of the map, and the processing of the electronic chart is shown in FIG. 2. Figure 3 The three-dimensional point cloud information is clustered to obtain the known obstacle distribution; the terrain geographic information is analyzed according to the hovercraft obstacle crossing ability to obtain the landing feasible region and land feasible region information, thereby constructing the feasible region layer of the map, and the hovercraft obstacle crossing ability is expressed as follows:

[0061]

[0062] wherein, θ max , are the maximum pitch angle and the maximum roll angle of the dynamic obstacle crossing of the hovercraft, respectively. Then, the terrain segmentation layer and the feasible region layer constructed are integrated to further construct the cost parameter layer of the hovercraft motion; finally, the above three map layers are constructed into a 2.5DDEM map using the GridMap library and stored, and the principle and schematic diagram are shown in FIG. 3 and FIG. 4. Figure 4 Figure 5

[0063] Step 2: The starting point and target point of trajectory planning are selected as required, and it is judged whether the starting point and target point in the 2.5DDEM are in the hovercraft reachable region. In order to ensure that the path planned by the algorithm is effective and collision-free, the following assumptions are made: the hovercraft is a rigid body with uniform mass; the hovercraft has a fixed and unchanged two-dimensional overhead geometric shape υ during motion. The collision avoidance hard constraint during planning is defined as follows:

[0064]

[0065] wherein, Ρ is the feasible region, which is given by the feasible region layer, ο d is the dynamic obstacle, and ο i is the static obstacle. The legal judgment of the starting and ending points is completed by the above formula, and global path search and local re-planning are performed under the constraint premise.

[0066] Step 3: The improved hybrid A* algorithm is used for path search, and the water surface kinematics equation and the land kinematics equation are used as constraints for node expansion according to different driving environments of the hovercraft. The water surface kinematics equation and the land kinematics equation have different parameters, and the kinematics equations corresponding to different driving environments of the hovercraft are as follows:​​

[0067]

[0068] where u is constant longitudinal velocity, ψ is heading angle, a is longitudinal acceleration, ψ max is the maximum turning angle of hovercraft in high speed.

[0069] In the path search process, the real-time switching of land path points and water surface path points can be realized; according to the actual motion characteristics of the hovercraft, when the hybrid A* algorithm is used for node expansion, only the forward search strategy is adopted. Thus, the expanded nodes conforming to the kinematic characteristics of the hovercraft are obtained, and the cost function calculation formula of the node is as follows:

[0070] f(n)=g(n)+h(n)(4)

[0071] wherein the calculation of the cost function f value includes path cost g and heuristic function cost h. In the application, the path length from the starting point to the current node multiplied by the penalty term is used as the path cost g, and the penalty term is mainly for the nodes producing obstacle crossing actions in the land environment, and the penalty term of each node is given by the feasible region map layer in the 2.5DDEM; the Dubins curve length from the current node to the target point is used as the heuristic function cost h, and thus the numerical value of the cost function f of each feasible node can be obtained. The node is expanded to the target point, thereby completing the search of the global path and obtaining the discrete path points.

[0072] Step 4: On the basis of the discrete path points, the discrete path points are optimized and fitted to the continuous trajectory conforming to the kinematic characteristics of the unmanned hovercraft in combination with the speed and acceleration restrictions of the unmanned hovercraft. Considering that the paths searched by the hybrid A* algorithm and the topological consistency RRT* algorithm are in discrete form, the kinematic constraints of the discrete form of the unmanned hovercraft are given as follows:

[0073] x k+1 -L(x k ,u k )=0,k=0,…,N-1,

[0074] u min ≤u k ≤u max ,k=0,…,N,(5)

[0075] a min ≤a k ≤a max ,k=0,…,N-1.

[0076] wherein x represents {x, y, z, ψ} T is the four-degree-of-freedom state of the hovercraft, and u represents {u, v, w, r} Tis the air cushion ship four-degree-of-freedom velocity, L is the kinematic mapping, a represents i.e. is the air cushion ship four-degree-of-freedom acceleration, N is the total number of discrete path points, u max and u min are the maximum and minimum values of the velocity, respectively, a max and a min are the maximum and minimum values of the acceleration, respectively.

[0077] Under the discrete kinematic constraints, path optimization is performed. First, the control problem is converted into an optimal problem with minimum time and minimum control variation as the optimization objectives, and the following optimal control problem is constructed for solving:

[0078]

[0079] where T f is the time of the entire path tracking control, i.e., the solution time item in the optimization problem, J t is the cost of the optimization problem, J input (u k ) is the control variation cost. The solution of the optimal problem can be quickly calculated by using an existing solver, and the solution time of the initial trajectory can be controlled within 500 ms, and the solution time of the local trajectory can be controlled within 100 ms, which can meet the real-time requirement. After the solution of the minimum time cost J t and the minimum control variation cost J input (u k ) of the path tracking control is completed, a continuous trajectory is obtained by using a polynomial fitting method, and time parameterization is completed, which is used as a reference trajectory, and then, the MPC algorithm is used for trajectory tracking control.

[0080] Step 5: The unmanned air cushion ship is controlled to track the trajectory, and the position information of the surrounding unknown obstacles is perceived and determined by using devices such as a laser radar and a depth camera, the continuous-time obstacle position information is analyzed, the motion prediction and collision judgment of the obstacles are realized. If the air cushion ship continues to track the current trajectory and a collision with the obstacle occurs, collision avoidance re-planning is needed, otherwise, the original trajectory is continuously tracked. When the unknown obstacle conflicts with the initial trajectory, the start and end points of the path re-planning are determined, and then the local re-planning of step 6 is performed. The start and end points of the re-planning are selected by the following method: the current position and attitude information of the unmanned air cushion ship is taken as the re-planning starting point; the position of the re-planning termination point is determined by the following formula:

[0081] S=a x l+S0(7)

[0082] where S is the length of the trajectory from the initial start point to the replanning end point, S0 is the length of the trajectory from the initial start point to the replanning start point, a is the limit distance of perception, and λ is the forward exploration parameter. The end point of the replanning is dynamically optimized on the reference trajectory, ensuring that the final state of the replanning overlaps with the initial trajectory. Figure 6 Fig. 2 shows a schematic diagram of dynamic selection of the replanning end point.

[0083] Step 6: Based on the start and end point information obtained in step 5, a topologically consistent kinematic RRT* algorithm is used for trajectory replanning search. Considering the motion characteristics of the air-cushion vehicle, a steering angle soft constraint penalty is added, and the penalty term is taken as part of the RRT* algorithm node expansion cost, which is mapped by the following formula:

[0084]

[0085] where f cost is the value of the expansion cost to the current node, given by the penalty function P; x near is the four-degree-of-freedom state of the neighboring node; a is the heading angle of the air-cushion vehicle, i.e., ψ; a near is the orientation of the neighboring node to the current node; a goal is the orientation of the current node to the target node; m and n are distance cost parameters and steering cost parameters, respectively, which change with the switching of water and land environments.

[0086] The optimization of the discrete path points is performed for path optimization using the kinematic constraints of step 4 (formula (5) and formula (6)).

[0087] Table 1 is the pseudo code of the RRT algorithm

[0088]

[0089] Fig. 2 shows a schematic diagram of dynamic selection of the replanning end point. Figure 7 Fig. 3 shows a topological consistency diagram when selecting a parent node in the above algorithm. The improved kinematic RRT* algorithm can ensure that the replanning trajectory remains topologically consistent with the initial trajectory within the specified time t, thereby eliminating the velocity discontinuity when switching trajectories and completing trajectory replanning. Steps 4, 5, and 6 are repeated until the unmanned air-cushion vehicle reaches the target point.

[0090] The application firstly constructs a three-layer 2.5D DEM map, and then utilizes a hybrid A-star algorithm to perform global path searching, and the discrete path points searched by the hybrid A-star algorithm are fitted to form a continuous initial trajectory after optimization; when it is perceived that an unknown obstacle conflicts with the initial trajectory in the trajectory tracking process, a kinematically consistent RRT* algorithm is utilized to perform local path re-planning, and the topological consistency parameter tau is utilized to ensure that the front path of the re-planning is consistent with the original trajectory, thereby avoiding sudden change of speed when the trajectory is switched; in addition, the search area of the path re-planning algorithm is spatially constrained by the start and end points and the feasible region layer of the map, thereby reducing unnecessary node growth.

[0091] In particular, in some preferred embodiments of the application, a computer device is also provided, which includes a memory, a processor, and a computer program stored in the memory, and the processor implements the steps of the method for planning a cross-domain trajectory and dynamically avoiding collision of an unmanned hovercraft in any of the above embodiments when executing the computer program.

[0092] In some other preferred embodiments of the application, a computer readable storage medium is also provided, which stores a computer program / instructions, and the computer program is executed by a processor to implement the steps of the method for planning a cross-domain trajectory and dynamically avoiding collision of an unmanned hovercraft in any of the above embodiments.

[0093] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by a computer program instructing related hardware, and the computer program can be stored in a non-volatile computer readable storage medium, and the computer program can include the processes of the above-mentioned embodiments of the method for planning a cross-domain trajectory and dynamically avoiding collision of an unmanned hovercraft when executed, and the description is not repeated here.

[0094] In the description of the present specification, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, the skilled in the art can combine and combine the different embodiments or examples described in the present specification and the features of the different embodiments or examples without contradiction.

[0095] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0096] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more N executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of the invention pertain.

[0097] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0098] It should be understood that each part of the present application can be realized by hardware, software, firmware or their combination. In the above-mentioned embodiments, the N steps or methods can be realized by software or firmware stored in the memory and executed by a suitable instruction execution system. As if realized by hardware and in another embodiment, any one or their combination of the following technologies known in the art can be used: discrete logic circuit with logic gate circuit for implementing logic function on data signal, application specific integrated circuit with suitable combination logic gate circuit, programmable gate array (PGA), field programmable gate array (FPGA) and the like.

[0099] Those skilled in the art of the present technology can understand that all or part of the steps carried out by the above-mentioned embodiment method can be completed by a program instructing the relevant hardware, and the program can be stored in a computer readable storage medium, and when the program is executed, it includes one or a combination of the steps of the method embodiment.

[0100] In addition, each functional unit in each embodiment of the present application can be integrated in one processing module, or each unit can exist physically, or two or more units can be integrated in one module. The above-mentioned integrated module can be realized in the form of hardware or in the form of software functional module. The integrated module, if realized in the form of software functional module and sold or used as an independent product, can also be stored in a computer readable storage medium.

[0101] The above-mentioned storage medium can be read-only memory, disk or optical disk, etc. Although the embodiments of the present application have been shown and described above, it should be understood that the above-mentioned embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1. A method for cross-domain trajectory planning and dynamic collision avoidance of an unmanned hovercraft, characterized in that, include: A three-layer map is constructed by fusing the terrain segmentation layer, the feasible region layer, and the cost parameter layer of the unmanned hovercraft's movement. A hybrid A* algorithm is used for global path planning. The kinematic equations of water surface and land are used as constraints for node expansion to obtain a cross-domain discrete path from the starting point to the target point. Combined with the speed and acceleration constraints of the unmanned hovercraft, the trajectory that conforms to the kinematic and dynamic characteristics of the unmanned hovercraft is optimized. When an unmanned hovercraft is tracking a trajectory, if an unknown obstacle is found to conflict with the trajectory, the endpoint of the path replanning is first determined. Then, the topological consistency kinematics RRT* algorithm is used to search for the trajectory replanning. Local path planning is performed in combination with the speed and acceleration constraints of the unmanned hovercraft until the unmanned hovercraft reaches the endpoint.

2. The unmanned hovercraft cross-domain trajectory planning and dynamic collision avoidance method according to claim 1, characterized in that, The topographic geographic information is contained in the two-dimensional electronic nautical chart data. The two-dimensional electronic nautical chart data is binarized and rendered using the open-source geographic library Shapely to obtain the topographic segmentation layer of the map. Based on the obstacle-crossing capability of the unmanned hovercraft, the topographic geographic information is used to obtain the map layer of the beach landing feasible area and the land feasible area, i.e., the feasible domain layer of the map, and then the cost parameter layer of the hovercraft's movement is constructed. The obstacle-crossing capability of the unmanned hovercraft is: In the formula, θ max , These are the maximum pitch angle and maximum roll angle for the hovercraft's dynamic obstacle crossing.

3. The unmanned hovercraft cross-domain trajectory planning and dynamic collision avoidance method according to claim 1, characterized in that, The hard constraints for obstacle avoidance during the global and local path planning processes are as follows: Where υ represents the geometry of the unmanned hovercraft; P represents the feasible region, obtained from the feasible region layer; o d For dynamic obstacles; o i It is a static obstacle.

4. The unmanned hovercraft cross-domain trajectory planning and dynamic collision avoidance method according to claim 1, characterized in that, The equations of motion for the unmanned surface hovercraft and its equations of motion on land are as follows: In the formula, u is a constant longitudinal velocity, ψ is the heading angle, a is the longitudinal acceleration, and ψ max This is the maximum turning angle for a hovercraft traveling at high speed.

5. The unmanned hovercraft cross-domain trajectory planning and dynamic collision avoidance method according to claim 1, characterized in that, The cost function f for node expansion is: f(n) = g(n) + h(n) Where g is the path cost, which is the path length from the starting point to the current node multiplied by the penalty term; h is the heuristic function cost, which is the length of the Dubins curve from the current node to the target point.

6. The method for cross-domain trajectory planning and dynamic collision avoidance of unmanned hovercraft according to claim 1, characterized in that, The endpoint of the path replanning path is: S=a×λ+S0 Where S is the trajectory length from the initial starting point to the replanning termination point; S0 is the trajectory length from the initial starting point to the replanning starting point; a is the limit distance perceived by the sensor; and λ is the forward exploration parameter.

7. The method for cross-domain trajectory planning and dynamic collision avoidance of unmanned hovercraft according to claim 1, characterized in that, The cost function of the topology-consistent kinematics RRT* algorithm is: Among them, f cost The cost extended to the current node is given by the penalty function P; near The four-degree-of-freedom state represents the nearest neighbor node; α is the bow angle of the hovercraft, i.e., ψ; α near The orientation from the nearest neighbor node to the current node; α goal represents the orientation from the current node to the target node; m and n are the distance cost parameter and turning cost parameter, respectively, which change with the switching between land and water environments.

8. The method for cross-domain trajectory planning and dynamic collision avoidance of unmanned hovercraft according to claim 1, characterized in that, The discrete path kinematic constraints in the hybrid A* algorithm and the topology-consistent kinematics RRT* algorithm, which combine the speed and acceleration limitations of the unmanned hovercraft, are as follows: x k+1 -L(x k ,u k )=0,k=0,…,N-1, u min ≤u k ≤u max ,k=0,…,N, a min ≤a k ≤a max ,k=0,…,N-1. Where x is represented as {x,y,z,ψ} T This represents the four-degree-of-freedom state of the hovercraft, where u is denoted as {u,v,w,r}. T Here, L is the four-DOF velocity of the hovercraft, a is the kinematic mapping, a is the four-DOF acceleration of the hovercraft, N is the total number of discrete path points, and u is the velocity of the hovercraft. max and u min These are the maximum and minimum speed values, a. max and a min These are the maximum and minimum values ​​of acceleration, respectively. Among them, T f The time for the entire path tracking control is the time term cost J in the solution term of the optimization problem. t J input (u k This refers to controlling the cost of change.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method according to any one of claims 1 to 8.

Images