Air-ground cross-domain unmanned system cooperative reconnaissance and path planning method and system

Abstract

The invention provides an air-ground cross-domain unmanned system cooperative reconnaissance and path planning method and system, and the method comprises the steps: dynamically constructing an environment map according to the obtained position information of a target collected by an unmanned plane in an air-ground cross-domain unmanned system; based on the position information of the target and the environment map, performing path planning between a ground unmanned vehicle in the air-ground cross-domain unmanned system and the target by using an A * algorithm to obtain a driving path between the ground unmanned vehicle and the target; based on the driving path between the ground unmanned vehicle and the target, positioning and tracking the target; according to the invention, the unmanned aerial vehicle collects the target position information to provide the global environment sensing capability and quickly lock the position of the target; and orientation information is introduced into an A * algorithm to be combined with an environment map, so that obstacles can be effectively avoided in a complex dynamic environment, and the accuracy of dynamic environment reconnaissance and path planning is improved, and therefore, the autonomous motion navigation can be accurately realized by the method disclosed by the invention.

Description

Technical Field

[0001] The present invention relates to the field of robot intelligent control technology, and in particular to a method and system for collaborative reconnaissance and path planning of an air-to-ground cross-domain unmanned system. Background Art

[0002] At present, with the rapid development of unmanned system technology, unmanned systems have gained wide attention and application in the civil and military fields. Intelligent unmanned systems such as drones and unmanned vehicles have many advantages such as high security, low cost, high flexibility and long working hours, which can effectively improve work efficiency and safety. For example, ground unmanned vehicles have the advantages of mobility, high load carrying and long endurance, and can perform tasks for a long time on uneven terrain. However, due to its large limitations in environmental perception, it is difficult to quickly locate targets in completely unknown environments. In contrast, aerial drones have many advantages such as wide field of view, easy deployment, and rapid maneuverability. They can scout targets from multiple directions and angles, so as to quickly obtain key information such as target location. With the continuous expansion of application needs, real-time, changeable and complex mission scenarios have put forward more and higher requirements for unmanned systems. Therefore, making full use of unmanned systems in different spatial domains to form an organic whole and form a cross-domain unmanned system cluster is conducive to giving full play to the functional redundancy and capability complementarity of cross-domain and heterogeneous unmanned platforms, so as to achieve the purpose of effect interoperability and efficiency enhancement.

[0003] Cross-domain unmanned swarm systems have become a hot topic in the current field of unmanned systems due to their many advantages. In patrol and search and rescue missions, the environment in which unmanned systems are located is dynamic and unknown. Limited by the limited perception range and motion control capabilities, how cross-domain unmanned systems can achieve accurate target detection and fast and safe path planning is the key to achieving autonomous motion navigation of unmanned systems and improving the level of autonomous collaborative intelligence of multiple unmanned systems. It is also an important issue that must be solved for autonomous collaboration of unmanned swarm systems. Summary of the invention

[0004] In order to solve the problem that the coordination efficiency of UAVs and ground unmanned vehicles in the cross-domain unmanned swarm system in the prior art is low, resulting in low path planning accuracy and poor dynamic environment adaptability, especially being susceptible to obstacles in complex scenes, so that the existing unmanned swarm system cannot accurately realize autonomous motion navigation, the present invention proposes a collaborative reconnaissance and path planning method for an air-ground cross-domain unmanned system, including:

[0005] Dynamically construct an environmental map based on the location information of the targets collected by the drones in the air-to-ground cross-domain unmanned system;

[0006] Based on the location information of the target and the environment map, an A* algorithm is used to plan a path between the ground unmanned vehicle and the target in the air-to-ground cross-domain unmanned system to obtain a driving path between the ground unmanned vehicle and the target;

[0007] Positioning and tracking the target based on the driving path between the ground unmanned vehicle and the target;

[0008] Among them, the environmental map is constructed by fusing the SLAM algorithm and the Kalman filter algorithm; the A* algorithm plans the path between the ground unmanned vehicle and the target based on the orientation information of the ground unmanned vehicle.

[0009] Optionally, the location information of the target collected by the drone in the air-to-ground cross-domain unmanned system includes the following acquisition process:

[0010] Through the UAV in the air-to-ground cross-domain unmanned system, the target detection algorithm is used to perform target detection in the preset target area to obtain the target to be tracked;

[0011] The target tracking algorithm is used to track the target to be tracked, and the GPS world coordinate positioning information of the target to be tracked is calculated by the target positioning algorithm;

[0012] The GPS world coordinate positioning information of the target to be tracked is collected as the position information of the target.

[0013] Optionally, dynamically constructing an environment map according to the acquired location information of the target collected by the drone in the air-to-ground cross-domain unmanned system includes:

[0014] Determine a target area corresponding to the target location information according to the location information of the target collected by the drone in the air-to-ground cross-domain unmanned system;

[0015] Acquire environmental data collected by each sensor in the target area according to a preset time period;

[0016] Using a Kalman filter algorithm to fuse the environmental data collected by the sensors to obtain fused environmental data;

[0017] Based on the fused environmental data, a SLAM algorithm is used to construct a map of the target area to obtain an environmental map of the target area;

[0018] The environment map includes a global environment map and a local environment map, and the environment map is stored in a Costmap_2D storage structure.

[0019] Optionally, the step of performing path planning between the ground unmanned vehicle and the target in the air-to-ground cross-domain unmanned system by using an A* algorithm based on the location information of the target and the environment map to obtain a driving path between the ground unmanned vehicle and the target includes:

[0020] According to the location information of the target and the environment map, the A* algorithm is used to calculate the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target;

[0021] Using a path management algorithm based on distance judgment, the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target is updated to obtain an updated track point sequence path;

[0022] A driving path between the ground unmanned vehicle and the target is generated according to the updated track point sequence path.

[0023] Optionally, generating a driving path between the ground unmanned vehicle and the target according to the updated track point sequence path includes:

[0024] According to the updated track point sequence path, obtaining the traveling direction information of the ground unmanned vehicle in real time;

[0025] Determining whether the ground unmanned vehicle will encounter an obstacle according to the traveling direction information of the ground unmanned vehicle;

[0026] When the ground unmanned vehicle will not encounter obstacles, a path following algorithm is used to follow the global track point sequence path;

[0027] When the ground unmanned vehicle encounters an obstacle, a vector field histogram algorithm is used to perform collision avoidance on the ground unmanned vehicle, and the latest track point sequence path is obtained as the driving path between the ground unmanned vehicle and the target.

[0028] Optionally, the calculating, based on the location information of the target and the environment map, an optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target using an A* algorithm includes:

[0029] According to the location information of the target, loading the static obstacle information and the dynamic obstacle information corresponding to the target from the environment map;

[0030] Using the position coordinates of the ground unmanned vehicle and the orientation information under the position coordinates as the search starting point;

[0031] The position information of the target is used as the end point, and when it is detected that the target is located inside the obstacle, the track point located outside the obstacle and closest to the target is updated as the new end point through the target point dynamic update method;

[0032] Generate a global track point sequence according to the static obstacle information, the dynamic obstacle information, the search start point and the search end point;

[0033] According to the global track point sequence, an optimal sequence is output as an optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target.

[0034] Optionally, outputting an optimal sequence as an optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target according to the global track point sequence includes:

[0035] Using the cost function corresponding to the A* algorithm, calculate the cost function value of each track point in the global track point sequence;

[0036] The path composed of track points with the smallest cost function value is used as the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target.

[0037] Optionally, the cost function expression corresponding to the A* algorithm is as follows:

[0038] f(n)=g(n)+w(n)*h(n);

[0039] Among them, f(n) represents the cost function value of the ground unmanned vehicle from the starting track point to the target via n track points; g(n) represents the cost function value of the ground unmanned vehicle from the starting track point to the nth track point; h(n) represents the cost function value of the ground unmanned vehicle from the nth track point to the target; w(n) represents the orientation influence weight of the ground unmanned vehicle from the nth track point to the target.

[0040] Based on the same inventive concept, the present invention also provides an air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning system, comprising:

[0041] A map construction module is used to dynamically construct an environment map based on the location information of the target collected by the drone in the air-to-ground cross-domain unmanned system;

[0042] A path planning module is used to plan a path between the ground unmanned vehicle and the target in the air-to-ground cross-domain unmanned system based on the location information of the target and the environment map, and obtain a driving path between the ground unmanned vehicle and the target by using an A* algorithm;

[0043] A positioning and tracking module, used for positioning and tracking the target based on the driving path between the ground unmanned vehicle and the target;

[0044] Among them, the environmental map is constructed by fusing the SLAM algorithm and the Kalman filter algorithm; the A* algorithm plans the path between the ground unmanned vehicle and the target based on the orientation information of the ground unmanned vehicle.

[0045] Optionally, the path planning system further includes: an information acquisition module, including:

[0046] The target detection submodule is used to detect targets in a preset target area using a target detection algorithm through the UAV in the air-to-ground cross-domain unmanned system to obtain the target to be tracked;

[0047] The target tracking submodule is used to track the target to be tracked by using a target tracking algorithm, and calculate the GPS world coordinate positioning information of the target to be tracked by using a target positioning algorithm;

[0048] The information collection submodule is used to collect the GPS world coordinate positioning information of the target to be tracked as the position information of the target.

[0049] Optionally, the map construction module includes:

[0050] Constructing a region determination submodule, for determining a target region corresponding to the position information of the target according to the position information of the target collected by the drone in the air-to-ground cross-domain unmanned system;

[0051] An environmental data collection submodule, used to obtain environmental data collected by each sensor in the target area according to a preset time period;

[0052] A data fusion submodule is used to fuse the environmental data collected by the sensors using a Kalman filter algorithm to obtain fused environmental data;

[0053] An environmental map construction submodule is used to construct a map of the target area based on the fused environmental data using a SLAM algorithm to obtain an environmental map of the target area;

[0054] The environment map includes a global environment map and a local environment map, and the environment map is stored in a Costmap_2D storage structure.

[0055] Optionally, the path planning module includes:

[0056] A track point generation submodule is used to calculate the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target using the A* algorithm according to the position information of the target and the environment map;

[0057] A path updating submodule, used to update the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target by using a path management algorithm based on distance judgment, so as to obtain an updated track point sequence path;

[0058] The driving path generation submodule is used to generate a driving path between the ground unmanned vehicle and the target according to the updated track point sequence path.

[0059] Optionally, the driving path generation submodule includes:

[0060] A direction acquisition unit, used for acquiring the traveling direction information of the ground unmanned vehicle in real time according to the updated track point sequence path;

[0061] An obstacle judgment unit, used to judge whether the ground unmanned vehicle will encounter an obstacle according to the traveling direction information of the ground unmanned vehicle;

[0062] A path following unit, configured to follow the global track point sequence path using a path following algorithm when the ground unmanned vehicle will not encounter obstacles;

[0063] The collision avoidance unit is used to use a vector field histogram algorithm to perform collision avoidance on the ground unmanned vehicle when the ground unmanned vehicle encounters an obstacle, and obtain the latest track point sequence path as the driving path between the ground unmanned vehicle and the target.

[0064] Optionally, the track point generation submodule includes:

[0065] An obstacle loading unit, used to load static obstacle information and dynamic obstacle information corresponding to the target from the environment map according to the position information of the target;

[0066] A starting point setting unit, used to use the position coordinates of the ground unmanned vehicle and the orientation information under the position coordinates as a search starting point;

[0067] An end point setting unit, configured to use the position information of the target as the end point, and when it is detected that the target is located inside an obstacle, update the track point located outside the obstacle and closest to the target as a new end point by a target point dynamic updating method;

[0068] A sequence generating unit, configured to generate a global track point sequence according to the static obstacle information, the dynamic obstacle information, the search start point and the search end point;

[0069] The optimal path output unit is used to output the optimal sequence as the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target according to the global track point sequence.

[0070] Optionally, the optimal path output unit includes:

[0071] A cost calculation subunit, used to calculate the cost function value of each track point in the global track point sequence using the cost function corresponding to the A* algorithm;

[0072] The optimal path selection unit is used to use the path composed of track points with the smallest cost function value as the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target.

[0073] Optionally, the cost function expression corresponding to the A* algorithm is as follows:

[0074] f(n)=g(n)+w(n)*h(n);

[0075] Among them, f(n) represents the cost function value of the ground unmanned vehicle from the starting track point to the target via n track points; g(n) represents the cost function value of the ground unmanned vehicle from the starting track point to the nth track point; h(n) represents the cost function value of the ground unmanned vehicle from the nth track point to the target; w(n) represents the orientation influence weight of the ground unmanned vehicle from the nth track point to the target.

[0076] In another aspect, the present invention further provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus;

[0077] The memory is used to store one or more programs;

[0078] When the one or more programs are executed by the at least one processor, the aforementioned air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning method is implemented.

[0079] On the other hand, the present invention also provides a computer-readable storage medium having an execution program stored thereon. When the execution program is executed, the aforementioned air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning method is implemented.

[0080] Compared with the prior art, the present invention has the following beneficial effects:

[0081] The present invention provides a method and system for collaborative reconnaissance and path planning of an air-to-ground cross-domain unmanned system, comprising: dynamically constructing an environmental map according to the position information of a target collected by a drone in the air-to-ground cross-domain unmanned system; based on the position information of the target and the environmental map, using an A* algorithm to plan a path between a ground unmanned vehicle in the air-to-ground cross-domain unmanned system and the target, and obtaining a driving path between the ground unmanned vehicle and the target; based on the driving path between the ground unmanned vehicle and the target, positioning and tracking the target; wherein the A* algorithm plans a path between the ground unmanned vehicle and the target based on the orientation information of the ground unmanned vehicle; the environmental map is constructed by fusing a SLAM algorithm and a Kalman filter algorithm; the present application can provide global environmental perception capability and quickly lock the position of the target by collecting target position information through a drone; by combining the A* algorithm that introduces orientation information with the environmental map, obstacles can be effectively avoided in a complex dynamic environment, and the accuracy of dynamic environment reconnaissance and path planning can be improved. Therefore, the method of the present invention can accurately realize autonomous motion navigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 A schematic diagram of a flow chart of a method for collaborative reconnaissance and path planning of an air-to-ground cross-domain unmanned system provided by the present invention;

[0083] Figure 2 A schematic diagram of a collaborative path planning framework for a UAV and a ground unmanned vehicle in an air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning method provided by the present invention;

[0084] Figure 3 A track point sequence diagram for dynamic path management in an air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning method provided by the present invention;

[0085] Figure 4 A track point sequence diagram of a dynamic update method of target points used in an air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning method provided by the present invention;

[0086] Figure 5 A schematic diagram of target detection in which a target point is on or inside an obstacle in an air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning method provided by the present invention;

[0087] Figure 6 A target detection schematic diagram of adjusting a target point to a position that is closest to and reachable from an original target point by a dynamic target updating method in an air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning method provided by the present invention;

[0088] Figure 7A schematic diagram of a vector field histogram collision avoidance algorithm used in an air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning method provided by the present invention;

[0089] Figure 8 A calculation result diagram of the optimal obstacle avoidance direction in the air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning method provided by the present invention, when the partition angle is 10 degrees, the obstacle density threshold is 2 degrees / meter, and the threshold angle is 50 degrees, is the direction in which the head of the ground unmanned vehicle rotates 81 degrees to the right;

[0090] Fig. 9 A calculation result diagram of the optimal obstacle avoidance direction in the air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning method provided by the present invention, when the partition angle is 1 degree, the obstacle density threshold is 2 degrees / meter, and the threshold angle is 50 degrees, is the direction in which the head of the ground unmanned vehicle rotates 76 degrees to the right;

[0091] Fig.10 A working principle diagram of a PID path following algorithm used in an air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning method provided by the present invention;

[0092] Fig.11 A schematic diagram of the overall framework for collaborative path planning in an air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning method provided by the present invention;

[0093] Fig.12 This is a schematic diagram of node expansion of the traditional A* algorithm;

[0094] Fig.13 A schematic diagram of node expansion of the A* algorithm in a method for collaborative reconnaissance and path planning of an air-to-ground cross-domain unmanned system provided by the present invention;

[0095] Fig.14 A collaborative workflow diagram of an aerial unmanned aerial vehicle and a ground unmanned vehicle in an air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning method provided by the present invention;

[0096] Fig.15 A schematic diagram of the structure of an air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning system provided by the present invention;

[0097] Fig.16 The present invention provides a schematic structural diagram of an electronic device. DETAILED DESCRIPTION

[0098] The present invention proposes a method, system, device and medium for collaborative reconnaissance and path planning of an air-to-ground cross-domain unmanned system. The specific implementation methods of the present invention are further described in detail below in conjunction with the accompanying drawings.

[0099] Embodiment 1:

[0100] The present invention provides an air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning method, the flow chart is as follows Figure 1 As shown, including:

[0101] Step 1: Dynamically construct an environment map based on the location information of the target collected by the drone in the air-to-ground cross-domain unmanned system;

[0102] Step 2: Based on the location information of the target and the environment map, the A* algorithm is used to plan the path between the ground unmanned vehicle and the target in the air-to-ground cross-domain unmanned system to obtain the driving path between the ground unmanned vehicle and the target;

[0103] Step 3: Locate and track the target based on the driving path between the ground unmanned vehicle and the target;

[0104] Among them, the environmental map is constructed by fusion of SLAM algorithm and Kalman filter algorithm; the A* algorithm is based on the orientation information of the ground unmanned vehicle to plan the path between the ground unmanned vehicle and the target.

[0105] Generally, the air-to-ground cross-domain unmanned system is of great significance in the field of target detection, positioning and navigation because it can integrate the advantages of unmanned aerial vehicles and ground unmanned vehicles. However, due to the dynamic nature of targets in complex environments and the limitations of environmental maps, ground unmanned vehicles face problems such as insufficient path planning accuracy and inaccurate target tracking when performing autonomous navigation tasks. At the same time, traditional path planning algorithms usually ignore the orientation information and actual motion constraints of unmanned vehicles when generating paths, further limiting the feasibility and efficiency of the paths. In order to solve the above problems, this embodiment uses unmanned aerial vehicles to collect target location information, performs path planning based on high-precision environmental maps and improved A* algorithms, and combines the path results to accurately locate and track the target. Specifically:

[0106] In one implementation, the location information of the target collected by the drone in the above-mentioned air-to-ground cross-domain unmanned system may include the following acquisition process:

[0107] Through the UAV in the air-to-ground cross-domain unmanned system, the target detection algorithm is used to perform target detection in the preset target area to obtain the target to be tracked;

[0108] The target tracking algorithm is used to track the target to be tracked, and the GPS world coordinate positioning information of the target to be tracked is calculated through the target positioning algorithm;

[0109] The GPS world coordinate positioning information of the target to be tracked is collected as the target's location information;

[0110] In this implementation, the drones in the air-to-ground cross-domain unmanned system can efficiently perform wide-area search and reconnaissance missions and conduct regional coverage patrols. At the same time, the coordinated operation of target detection, tracking and positioning algorithms can achieve accurate identification of targets, continuous tracking and high-precision calculation of location information. When the drone quickly performs target detection tasks in a preset area, it can use advanced target detection algorithms (such as the YOLO algorithm). After detecting suspicious targets, it switches to target tracking mode and uses target tracking algorithms (such as the DaSiameseRPN algorithm) to continuously and stably track the target dynamically, ensuring real-time control of the target status in complex scenarios. If the tracking fails, the drone will automatically switch back to the target detection mode to continue the mission; if the tracking is successful, the monocular vision positioning algorithm is used to calculate the GPS world coordinate position of the target. In this process, the target positioning algorithm based on monocular vision plays a core role in the world coordinate positioning of the target detection results. The camera internal parameter information is obtained through the checkerboard calibration method, and the conversion relationship between the image coordinate system and the camera coordinate system is established. Combined with the external parameter information such as the camera pitch angle, roll angle, and heading angle, the relationship conversion between the camera coordinate system and the carrier coordinate system is completed. Finally, through the attitude information of the carrier platform and the world coordinate system conversion, the detection results are accurately converted to the GPS world coordinates of the target. This implementation method uses the rigor and accuracy of multi-level coordinate transformation to ensure the high reliability of the positioning results. After completing GPS positioning, the target data information is shared to the ground unmanned vehicle in real time through the target sharing mechanism within the cross-domain unmanned cluster, ensuring the rapid identification and accurate positioning of suspicious targets, and comprehensively improving the reconnaissance efficiency of collaborative targets. Therefore, in this implementation method, by leveraging the drone's global vision advantage and multi-algorithm collaboration, the drone can achieve accurate identification, dynamic tracking and location sharing of suspicious targets, while the ground unmanned vehicle uses high-quality target information for path planning and task execution, greatly improving the collaborative efficiency and task completion of the air-to-ground cross-domain unmanned system, and through the seamless connection of regional coverage patrol and target reconnaissance modules, it demonstrates excellent adaptability and task execution capabilities in complex dynamic scenarios.

[0111] Through the above implementation, the UAV can efficiently detect, track and locate the target, and share the target's location information with the ground unmanned vehicle in real time, providing a reliable basis for path planning. However, only the target location information cannot fully cope with the dynamic and complex environmental constraints, so it is necessary to further rely on high-precision environmental maps as the core support for path planning and target tracking. The construction of environmental maps requires the comprehensive use of multi-sensor data fusion technology and SLAM algorithms to achieve accurate modeling of the global and local environment. Specifically:

[0112] In one implementation, the process of dynamically constructing the environment based on the acquired location information of the target collected by the drone in the air-to-ground cross-domain unmanned system may include:

[0113] Determine a target area corresponding to the target location information according to the location information of the target collected by the drone in the air-to-ground cross-domain unmanned system;

[0114] Acquire environmental data collected by each sensor in the target area according to a preset time period;

[0115] The Kalman filter algorithm is used to fuse the environmental data collected by each sensor to obtain fused environmental data;

[0116] Based on the fused environmental data, the SLAM algorithm is used to construct a map of the target area to obtain an environmental map of the target area;

[0117] The environment map includes a global environment map and a local environment map, and the environment map is stored in a Costmap_2D storage structure.

[0118] In this implementation, in order to solve the problem of building environmental maps in complex and unknown situations, the Kalman filter algorithm is used to fuse multiple sensor data (such as GPS, inertial measurement unit IMU and odometer), combined with classic Gmapping SLAM and other algorithms, to efficiently build a high-precision unknown environmental map, which can capture and process data from different sensors, build an accurate environmental map, and can operate stably in a changeable and unknown environment, providing a solid data foundation for the autonomous navigation, path planning, navigation obstacle avoidance and intelligent decision-making of unmanned vehicles. When building the environmental map, the global environmental map and the local environmental map can be included, which respectively support the global path planning and local navigation obstacle avoidance of the unmanned system, ensuring that the system can also achieve accurate task execution in complex dynamic scenes. In addition, the construction of the environmental map adopts a layered architecture mode based on Costmap_2D for storage. The layered design includes a static map layer, a dynamic obstacle layer, an expansion layer and a main map layer, which can flexibly handle the information update of static known maps and dynamic perception maps, and improve the obstacle avoidance ability of the system through the configuration of obstacle safety distance. This layered architecture provides flexible and comprehensive environmental data support for unmanned systems, ensuring real-time adaptability in dynamic environments. Therefore, this implementation method has achieved significant improvements in environmental perception accuracy, real-time performance, and data management efficiency through the design of multi-sensor fusion and layered map storage. The global environmental map provides wide-area path planning capabilities, while the local environmental map enhances the obstacle avoidance capability of ground unmanned vehicles in complex scenarios through dynamic updates. In addition, the introduction of the expansion layer provides strong support for the dynamic adjustment of the safety distance, avoiding sudden risks caused by the approach of obstacles.

[0119] Through the above-mentioned environmental map construction process, multi-sensor data fusion and SLAM algorithm can be used to generate accurate global and local environmental maps, providing key support for path planning. The global environmental map provides a broad field of view for overall path planning, while the local environmental map is used to handle local obstacles and path adjustments in dynamic environments. Based on these high-precision maps, the location information of the target can be further combined to achieve autonomous navigation optimization of ground unmanned vehicles through path planning algorithms. Specifically:

[0120] In one implementation, the above-mentioned process of planning a path between a ground unmanned vehicle and a target in an air-to-ground cross-domain unmanned system based on the target location information and the environment map using the A* algorithm to obtain a driving path between the ground unmanned vehicle and the target may include:

[0121] According to the location information of the target and the environment map, the A* algorithm is used to calculate the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target;

[0122] Using the path management algorithm based on distance judgment, the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target is updated to obtain an updated track point sequence path;

[0123] Generate a driving path between the ground unmanned vehicle and the target based on the updated track point sequence path;

[0124] In this implementation, global path planning, path management, path following and collision avoidance are integrated. The framework diagram is shown in the figure. Figure 2 As shown in the figure, collaborative path planning adopts advanced global path planning technology, combined with dynamic path management and real-time path following, as well as efficient collision avoidance strategies, to achieve intelligent navigation and obstacle avoidance of unmanned systems. When faced with a situation where the target point is unreachable, the strategy can be flexibly adjusted to ensure that the unmanned system can quickly find an alternative path, significantly improving the adaptability of the unmanned system in complex environments and the success rate of task execution. In addition, by adopting the SLAM environment map based on multi-source heterogeneous sensor fusion and cross-domain collaborative sharing of target information, it not only ensures that the path planning results meet the motion constraints of the ground unmanned vehicle, but also can dynamically adapt to complex environmental changes. After receiving the target information, the ground unmanned vehicle first uses the improved A* algorithm to calculate the optimal global track point sequence path from the current point to the target point based on the environmental map, as shown in Figure 3 As shown in the figure, the blue dots are the track point sequences planned by the global path. When the unmanned system reaches the range of the blue circle shown in the figure, the target point followed by the current path is updated to the next track point. There is no direction information between the two track points. The path management realizes path following by adding orientation information. When generating global track points, Figure 4As shown in the figure, by using the dynamic update method of the target point, the unmanned system can be ensured to be as close to the target point as possible. The target point update method takes a point at a certain distance from the target point to the unmanned system, and then determines whether each point is located on an obstacle. If the new point is determined not to be inside the obstacle, then the point will replace the previous target point as the new target point. By combining the improved A* algorithm with the bow area coverage patrol strategy, the accuracy and efficiency of path planning are greatly improved, and the optimal global path can be generated. After the path planning is completed, the track point sequence path is dynamically updated through the path management method based on distance judgment to ensure that the path is always consistent with the real-time position and environmental conditions of the target. If the track point to be tracked is found to be inside an obstacle or unreachable during the update process, such as Figure 5 As shown in the figure, the target point is usually located in the center of the target ROI area, and the center of the target is surrounded by the target's own obstacle area, which causes the environment map construction module to construct the SLAM environment map equivalent to building the target point inside the obstacle, which ultimately leads to the failure of global path planning and the inability to effectively complete the cross-domain collaborative task. In this implementation, the dynamic target update method is used to adjust the target point to the nearest and reachable position from the original target point, such as Figure 6 As shown in the figure, the consistency and executability of the path are guaranteed. After the path is generated, the ground unmanned vehicle further performs the navigation task through path following and collision avoidance strategies, so that it can not only accurately reach the target, but also effectively avoid potential risks caused by obstacles or dynamic environmental changes. Therefore, this implementation method forms a full-process closed-loop optimization from path generation to dynamic adjustment and task execution through the organic combination of global path planning and path management, and has achieved significant improvements in path planning accuracy, environmental dynamic adaptability, and path safety. In particular, it effectively solves the problem of path unavailability caused by environmental changes or target position movement in dynamic and complex environments, fully reflecting the creative value and practical significance of this solution in the field of unmanned system path planning and control.

[0125] Through the above path planning and path management algorithm based on distance judgment, an updated track point sequence path can be generated and the path can be ensured to be real-time and effective in a dynamic environment. However, in actual driving, ground unmanned vehicles may encounter dynamic obstacles or other unforeseen situations. Therefore, it is necessary to further judge the potential collision risk and dynamically adjust the path based on the direction of travel of the unmanned vehicle. Specifically:

[0126] In one implementation, the process of generating a driving path between the ground unmanned vehicle and the target according to the updated track point sequence path may include:

[0127] According to the updated track point sequence path, the moving direction information of the ground unmanned vehicle is obtained in real time;

[0128] According to the moving direction information of the ground unmanned vehicle, determine whether the ground unmanned vehicle will encounter obstacles;

[0129] When the ground unmanned vehicle will not encounter obstacles, a path following algorithm is used to follow the global track point sequence path;

[0130] When the ground unmanned vehicle encounters an obstacle, the vector field histogram algorithm is used to avoid the collision of the ground unmanned vehicle and obtain the latest track point sequence path as the driving path between the ground unmanned vehicle and the target;

[0131] In this implementation, by obtaining the moving direction information of the ground unmanned vehicle in real time and dynamically judging the obstacle situation, combined with multiple algorithm designs for path following and collision avoidance, efficient path execution and dynamic adjustment of the ground unmanned vehicle in complex environments are achieved. In the process of following the global path planning track points, based on the dynamic target point update management mechanism, when the target point is identified to be inside the obstacle, the target point dynamic update algorithm will be automatically adopted. The ground unmanned vehicle will continue to detect obstacles in the forward direction. After taking points at equal distances, the obstacle cost value of the unmanned system at each point is calculated to determine whether there is an obstacle at the point. If the obstacle cost value is less than 0, it means that this target point is not located on the obstacle. The algorithm will replace the dynamically calculated target point with an obstacle cost value that is not 0 and is closest to the target point with a new target point, making the unreachable target point reachable, ensuring that the unmanned system reaches the target point as close as possible to perform the next stage of the task. If there is an obstacle ahead, the Vector Field Histogram (VFH) algorithm will be called to perform collision avoidance actions and dynamically adjust the current track point sequence path to ensure the safety and feasibility of the path. The principle diagram of the collision avoidance algorithm using the Vector Field Histogram algorithm is as follows: Figure 7As shown in the figure (Robot represents a ground unmanned vehicle; Histogram Grid represents the histogram grid part, which is used to represent the obstacle position in the surrounding environment; Polar Histogram represents the polar coordinate histogram part, which is used to represent the density of obstacles. The polar coordinate histogram is specifically divided into multiple Sectors k, which represents k direction sectors; ActiveWindow represents the active window, which is used to focus the sensor to collect data to capture the nearest obstacle information; CertaintyValues ​​represents the certainty value, which is used to indicate the possibility that the position is occupied by an obstacle; Ws represents the length of the scanning area; (Ws-1) / 2 represents the width of the scanning area), the VFH algorithm detects unknown obstacles through the vector field histogram to achieve real-time collision avoidance effect. The algorithm mainly uses a two-dimensional Cartesian histogram grid as a world model, divides the environment around the unmanned system in the form of a grid, and defines a certain range around the unmanned system as an active window for analysis. The VFH algorithm mainly includes three parts: (1) continuously collect and update sensor data to obtain information about surrounding obstacles; (2) construct a polar coordinate obstacle density histogram based on sensor data; (3) select a safe polar coordinate angle interval based on the polar coordinate obstacle density histogram and density threshold, combined with the minimum angle threshold for the unmanned platform to pass, to determine the subsequent movement direction and generate unmanned system control instructions. When there are no obstacles ahead of the path, the path following algorithm (such as PID or dynamic window DWA algorithm) is directly called to perform intelligent trajectory following on the track points, and control the ground unmanned vehicle to continuously approach the target point. In addition, when the ground unmanned vehicle is moving towards the target point, if the target image appears in the detection field of view, it will switch to the PID-based visual target guidance mode. By accurately controlling the motion posture and direction of the ground unmanned vehicle, it provides precise guidance when the unmanned vehicle approaches the target, effectively overcoming the target positioning deviation caused by collaborative target reconnaissance, achieving direct approach to the target, ensuring the accuracy and reliability of the unmanned system when performing reconnaissance or disposal tasks, and significantly improving the success rate and efficiency of task execution. This mechanism effectively solves the problem that the ground unmanned vehicle cannot accurately approach the target due to collaborative positioning deviation, thereby ensuring the smooth completion of cross-domain collaborative tasks; the path following algorithm ensures the navigation accuracy of the ground unmanned vehicle by intelligently following the trajectory of the track points after path management, and the VFH algorithm provides efficient dynamic obstacle avoidance capabilities during trajectory execution, enabling the unmanned vehicle to adapt to changes in complex dynamic environments in real time; in the process of collision avoidance of the unmanned vehicle using the vector field histogram algorithm, the polar coordinate obstacle density histogram under different parameter configurations is calculated by using the calculation sensor as the data source, such as Figure 8 The figure shows the calculation result of the optimal obstacle avoidance direction when the partition angle is 10 degrees, the obstacle density threshold is 2 degrees / meter, and the threshold angle is 50 degrees, and the unmanned vehicle's head rotates 81 degrees to the right; Fig. 9 The figure shows the calculation result when the partition angle is 1 degree, the obstacle density threshold is 2 degrees / meter, and the threshold angle is 50 degrees. The best obstacle avoidance direction is when the front of the unmanned vehicle rotates 76 degrees to the right. In practical applications, for safety and efficiency reasons, it is necessary to determine the partition angle, the angle threshold required for the passage of the unmanned system, and the obstacle density threshold based on the complexity of the environment and the characteristics of the unmanned platform itself (such as volume), so that the unmanned system can avoid obstacles smoothly while reducing the amount of calculation as much as possible to ensure that the unmanned system can respond to environmental changes in real time. The working principle diagram of the PID path following algorithm is shown in the figure. Fig.10 As shown in the figure, first, according to the image center coordinates and the image target position estimated coordinates, the visual guidance controller calculates the deviation value and solves the motion command to output to the unmanned system environment. The unmanned system then obtains the image stream from the environment through the onboard sensor, extracts and processes the image features, and feeds back the image target position estimated coordinates to the visual servo controller. This feedback process constitutes the main negative feedback system. The PID-based visual target guidance algorithm has slightly different visual guidance processes due to the different kinematic and dynamic limitations of aerial and ground unmanned systems: for aerial unmanned systems equipped with pods, the visual guidance collects images through the pods, and solves the linear speed of the drone in the X-axis and Y-axis directions according to the deviation values ​​of the image center point coordinates and the target estimated position coordinates on the X-axis and Y-axis; for ground unmanned systems equipped with depth cameras, the visual guidance algorithm collects images through the pods, and solves the linear speed of the drone in the X-axis and Y-axis directions according to the deviation values ​​of the image center point coordinates and the target estimated position coordinates on the X-axis and Y-axis. System, visual guidance uses a depth camera to collect depth images, and according to the deviation between the coordinates of the center point of the image and the estimated position coordinates of the target on the X-axis, combined with the depth information of the unmanned vehicle from the target point, the linear velocity of the ground unmanned system in the X-axis direction and the angular velocity of rotation in the Z-axis direction are calculated; the ground unmanned vehicle can approach dangerous goods under the control of the visual guidance algorithm and keep the dangerous goods target in the center of the horizontal field of view of the unmanned vehicle, so that the deviation of the distance between the unmanned system and the target center point is no more than 0.5 meters. The motion control under visual guidance is a small negative feedback system. When the visual servo controller gives the current motion command, the unmanned system controller controls the unmanned system to adjust the motion direction and speed in the current environment under the dynamic constraints of the unmanned system, and continuously approaches the target. The joint / angle sensor then feeds back to the unmanned system controller through the motion process. Therefore, by Fig.11As shown in the collaborative path planning framework diagram, this implementation method can realize global path planning, path management, path following and collision avoidance of cross-domain unmanned systems. The global path planning adopts the improved A* algorithm and the arc area coverage patrol algorithm to optimize the path calculation; secondly, the path management realizes the dynamic adjustment of the path by updating based on distance judgment and dynamic target update; the path following module uses algorithms such as PID and dynamic window (DWA) to ensure accurate path tracking; finally, the collision avoidance module uses algorithms based on vector field histogram (VFH) and artificial potential field (APF) to enhance the obstacle avoidance ability of the system. Therefore, this implementation method seamlessly connects path following and dynamic obstacle avoidance through the collaborative design of multiple algorithms organically combined with path following, collision avoidance and visual target guidance, providing a comprehensive and flexible path execution solution for unmanned vehicles, especially in complex dynamic environments, solving the path failure or task interruption caused by obstacles, environmental changes and target position offset, which fully reflects its innovation and practical value in the field of unmanned vehicle navigation and task execution.

[0132] Through the above path generation process, the ground unmanned vehicle can adjust the path according to the real-time direction of travel and obstacle information in a dynamic environment, and realize dynamic obstacle avoidance by combining the path following algorithm and the vector field histogram algorithm to ensure the safety and continuity of navigation. However, in order to further improve the efficiency and accuracy of path planning, it is necessary to combine the target location information and the dynamically constructed environment map, and use the A* algorithm to calculate the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target. Specifically:

[0133] In one implementation, the process of calculating the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target using the A* algorithm based on the target's location information and the environment map may include:

[0134] According to the location information of the target, the static obstacle information and dynamic obstacle information corresponding to the target are loaded from the environment map;

[0135] The position coordinates of the ground unmanned vehicle and the orientation information under the position coordinates are used as the search starting point;

[0136] The position information of the target is used as the end point. When it is detected that the target is located inside the obstacle, the track point closest to the target outside the obstacle is updated as the new end point through the target point dynamic update method;

[0137] Generate a global track point sequence based on static obstacle information, dynamic obstacle information, search start point and end point;

[0138] According to the global track point sequence, the optimal sequence is output as the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target;

[0139] In this implementation, the optimal path planning in a complex dynamic environment is realized by combining the current position information and orientation information of the ground unmanned vehicle with the position information of the target, and based on the dynamically constructed environment map, the improved A* algorithm is used. By loading the information of static obstacles and dynamic obstacles from the environment map, the factors that may affect the path planning in the environment can be accurately identified and evaluated, providing comprehensive data support for path planning; the orientation information of the unmanned vehicle is introduced into the search starting point of the path planning, which effectively solves the problem of low path execution efficiency caused by ignoring motion constraints in the traditional path planning method, making the generated path smoother and in line with the motion characteristics of the unmanned vehicle; through the improved A* algorithm cost function design, the cost weight is dynamically adjusted in the path planning, and the weight value is flexibly adjusted for grids in different directions, and the grid path consistent with the direction of the unmanned vehicle is preferentially selected. This design not only significantly reduces the phenomenon of sharp turns, reversing or rotating in place, but also optimizes the execution efficiency of path planning. After generating the global track point sequence, the optimal sequence path is further output through screening and optimization, thereby ensuring that the unmanned vehicle reaches the target point in the shortest time and the optimal path. This implementation method achieves a deep combination of environmental perception and path planning through multi-level path optimization: from the perception of dynamic obstacles to the generation of global paths, and then to the output of the optimal sequence path. Compared with the traditional A* algorithm, the improved algorithm significantly improves the efficiency and execution effect of path planning in complex environments, enabling unmanned vehicles to complete navigation tasks in a more efficient way in dynamic environments. This technical solution solves the problem of path unavailability caused by the disconnection between motion constraints and path planning in traditional methods, and at the same time demonstrates the strong adaptability of unmanned system path planning in dynamic and complex scenarios, fully reflecting the creative value and practical application potential of the technical solution.

[0140] Through the above process, the system can generate a global trackpoint sequence path that includes the influence of static and dynamic obstacles based on the target location information and the dynamically constructed environment map, and use it as the path planning basis for the unmanned vehicle from the current location to the target point. However, in order to ensure the optimization and optimality of the final path, it is necessary to further evaluate and screen the generated global trackpoint sequence according to the cost function of the A* algorithm. Specifically:

[0141] In one implementation, the process of outputting the optimal sequence as the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target according to the global track point sequence may include:

[0142] Using the cost function corresponding to the A* algorithm, calculate the cost function value of each track point in the global track point sequence;

[0143] The path composed of track points with the smallest cost function value is taken as the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target.

[0144] For example, the cost function expression corresponding to the above A* algorithm can be as follows:

[0145] f(n)=g(n)+w(n)*h(n);

[0146] Among them, f(n) represents the cost function value of the ground unmanned vehicle from the starting track point to the target via n track points; g(n) represents the cost function value of the ground unmanned vehicle from the starting track point to the nth track point; h(n) represents the cost function value of the ground unmanned vehicle from the nth track point to the target; w(n) represents the direction influence weight of the ground unmanned vehicle from the nth track point to the target; Fig.12 The figure shows a node expansion diagram of the traditional A* algorithm track point search method. Fig.13 The figure shows a node expansion diagram of the improved A* algorithm obtained by introducing the current orientation of the unmanned system into the traditional A* algorithm. In this example, the improved A* algorithm adds a dynamic adjustment mechanism of the weight w(n) in the cost function, so that the algorithm can more intelligently handle the orientation optimization problem in path planning. Within the exploration area, the grids that are consistent with the orientation of the unmanned vehicle are set with a lower weight value (w(n)<1), as shown in the yellow area, thereby reducing the path planning cost and giving priority to smoother paths; grids that are inconsistent with the orientation of the unmanned vehicle are set with a higher weight value (w(n)>1), as shown in the blue area, to avoid unnecessary sharp turns or changes in direction. Outside the exploration area, the weight value is set to 1, which is equivalent to the classic A algorithm. This design greatly optimizes the execution of the path, making the unmanned vehicle more in line with its motion constraints when planning the path. Through physical and simulation tests, it is verified that the improved A* algorithm can effectively avoid the phenomenon of reversing or rotating in place when excluding scenes with particularly dense obstacles, significantly shortening the path planning time of the unmanned vehicle from the starting point to the target point, and greatly improving the navigation efficiency. Compared with the classic A* algorithm, when there are multiple obstacles in the environment, the improved A* algorithm significantly reduces the frequency of vehicle head shaking after avoiding obstacles (the number of times the vehicle head sways left and right more than 3 times is significantly reduced), and the frequency of large steering angles (>120°) is also greatly reduced. In addition, due to the effective avoidance of vehicle head shaking, in the face of more obstacles, the improved A algorithm can complete the navigation task from the starting point to the target point in a shorter total time.

[0147] In summary, the present invention aims at the problem that the coordination efficiency of UAVs and ground unmanned vehicles in the cross-domain unmanned swarm system in the prior art is low, resulting in low path planning accuracy and poor adaptability to dynamic environments, especially being susceptible to interference from obstacles in complex scenes, making it impossible for the existing unmanned swarm system to accurately realize autonomous motion navigation. A method for coordinated reconnaissance and path planning of air-ground cross-domain unmanned systems is proposed, such as Fig.14 The figure shows the collaborative workflow between aerial drones and ground unmanned vehicles. The aerial drone patrols through the arched area, performs target detection and tracking tasks, and locates the target after target identification and confirmation, ensuring the rapid identification and precise positioning of suspicious targets, and comprehensively improving the accuracy and efficiency of collaborative target reconnaissance. At the same time, the target information is shared with the ground unmanned vehicle to achieve cross-domain collaboration. The ground unmanned vehicle uses the SLAM algorithm to build an environmental map and stores the path through Costmap_2D. Based on the target information provided by the drone, global path planning is carried out, and the improved A* algorithm is used for path calculation; then, through path management, the ground unmanned vehicle implements dynamic path adjustment to adapt to the changes in the real-time environment and target position. If there is an obstacle in front of the path, the VFH obstacle avoidance algorithm is triggered to ensure that the unmanned vehicle safely stops or detours; when the obstacle is removed, the ground unmanned vehicle continues to perform path tracking tasks based on algorithms such as PID until it successfully arrives or detects the target again, completes visual guidance control, and achieves accurate target tracking. The overall process reflects the advanced collaborative capability and real-time adaptability of the cross-domain unmanned cluster system. Therefore, the method of the present invention can improve the speed and accuracy of target positioning, enhance the adaptability and intelligence of path planning, and significantly improve the navigation capability and safety of unmanned vehicles in dynamic environments, thereby promoting the practicality and reliability of the cross-domain cluster unmanned system in diversified application scenarios.

[0148] Embodiment 2:

[0149] Based on the same inventive concept, the present invention also provides an air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning system, the structural composition diagram is as follows: Fig.15 As shown, including:

[0150] A map construction module is used to dynamically construct an environment map based on the location information of the target collected by the drone in the air-to-ground cross-domain unmanned system;

[0151] The path planning module is used to plan the path between the ground unmanned vehicle and the target in the air-to-ground cross-domain unmanned system based on the target's location information and environmental map, and obtain the driving path between the ground unmanned vehicle and the target using the A* algorithm;

[0152] A positioning and tracking module is used to locate and track the target based on the driving path between the ground unmanned vehicle and the target;

[0153] Among them, the environmental map is constructed by fusion of SLAM algorithm and Kalman filter algorithm; the A* algorithm is based on the orientation information of the ground unmanned vehicle to plan the path between the ground unmanned vehicle and the target.

[0154] In one implementation, the above-mentioned path planning system may further include: an information acquisition module, specifically including:

[0155] The target detection submodule is used to detect targets in a preset target area using a target detection algorithm through the UAV in the air-to-ground cross-domain unmanned system to obtain the target to be tracked;

[0156] The target tracking submodule is used to track the target to be tracked by using the target tracking algorithm, and calculate the GPS world coordinate positioning information of the target to be tracked by using the target positioning algorithm;

[0157] The information collection submodule is used to collect the GPS world coordinate positioning information of the target to be tracked as the position information of the target.

[0158] In one implementation, the map construction module includes:

[0159] Constructing a region determination submodule, for determining a target region corresponding to the position information of the target according to the position information of the target collected by the drone in the air-to-ground cross-domain unmanned system;

[0160] The environmental data acquisition submodule is used to obtain environmental data collected by each sensor in the target area according to a preset time period;

[0161] The data fusion submodule is used to fuse the environmental data collected by each sensor using the Kalman filter algorithm to obtain fused environmental data;

[0162] The environmental map construction submodule is used to construct a map of the target area based on the fused environmental data using the SLAM algorithm to obtain an environmental map of the target area;

[0163] The environment map includes a global environment map and a local environment map, and the environment map is stored in a Costmap_2D storage structure.

[0164] In one implementation, the above-mentioned path planning module may include:

[0165] The track point generation submodule is used to calculate the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target using the A* algorithm based on the target's location information and the environment map;

[0166] The path update submodule is used to update the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target by using the path management algorithm based on distance judgment, and obtain the updated track point sequence path;

[0167] The driving path generation submodule is used to generate the driving path between the ground unmanned vehicle and the target according to the updated track point sequence path.

[0168] In this implementation, the above-mentioned driving path generation submodule may include:

[0169] A direction acquisition unit, used to obtain the traveling direction information of the ground unmanned vehicle in real time according to the updated track point sequence path;

[0170] An obstacle judgment unit, used to judge whether the ground unmanned vehicle will encounter an obstacle based on the traveling direction information of the ground unmanned vehicle;

[0171] The path following unit is used to follow the global track point sequence path using a path following algorithm when the ground unmanned vehicle will not encounter obstacles;

[0172] The collision avoidance unit is used to avoid collisions of the ground unmanned vehicle when the ground unmanned vehicle encounters obstacles. The vector field histogram algorithm is used to obtain the latest track point sequence path as the driving path between the ground unmanned vehicle and the target.

[0173] In one implementation, the track point generation submodule may include:

[0174] An obstacle loading unit, used to load static obstacle information and dynamic obstacle information corresponding to the target from the environment map according to the position information of the target;

[0175] A starting point setting unit, used to use the position coordinates of the ground unmanned vehicle and the orientation information under the position coordinates as the search starting point;

[0176] An end point setting unit is used to use the position information of the target as the end point, and when it is detected that the target is located inside the obstacle, update the track point located outside the obstacle and closest to the target as the new end point through a target point dynamic update method;

[0177] A sequence generation unit, used to generate a global track point sequence according to static obstacle information, dynamic obstacle information, a search start point and an end point;

[0178] The optimal path output unit is used to output the optimal sequence as the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target according to the global track point sequence.

[0179] In this implementation, the above-mentioned optimal path output unit may include:

[0180] The cost calculation subunit is used to calculate the cost function value of each track point in the global track point sequence using the cost function corresponding to the A* algorithm;

[0181] The optimal path selection unit is used to use the path composed of track points with the smallest cost function value as the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target.

[0182] For example, the cost function expression corresponding to the above A* algorithm can be as follows:

[0183] f(n)=g(n)+w(n)*h(n);

[0184] Among them, f(n) represents the cost function value of the ground unmanned vehicle from the starting track point to the target via n track points; g(n) represents the cost function value of the ground unmanned vehicle from the starting track point to the nth track point; h(n) represents the cost function value of the ground unmanned vehicle from the nth track point to the target; w(n) represents the direction influence weight of the ground unmanned vehicle from the nth track point to the target.

[0185] The system adopts a flexible modular architecture, allowing each module to be independently upgraded and optimized to adapt to the ever-changing technical requirements and mission environment. Through decoupling between modules, rapid iteration and seamless integration can be achieved, ensuring the long-term adaptability and continuous update capability of the system.

[0186] Embodiment 3:

[0187] like Fig.16 As shown, the present invention also provides an electronic device, which may be a computer device, a single-chip device, an intelligent mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, the processor, and the transceiver component are connected via a bus; the memory may be used to store an execution program, and an exemplary execution program may include instructions; the processor is used to execute the instructions stored in the memory. The memory may also be used to store data, which may be called and / or modified when the instructions are executed.

[0188] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in a storage medium to implement the corresponding method flow or corresponding functions, so as to realize the steps of a method for collaborative reconnaissance and path planning of an air-to-ground cross-domain unmanned system in the above-mentioned embodiment.

[0189] Embodiment 4:

[0190] Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It is understandable that the storage medium here may include both a built-in storage medium in an electronic device and an extended storage medium supported by the electronic device. The storage medium provides a storage space that stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions may be one or more execution programs (including program codes). It should be noted that the storage medium here may be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a method for collaborative reconnaissance and path planning of an air-to-ground cross-domain unmanned system in the above-mentioned embodiment.

[0191] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0192] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0193] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0194] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0195] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit its protection scope. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that after reading the present invention, those skilled in the art can still make various changes, modifications or equivalent substitutions to the specific implementation methods of the application, but these changes, modifications or equivalent substitutions are all within the protection scope of the claims to be approved.

Claims

1. A method for collaborative reconnaissance and path planning of air-to-ground cross-domain unmanned systems, characterized in that: include: Dynamically construct an environmental map based on the location information of the targets collected by the drones in the air-to-ground cross-domain unmanned system; Based on the location information of the target and the environment map, an A* algorithm is used to plan a path between the ground unmanned vehicle and the target in the air-to-ground cross-domain unmanned system to obtain a driving path between the ground unmanned vehicle and the target; Positioning and tracking the target based on the driving path between the ground unmanned vehicle and the target; Among them, the environmental map is constructed by fusing the SLAM algorithm and the Kalman filter algorithm; the A* algorithm plans the path between the ground unmanned vehicle and the target based on the orientation information of the ground unmanned vehicle.

2. The method according to claim 1, characterized in that The location information of the target collected by the drone in the air-to-ground cross-domain unmanned system includes the following acquisition process: Through the UAV in the air-to-ground cross-domain unmanned system, the target detection algorithm is used to perform target detection in the preset target area to obtain the target to be tracked; The target tracking algorithm is used to track the target to be tracked, and the GPS world coordinate positioning information of the target to be tracked is calculated by the target positioning algorithm; The GPS world coordinate positioning information of the target to be tracked is collected as the position information of the target.

3. The method according to claim 1 or 2, characterized in that The method of dynamically constructing an environment map based on the acquired location information of the target collected by the drone in the air-to-ground cross-domain unmanned system includes: Determine a target area corresponding to the target location information according to the location information of the target collected by the drone in the air-to-ground cross-domain unmanned system; Acquire environmental data collected by each sensor in the target area according to a preset time period; Using a Kalman filter algorithm to fuse the environmental data collected by the sensors to obtain fused environmental data; Based on the fused environmental data, a SLAM algorithm is used to construct a map of the target area to obtain an environmental map of the target area; The environment map includes a global environment map and a local environment map, and the environment map is stored in a Costmap_2D storage structure.

4. The method according to claim 1, characterized in that The method of performing path planning between the ground unmanned vehicle and the target in the air-to-ground cross-domain unmanned system based on the location information of the target and the environment map using the A* algorithm to obtain a driving path between the ground unmanned vehicle and the target includes: According to the location information of the target and the environment map, the A* algorithm is used to calculate the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target; Using a path management algorithm based on distance judgment, the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target is updated to obtain an updated track point sequence path; A driving path between the ground unmanned vehicle and the target is generated according to the updated track point sequence path.

5. The method according to claim 4, characterized in that The step of generating a driving path between the ground unmanned vehicle and the target according to the updated track point sequence path includes: According to the updated track point sequence path, obtaining the traveling direction information of the ground unmanned vehicle in real time; Determining whether the ground unmanned vehicle will encounter an obstacle according to the traveling direction information of the ground unmanned vehicle; When the ground unmanned vehicle will not encounter obstacles, a path following algorithm is used to follow the global track point sequence path; When the ground unmanned vehicle encounters an obstacle, a vector field histogram algorithm is used to perform collision avoidance on the ground unmanned vehicle, and the latest track point sequence path is obtained as the driving path between the ground unmanned vehicle and the target.

6. The method according to claim 4, characterized in that The method of calculating the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target using the A* algorithm according to the location information of the target and the environment map includes: According to the location information of the target, loading the static obstacle information and the dynamic obstacle information corresponding to the target from the environment map; Using the position coordinates of the ground unmanned vehicle and the orientation information under the position coordinates as the search starting point; The position information of the target is used as the end point, and when it is detected that the target is located inside the obstacle, the track point located outside the obstacle and closest to the target is updated as the new end point through the target point dynamic update method; Generate a global track point sequence according to the static obstacle information, the dynamic obstacle information, the search start point and the search end point; According to the global track point sequence, an optimal sequence is output as an optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target.

7. The method according to claim 6, characterized in that Outputting an optimal sequence as an optimal global track point sequence path from the current coordinate point of the ground unmanned vehicle to the target according to the global track point sequence includes: Using the cost function corresponding to the A* algorithm, calculate the cost function value of each track point in the global track point sequence; The path composed of track points with the smallest cost function value is used as the optimal global track point sequence path between the current coordinate point of the ground unmanned vehicle and the target.

8. The method according to claim 7, characterized in that The cost function expression corresponding to the A* algorithm is as follows: f(n)=g(n)+w(n)*h(n); Among them, f(n) represents the cost function value of the ground unmanned vehicle from the starting track point to the target via n track points; g(n) represents the cost function value of the ground unmanned vehicle from the starting track point to the nth track point; h(n) represents the cost function value of the ground unmanned vehicle from the nth track point to the target; w(n) represents the orientation influence weight of the ground unmanned vehicle from the nth track point to the target.

9. An air-to-ground cross-domain unmanned system collaborative reconnaissance and path planning system, characterized in that: include: A map construction module is used to dynamically construct an environment map based on the location information of the target collected by the drone in the air-to-ground cross-domain unmanned system; A path planning module is used to plan a path between the ground unmanned vehicle and the target in the air-to-ground cross-domain unmanned system based on the location information of the target and the environment map, and obtain a driving path between the ground unmanned vehicle and the target by using an A* algorithm; A positioning and tracking module, used for positioning and tracking the target based on the driving path between the ground unmanned vehicle and the target; Among them, the environmental map is constructed by fusing the SLAM algorithm and the Kalman filter algorithm; the A* algorithm plans the path between the ground unmanned vehicle and the target based on the orientation information of the ground unmanned vehicle.

10. The system according to claim 9, characterized in that The path planning system further includes: an information acquisition module, including: The target detection submodule is used to detect targets in a preset target area using a target detection algorithm through the UAV in the air-to-ground cross-domain unmanned system to obtain the target to be tracked; The target tracking submodule is used to track the target to be tracked by using a target tracking algorithm, and calculate the GPS world coordinate positioning information of the target to be tracked by using a target positioning algorithm; The information collection submodule is used to collect the GPS world coordinate positioning information of the target to be tracked as the position information of the target.

Images