Intelligent agent stealth maneuvering method, system, device, and readable storage medium

Abstract

The application provides an agent concealment maneuvering method, system, device and readable storage medium, which comprises the following steps: determining a tangent plane of an obstacle in a target area, the tangent plane being perpendicular to a line connecting a geometric center of the obstacle and a detection point; determining a first projection of the obstacle on the tangent plane with the detection point as a projection center, and determining a blind area corresponding to the obstacle based on a boundary of the first projection and a straight line connecting the detection point and the boundary; determining a reference path in the target area based on a path cost function, the path cost function comprising a cost of the path passing through the blind area, the obstacle and an exposure area, the exposure area being an area in the target area except the blind area and the obstacle; and constructing a simulation force field in the target area with the reference path as an attractive force source and the obstacle as a repulsive force source to generate a concealment maneuvering path of the agent. Compared with the prior art, the application solves the problems of insufficient concealment consideration and weak environmental adaptability in the prior art.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of intelligent control, and particularly relates to an agent concealment maneuver method, system, device and readable storage medium. BACKGROUND

[0002] Agent concealment maneuver is a key means to improve task success rate. With the rapid development of unmanned platforms and simulation technology, agents are widely used to replace or assist personnel in performing concealment tasks. How to make the agent autonomously plan and execute a maneuver path with optimal concealment in a complex environment has become a core technical challenge to improve the realism of unmanned system combat and simulation deduction.

[0003] When facing the specific needs of agent concealment maneuver, the existing technology still has obvious deficiencies: in the aspect of environment perception and modeling, most technologies focus on geometric accessibility analysis, and lack of detailed modeling and fusion of detection threats. The agent can only avoid physical obstacles, but cannot quantitatively evaluate the "exposure" risk of being discovered during movement, resulting in that the planned path is not optimal in tactics. In the aspect of concealment area calculation, the existing methods are mostly based on two-dimensional plane or simplified line-of-sight analysis, and it is difficult to accurately calculate the global blind area under the joint action of multiple detection points in a three-dimensional complex environment, especially the shape and range of the blind area generated by irregular obstacles are not accurate enough, which affects the reliability of the concealment path. In the aspect of path planning, traditional algorithms take the shortest path or the shortest time as the optimization target, and do not take "exposure" as the core cost function. There is a lack of online re-planning and control ability to deal with real-time changes in the environment.

[0004] Therefore, in view of the above technical problems, it is necessary to provide an agent concealment maneuver method, system, device and readable storage medium. SUMMARY

[0005] The purpose of the present application is to provide an agent concealment maneuver method, system, device and readable storage medium, which can effectively improve the concealment of the simulation path.

[0006] In order to achieve the above-mentioned purpose, the technical solutions provided by the present application are as follows:

[0007] In a first aspect, the present application provides an agent concealment maneuvering method, comprising: determining a tangent plane of an obstacle in a target region, the tangent plane being perpendicular to a line connecting a geometric center of the obstacle and a probe point; determining a first projection of the obstacle on the tangent plane with the probe point as a projection center, and determining a blind area corresponding to the obstacle based on a boundary of the first projection and a straight line connecting the probe point and the boundary; determining a reference path in the target region based on a path cost function, the path cost function including costs of the path passing through the blind area, the obstacle and an exposure area, the exposure area being an area in the target region excluding the blind area and the obstacle; and constructing a simulation force field in the target region to generate a concealment maneuvering path of an agent, with the reference path as an attractive force source and the obstacle as a repulsive force source.

[0008] In one or more embodiments of the present application, constructing a simulation force field in the target region to generate a concealment maneuvering path comprises: determining a resultant force on the current position of the agent in the simulation force field; determining a position of the agent at a next time based on the magnitude of the resultant force, the direction of the resultant force, a preset speed and an iteration period; and updating the position at the next time as the current position until the current position is a preset terminal point.

[0009] In one or more embodiments of the present application, the method further comprises: determining a nearest point of the current position of the agent on the reference path; determining that the attractive force of the reference path is invalid in an interval from a starting point to the nearest point; and determining the resultant force on the current position in the simulation force field based on the repulsive force source and the remaining valid attractive force sources.

[0010] In one or more embodiments of the present application, determining a tangent plane of an obstacle in a target region comprises: determining a plane that is perpendicular to a line connecting a probe point and a geometric center of the obstacle and tangent to the obstacle, to form a candidate plane set; and determining a plane farthest from the probe point in the candidate plane set as the tangent plane of the obstacle.

[0011] In one or more embodiments of the present application, determining a reference path in the target region based on a path cost function comprises: generating a navigation grid in the target region, the navigation grid including a plurality of nodes and edges connecting the nodes; determining a cost corresponding to each edge based on a region type in which the edge is located, the region type including a blind area, an obstacle area and an exposure area; and determining a path with a minimum cumulative cost from a starting point to a terminal point as the reference path based on the costs corresponding to the edges.

[0012] In one or more embodiments of the present application, the cost corresponding to the edge is:

[0013]

[0014]

[0015]

[0016]

[0017] in, For point Euclidean distance to the detection point; It is the distance decay function; The attenuation coefficient; The benchmark cost coefficient; , These are the start and end points of an edge in the navigation mesh, respectively; A parameterized representation of any point on an edge in the navigation mesh; These are linear interpolation parameters; , These represent the costs of the blind spot and the exposure zone, respectively.

[0018] In one or more embodiments of the present invention, the cost function corresponding to the edge is:

[0019]

[0020]

[0021]

[0022]

[0023]

[0024] in, , These are the start and end points of an edge in the navigation mesh, respectively; A parameterized representation of any point on an edge in the navigation mesh; These are linear interpolation parameters; It is a point by Detection points The instantaneous probability of discovery; For point To the Detection points The Euclidean distance; Coefficient and ; Any point on the edge affected by multiple probe points Cost density; a risk cost scaling coefficient, a base movement cost constant, a total number of probe points; a cost function corresponding to an edge of the connection node.

[0025] In one or more embodiments of the present application, the path with the minimum cumulative cost from the start point to the end point is determined as the reference path based on the cost corresponding to each edge, comprising: in the navigation grid, taking the start point as an expansion node, traversing the neighbor nodes connected thereto, determining a first path cost and a second path cost, wherein the first path cost represents the cost from the start point to the neighbor node via the expansion node; the second path cost represents the estimated cost from the neighbor node to the end point; determining whether to update the optimal path cost of the neighbor node based on the first path cost; if yes, determining the expansion node as the parent node of the neighbor node; based on the first path cost and the second path cost, selecting the expansion node with the minimum sum of the first path cost and the second path cost from the backup topology nodes with parent nodes and not yet as expansion nodes, until the parent node of the end point is determined; based on the parent node relationship, determining the reference path in the navigation grid.

[0026] In one or more embodiments of the present application, the second path cost is:

[0027]

[0028]

[0029]

[0030]

[0031]

[0032] wherein, , are the neighbor node and the end point, respectively; is a parametric representation of any point between the neighbor node and the end point; is a linear interpolation parameter; is a point discovered by the first probe point at the first time; is the Euclidean distance from the point to the first probe point ; is a coefficient and ; Any point on the edge affected by multiple probe points Cost density; This is a scaling factor for risk costs. Based on the fundamental movement cost constant, The total number of detection points; This is the cost of the second path.

[0033] In one or more embodiments of the present invention, determining the blind zone corresponding to the obstacle includes: taking the detection point as the projection center, determining the second projection of the obstacle on the boundary of the region; determining the region surrounded by the side surface, the first projection, and the second projection as the blind zone of the obstacle; wherein, the side surface is a curved surface or plane formed by a straight line connecting the detection point and the boundary of the first projection sweeping between the first projection and the second projection.

[0034] Secondly, the present invention provides a stealthy maneuvering system for an intelligent agent, comprising: a first blind zone calculation module, configured to determine a tangent plane of an obstacle within a target area, the tangent plane being perpendicular to the line connecting the geometric center of the obstacle and a detection point; a second blind zone calculation module, configured to determine a first projection of the obstacle on the tangent plane with the detection point as the projection center, and to determine a blind zone corresponding to the obstacle based on the boundary of the first projection and a straight line connecting the detection point and the boundary; a path planning module, configured to determine a reference path within the target area based on a path cost function, the path cost function including the cost of the path passing through the blind zone, the obstacle, and the exposure zone, the exposure zone being the area within the target area excluding the blind zone and the obstacle; and a maneuvering implementation module, configured to construct a simulated force field within the target area using the reference path as a gravitational source and the obstacle as a repulsive source to generate a stealthy maneuvering path for the intelligent agent.

[0035] Thirdly, the present invention provides an electronic device comprising: at least one processor; and a memory storing instructions that, when executed by the at least one processor, cause the at least one processor to perform the agent covert maneuvering method.

[0036] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the aforementioned agent covert maneuvering method.

[0037] Compared with the prior art, the intelligent body concealment maneuvering method provided by the application realizes the leap of intelligent body path planning from "physical obstacle avoidance" to "tactical concealment". The method can not only generate a globally optimal reference path for concealment, but also realize real-time obstacle avoidance and smooth tracking in a dynamic environment through simulation of a force field, and can improve the motion efficiency by invalidating the gravity of the path segment that has been passed. The method solves the problems of insufficient concealment consideration, weak environmental adaptability and inaccurate model in the prior art, forms a complete and reliable intelligent body concealment maneuvering technology closed loop, and significantly improves the survivability, success rate and application value of the intelligent body in anti-terrorism simulation and actual combat tasks. BRIEF DESCRIPTION OF DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0039] Figure 1 is a schematic diagram of the implementation scene of the intelligent body concealment maneuvering method in an embodiment of the present application;

[0040] Figure 2 is a flowchart of the intelligent body concealment maneuvering method in an embodiment of the present application;

[0041] Figure 3 is a structural block diagram of the intelligent body concealment maneuvering system in another embodiment of the present application;

[0042] Figure 4 is a structural block diagram of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0043] In order to make the technical personnel in the art better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should be within the scope of protection of the present application.

[0044] Unless otherwise clearly indicated, in the entire specification and claims, the term "comprise" or its variants such as "contain" or "include" and the like will be understood to include the stated element or component, but not exclude other elements or components.

[0045] In the simulation scene, the path planning of the agent usually takes physical obstacle avoidance as the main optimization goal, and develops an avoidance strategy for obstacles and the like on this basis. The obstacle avoidance mechanism is highly dependent on the geometric modeling of the environment and obstacles, and the search and generation of feasible paths are realized through corresponding algorithms. However, the inventors found through in-depth analysis of the actual environment and diversified needs that the existing technology generally has a key defect: it fails to take "detected risk" as the core optimization goal of path planning. In actual combat, damage often comes from discovery, and detection is the fundamental prerequisite for damage to occur. Therefore, on the one hand, the path planned by the existing method may be too high in exposure; on the other hand, from the efficiency point of view, actively implementing agent concealment maneuver can significantly reduce the loss rate compared to passive obstacle avoidance after exposure.

[0046] Since traditional path planning mainly focuses on geometric accessibility, shortest path or shortest time and the like, the key factor of "detected risk" is often excluded from the decision system. Even if the user tries to intervene in the path of the agent, it is often difficult to systematically and scientifically generate a motion trajectory with concealment and feasibility without weakening its inherent obstacle avoidance ability.

[0047] Based on the above understanding, the present application proposes a new technical implementation path from the fundamental needs of agent concealment maneuver. First, by accurately solving the spatial relationship between obstacles and detection points in the environment, the distribution and range of the blind area are determined, and then the entire action space is divided into three types of regions: blind area, obstacle area and exposure area. On this basis, different cost functions are assigned to different region types, and the cost is taken as the core weight in the path planning algorithm to guide the generation of an initial maneuver route with optimal concealment.

[0048] To further maintain and enhance the real-time obstacle avoidance ability in traditional path planning, the present application introduces a simulated force field mechanism after generating the initial path. The force field takes the planned path and the end position as the source of gravity, and the obstacles perceived in real time as the source of repulsion. By calculating the direction and size of the resultant force, the actual motion of the agent is dynamically guided. In this way, the system not only inherits the tactical advantages of the global concealed path, but also has the ability to respond and avoid local obstacles, thereby realizing stable, safe and concealed maneuver control in complex dynamic environments.

[0049] Please refer to Figure 1 , which shows an application scene schematic diagram of the agent concealment maneuver method provided by the present application in an embodiment. The scene specifically includes a simulation engine 101, a server 102 and a user terminal 103.

[0050] It is to be noted that a communication connection is provided between the simulation engine 101, the server 102 and the user terminal 103, and the communication network extended by the communication connection can include various connection types, including but not limited to wired connection, wireless connection or fiber cable connection, etc. Meanwhile, the communication network can be a local area network, a metropolitan area network, a wide area network or any combination of the three.

[0051] The simulation engine 101 is a core computing and simulation deduction hub. It is responsible for loading and running a high-precision digital environment model, and accurately executing the complete set of intelligent agent concealment maneuver algorithms proposed in the present application. Specifically, it includes receiving and processing site data, performing three-dimensional blind area calculation; performing global path planning based on a minimum exposure cost function; constructing and updating a simulation force field in real time, dynamically generating an execution trajectory of the intelligent agent; and finally driving the virtual intelligent agent to perform realistic intelligent agent concealment maneuver in the simulation environment. It is a direct carrier of algorithm logic and physical simulation.

[0052] The server 102 is a data, logic and communication transfer and coordination hub. It is used to manage user sessions, schedule computing tasks, maintain consistency of simulation states, and handle network communication between the simulation engine 101 and multiple user terminals 103. The specific functions mainly include: distributing initial tasks and parameters to the simulation engine 101; receiving instructions from the user terminal 103; pushing intermediate results and final states generated by the simulation engine 101 to the corresponding user terminal 103 in real time.

[0053] The user terminal 103 is a man-machine interaction and decision monitoring interface. It is usually an operator terminal for commanders or trainers. The user imports or defines the training scenario through a graphical interface; monitors the three-dimensional visualization of the entire intelligent agent concealment maneuver process in real time; makes limited intervention or adjusts parameters during the deduction process; receives and analyzes simulation result reports for tactical evaluation and decision support.

[0054] It is to be noted that the user terminal 103 can be installed with a computer software program matched with the intelligent agent concealment maneuver method provided by the present method; the user terminal 103 can include but is not limited to a desktop computer (PC terminal), a desktop computer, a smart phone, a handheld computer, a tablet computer, a personal digital assistant (PDA) and other portable electronic devices or wearable electronic devices, and the embodiments of the present application do not limit the above content.

[0055] It should be noted that the server 102 can be a stand-alone physical server, a server cluster composed of multiple physical servers, or a distributed system, and can also be a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and basic cloud computing services such as big data and artificial intelligence platforms, and the embodiments of the present application do not limit this.

[0056] Please refer to Figure 2 Fig. 1 shows a flowchart of an embodiment of the present application. The intelligent agent concealment maneuvering method specifically includes the following steps:

[0057] S201: Determine the tangent plane of the obstacle in the target area, and the tangent plane is perpendicular to the line connecting the geometric center of the obstacle and the detection point;

[0058] This step aims to establish a key geometric reference plane, i.e., the tangent plane, for subsequent blind area calculation. The determination of the tangent plane is the basis for accurately calculating the spatial range of the blind area formed by the obstacle.

[0059] The tangent plane is used for subsequent perspective projection of the obstacle to generate the proximal boundary of the blind area. The core geometric constraint is that the plane must satisfy the perpendicularity condition and the tangent condition. The perpendicularity condition means that the plane must be perpendicular to the line connecting the detection point and the geometric center of the obstacle, which is defined as the first coordinate axis of the reference coordinate system in the embodiment of the present application, representing the main line-of-sight direction from the detection point to the obstacle body or the average spatial position; the tangent condition means that the plane must be tangent to the surface of the obstacle. Selecting a plane that satisfies both conditions can ensure that the line of sight from the detection point is at the spatial position that just grazes the edge of the obstacle, thereby providing a computationally feasible reference plane for accurately defining the boundaries between visible and invisible.

[0060] It should be noted that the line connecting the detection point and the geometric center of the obstacle is selected as the reference direction of the perpendicularity condition based on the comprehensive consideration of computational robustness and geometric representativeness. For an arbitrarily shaped obstacle, the visual occlusion effect from a single viewpoint corresponds to a range of spatial angles, rather than a single direction. The geometric center is a stable, low-sensitivity statistic that represents the overall spatial distribution of the obstacle. The reference direction defined in this way can effectively smooth the directional noise caused by local concave-convex features of the obstacle surface, avoid reference direction mutations caused by individual extreme vertices, and thus provide a consistent, reliable, and easily automated spatial reference framework for the entire blind area calculation process, ensuring that the method still has good adaptability and computational stability for complex irregular obstacles.

[0061] It should be noted that, in an embodiment, there can be multiple planes that satisfy the perpendicularity condition and the tangency condition at the same time. Since the tangent plane will serve as the starting plane of the blind zone, i.e., the blind zone extends from the tangent plane along the direction from the detection point to the geometric center of the obstacle, and the blind zone will be divided into a priority passage area in subsequent path planning, therefore, if a plane that is too close to the detection point is selected from the multiple tangent planes that satisfy the conditions as the tangent plane, it will result in the generated blind zone space containing part of the obstacle volume at the near end, causing a physical contradiction that the obstacle exists in the priority passage area, thereby interfering with the judgment of the path search algorithm and reducing the feasibility of the planned path. Based on this, the present application provides another exemplary embodiment as follows:

[0062] In an exemplary embodiment, determining the tangent plane of the obstacle comprises: determining all planes that are perpendicular to the line connecting the detection point and the geometric center of the obstacle and tangent to the surface of the obstacle, to form a candidate plane set; and selecting the plane farthest from the detection point from the candidate plane set as the tangent plane to be finally used.

[0063] In actual determination of the candidate plane set, since the tangent planes close to the detection point can be excluded, a first plane passing through the geometric center of the obstacle and perpendicular to the line connecting the detection point and the geometric center of the obstacle can be generated; and the first plane is translated along the front line pointing from the detection point to the geometric center of the obstacle, and the planes tangent to the obstacle during the translation are recorded.

[0064] That is, in another exemplary embodiment, determining the tangent plane of the obstacle comprises: generating a first plane passing through the geometric center of the obstacle and perpendicular to the line connecting the detection point and the geometric center of the obstacle; and translating the first plane along the direction from the detection point to the geometric center of the obstacle, and determining a plane tangent to the obstacle and farthest from the detection point during the translation as the tangent plane.

[0065] S202: determining a first projection of the obstacle on the tangent plane with the detection point as the projection center, and determining a blind zone corresponding to the obstacle based on the boundary of the first projection and the straight line connecting the detection point and the boundary;

[0066] This step is based on the determination of the aforementioned tangent plane to perform blind zone calculation and obtain a well-defined and calculable blind zone space model. This model can accurately describe the area that cannot be directly observed by the detection point due to the obstruction of the obstacle, and provides a calculation basis for the cost allocation and path planning in the following.

[0067] In an exemplary embodiment, determining the blind area corresponding to the obstacle comprises: determining a second projection of the obstacle on the region boundary with the detection point as the projection center; and determining the region surrounded by the side surface, the first projection and the second projection as the blind area of the obstacle. The side surface is a straight line connecting the detection point and the boundary of the first projection, sweeping the curved surface or plane between the first projection and the second projection.

[0068] It should be noted that the shape of the obstacle is not limited in the present application, which can include but is not limited to regular cuboid, cuboid or irregular object, etc. When solving the projection of the obstacle, different projection solving methods can be used for different types of obstacles.

[0069] In an exemplary embodiment, if the obstacle is a convex polyhedron, that is, the line connecting any two points on the surface of the obstacle is completely located inside or on the surface, when observed from any external point, the contour is composed of a part of edges, and the endpoints of these edges are the vertices of the polyhedron. Under perspective projection, the polygon formed by the projection of these contour vertices is the exact occlusion contour of the obstacle on the projection plane. Therefore, only the perspective projection of all vertices of the convex polyhedron is needed, and the convex hull of these projected points is calculated to obtain the first projection or the second projection, without processing the faces or edges.

[0070] In another exemplary embodiment, if the obstacle is a concave polyhedron or the detection point is located near the concave part of the obstacle, the contour of the concave polyhedron can include points other than the vertices. These points are the intersection points of the edges of some "back surface" and the line of sight, which produce new concave turns on the contour line. Only projecting the original vertices will lose these key contour points, resulting in a polygon that cannot reflect the true occlusion shape, which is usually smaller or deviates from the true contour. Therefore, contour extraction can be performed to determine the contour edges based on the geometric data of the obstacle, project these contour edges onto the tangent plane, and connect them to form a projected polygon.

[0071] In addition, if the obstacle is a smooth curved surface body or there is a smooth curved surface body in the obstacle, there can be no vertex for the curved surface body, or the vertices of the discretized grid of the curved surface body are not directly located on the visual contour line. The contour is a smooth curve composed of points where all lines of sight are tangent to the curved surface. Therefore, in one embodiment, a contour line solving algorithm can be used, such as solving the equation set (line of sight direction · curved surface normal vector) = 0; or the smooth curved surface is discretized into a sufficiently fine polygonal grid in advance, which is treated as a polygon, and the contour edges of the grid are used to approximate the smooth contour.

[0072] It should be noted that in the actual application of the agent concealment maneuver method of the present application for blind area calculation, there are usually multiple obstacles and multiple detection points in the target space. For each combination of obstacles and detection points, a corresponding local blind area can be calculated according to the above method. However, the agent needs to achieve comprehensive concealment of all detection threats in the environment in actual maneuver. Therefore, the global blind area used for subsequent path planning is not a simple collection of these local blind areas, but the intersection of the space regions they surround.

[0073] The blind area formed by a single obstacle can only ensure that the agent is concealed with respect to a single detection point blocked by the obstacle; and the intersection region of all blind areas represents the global blind area that can simultaneously avoid the line of sight of all detection points. Only path points located in the intersection region can ensure that the agent is not discovered by any detection point during the maneuver process. Therefore, the intersection processing of multiple independently calculated blind areas can ensure the overall concealment and tactical feasibility of the path planning result in a complex multi-obstacle environment.

[0074] S203: determining a reference path in the target region based on a path cost function, the path cost function including the cost of the path passing through the blind area, the obstacle and the exposure area;

[0075] It should be noted that in an embodiment, after the blind area calculation is completed, the entire digital field can be divided into three types of regions, namely the obstacle region, the exposure region and the blind region. The obstacle region can refer to the space region occupied by a physical entity. This region is not passable in geometry and the agent cannot pass through. The blind region can refer to a space region that cannot be directly detected by one or more detection points due to the blocking of one or more obstacles. The exposure region can refer to a passable region that is not blocked by any obstacle and is exposed to the direct field of view or detection range of one or more detection points. In the present application, it refers to other regions in the target region except the obstacle region and the blind region. When the agent is located in this region, there is a certain risk of being discovered.

[0076] In an exemplary embodiment, the cost functions of the blind area, the obstacle region and the exposure region are determined respectively, and the reference path is determined based on the cost functions, including: generating a navigation grid, the navigation grid including a plurality of nodes and edges connecting the nodes; determining the cost function corresponding to each edge based on the type of the region where the edge is located, the region type including the blind area, the obstacle region and the exposure region; calculating the path with the minimum cumulative cost from the starting point to the ending point as the reference path based on the cost function.

[0077] The generation of the navigation mesh is the process of discretizing the continuous, typed three-dimensional space into a topological network structure suitable for graph search algorithm processing. Each vertex of the navigation mesh is regarded as a node in the graph; the mesh edge connecting two nodes is regarded as an edge in the graph. This edge represents that the agent can move directly between the two positions. Thus, the continuous navigation mesh is converted into an undirected graph or weighted graph composed of nodes and edges.

[0078] In addition, the cost function is used to quantify the mathematical cost that can be directly processed and optimized by the path search algorithm. The core is to assign a weight value to each edge of the navigation mesh, which reflects the "exposure risk cost" of moving along this edge.

[0079] For example, in an embodiment, if the navigation mesh is located in the blind area, the cost function is set to , if it is located in the exposure area, it is set to , and if it is located in the obstacle area, it is set to infinity. Wherein, to indicate the low risk cost of moving in the blind area and the high risk cost of moving in the exposure area.

[0080] It should be noted that in actual agent concealment maneuver combat, the battlefield environment often has highly dynamic and complex characteristics. On the one hand, the number and spatial distribution of enemy detection points are not fixed and unchanging, but may change dynamically from one to multiple as the combat progresses, forces are mobilized, or tactics are needed; on the other hand, in the navigation mesh constructed for agent path planning, adjacent nodes may be located in different types of tactical areas such as blind areas, exposure areas, etc., resulting in that the edge connecting them actually spans different area types, thereby simultaneously existing differentiating concealment attributes and exposure risks on a single path segment.

[0081] Based on the above complex and variable battlefield reality, if a single or simplified cost evaluation model is used, it will be difficult to accurately depict the true tactical cost of the agent moving in different areas, and it will also be difficult to respond sensitively to the integrated threat situation composed of multiple detection points. Therefore, the present application aims to set the cost function, and is specifically implemented by the following multiple embodiments of different types to systematically cope with the above complexity and ensure that the planned path is mathematically optimal while meeting the tactical concealment requirements in a highly hostile environment.

[0082] In an exemplary embodiment, for edges that are completely located in a single area (blind area or exposure area) and are mainly affected by a single detection point, the cost function of the edge can be designed as a continuous function related to the distance between the point and the detection point:

[0083]

[0084]

[0085]

[0086]

[0087] wherein, is the Euclidean distance from point to the nearest probe point; is the distance decay function; is the decay coefficient; is the reference cost coefficient; , are the start and end points of an edge in the navigation mesh, respectively; is the parametric representation of a point on an edge in the navigation mesh; is the linear interpolation parameter; , are the cost of the blind zone and the exposure zone, respectively.

[0088] In another exemplary embodiment, when an edge is split into two segments by a region boundary, for example, the first segment is in the blind zone, and the second segment is in the exposure zone, the total cost can be the length-weighted sum of the costs of the edge in each region segment:

[0089]

[0090] wherein, denotes the cost of edge in the segment of the parameter interval , which can be calculated using the integral form in the previous exemplary embodiment, but the integral interval is changed accordingly, and the value of corresponding to the segment region is used.

[0091] In addition, in another implementation, for each region segment , the distance from the midpoint of the segment to the nearest probe point can be selected as an approximate evaluation of the distance cost of the segment to simplify the calculation, that is, mathematically expressed as:

[0092]

[0093] wherein, is the distance from point to the nearest probe point; is the distance decay function; is the parametric representation of the midpoint of the region segment , and is the cost function of the corresponding region.

[0094] ​In a third exemplary embodiment, for situations where there are multiple detection points in the environment, the edge At a certain point Overall exposure risk value discovered It can be defined as a combination of the exposure risks posed to it by all detection points.

[0095] A valid definition is that, assuming the discovery events at each detection point are independent of each other, then... The probability that a point is not detected by any probe point is:

[0096]

[0097] in, It is a point Detected point The instantaneous probability (or risk) of discovery is defined based on the point. Compared to Location:

[0098]

[0099] At this point, point The total risk cost can be defined as 1- That is, the probability of being detected by at least one probe point. Therefore, for an edge... For any point affected by multiple probe points, its cost density can be defined as:

[0100]

[0101] At this point, the entire edge The cost is the integral of the cost density of all points on it:

[0102]

[0103] In the above formula, , These are the start and end points of an edge in the navigation mesh, respectively; A parameterized representation of any point on an edge in the navigation mesh; These are linear interpolation parameters; It is a point by Detection points The instantaneous probability of discovery; For point To the Detection points The Euclidean distance; Coefficient and ; Any point on the edge affected by multiple probe points a cost density of the edge between the node and the neighbor node; a risk cost scaling factor, a base movement cost constant, a total number of the probe points; a cost function corresponding to the edge of the connection node.

[0104] Further, the algorithm used for planning the reference path based on the cost function can include, but is not limited to, Dijkstra algorithm, best-first search algorithm, etc., and the embodiments of the present application do not limit this.

[0105] In an exemplary embodiment, based on the cost function, the path with the minimum cumulative cost from the start point to the end point is calculated as the reference path, including: in the navigation grid, taking the start point as an expansion node to traverse the neighbor nodes connected thereto, determining a first path cost and a second path cost, wherein the first path cost represents the cost from the start point to the neighbor node via the expansion node; the second path cost represents the estimated cost from the neighbor node to the end point; determining whether to update the optimal path cost of the neighbor node based on the first path cost; if yes, determining the expansion node as the parent node of the neighbor node; based on the first path cost and the second path cost, selecting the expansion node with the minimum sum of the first path cost and the second path cost from the standby topology nodes with parent nodes and not as expansion nodes, until the parent node of the end point is determined; based on the parent node relationship, determining the reference path in the navigation grid.

[0106] wherein, in an embodiment, the estimated cost (second path cost) in the best-first search algorithm such as A-star can be solved in a similar manner as the cost function of the edge between the nodes described above, specifically including:

[0107] The second path cost is:

[0108]

[0109]

[0110]

[0111]

[0112]

[0113] wherein, , are the neighbor node and the end point, respectively; is the parametric representation of any point between the neighbor node and the end point. is a linear interpolation parameter; is a point is a first probe point discovered instantaneous probability; is a point to a first probe point Euclidean distance; is a coefficient and ; is a cost density of an arbitrary point on the edge affected by multiple probe points is a risk cost scaling coefficient, is a base movement cost constant, is a total number of probe points; is the second path cost.

[0114] S204: Construct a simulated force field in the target area with the reference path as an attractive force source and the obstacles as repulsive force sources to generate a stealthy maneuver path of the agent.

[0115] It needs to be specially pointed out that although the reference path realizes the theoretical optimum of stealthiness at the global level, its essence is a fixed route generated based on the static environment model constructed before the task. In the actual dynamic task environment, this path is difficult to effectively respond to unknown obstacles or environmental mutations that appear in real time. If a complete global re-planning is triggered once every time a new obstacle is perceived to update the reference path, high computational resource consumption and system load pressure will be inevitably introduced, which will damage the real-time response performance of the system.

[0116] In addition, a large number of small or temporary obstacles in the environment that do not affect global stealth are not worth triggering a time-consuming global re-planning, but if they are ignored, the agent may directly collide or fail to maneuver.

[0117] To solve the above-mentioned contradictions, the application introduces a simulated force field mechanism. This mechanism converts the static end point and the reference path into attractive force sources with a guiding effect. In an embodiment, the end point can be set as a point-like attractive force source, and the reference path can be set as a line-like attractive force source. At the same time, the obstacles perceived in real time are converted into repulsive force sources with a repulsive effect, thereby constructing a dynamic virtual force field in the environment. The agent generates a second path under the real-time guidance of the force in this force field. The generation of this path is a lightweight, continuous physical simulation process, rather than repeated global search. This enables the agent to be guided within the global optimal path defined by the reference path at all times, while also having real-time, smooth local obstacle avoidance and bypassing capabilities.

[0118] It can be understood that the simulation force field is a virtual physical force field, and the agent is modeled as a force point in the field, and the motion of the agent is determined by the resultant force generated by all force sources in the field. In an embodiment, the end position can be configured as a point-like attractive force source, the attractive force function of which is proportional to the distance or has a constant strength, and the direction always points to the end point; the reference path is configured as a linear attractive force source or a combination of multiple linear attractive force sources; the obstacle is set as a planar repulsive force source to repel the agent and prevent collision with the obstacle. The obstacle repulsive force source can include original obstacles and newly detected obstacles in real time.

[0119] Specifically, in an exemplary embodiment, the generating a second path based on the simulation force field includes: determining a resultant force suffered by the current position in the simulation force field; determining a position at a next time based on the size of the resultant force, the direction of the resultant force, a preset speed and an iteration period length; and updating the position at the next time as the current position until the current position is a preset end point.

[0120] It should be noted that in the simulation force field, if the entire reference path is always regarded as a fixed attractive force source, the agent will continue to generate a backward attractive force component in the path segment that has been passed through during the advancement. This component is in conflict with the attractive force pointing to the end point and the repulsive force for obstacle avoidance, resulting in a deviation of the calculated resultant force direction from the expected optimal path. Therefore, the present application provides an embodiment as follows:

[0121] In this embodiment, before determining the resultant force suffered by the current position each time, the effectiveness of the attractive force source can be determined. Specifically, it includes: determining the nearest point of the current position on the reference path; determining the reference path between the start point and the nearest point as an ineffective attractive force source; and determining the resultant force suffered by the current position in the simulation force field based on the repulsive force source and the remaining effective attractive force source. Based on this, by dynamically determining and disabling the attractive force of the passed path segment, the interference source is eliminated, the force field guidance is always focused on the uncompleted path in front and the target, and the purpose of optimizing the motion quality and control efficiency is achieved.

[0122] In an embodiment, as known from the above, only the distance from the blind area starting face to the detection point, the projection point of the geometric center point of the obstacle on the blind area starting face, and all the vertices of the blind area starting face enclosed by the extension direction of the blind area are needed to determine the blind area.

[0123] The definition of the blind area is given as follows.

[0124]

[0125] In the above formula, represents the distance from the blind area starting face (first projection) to the detection point, This represents the projection point of the obstacle's geometric center onto the starting surface of the blind zone. Indicates the direction of blind spot extension. This represents all the vertices that form the starting face of the blind zone.

[0126] To simplify blind spots The definition of can be agreed upon. By arranging the elements in a fixed winding sequence and using the left-handed spiral rule to determine the positive direction of the normal to the starting surface of the blind zone, the direction of blind zone extension can be determined. can be It can be obtained directly. The formula is as follows.

[0127]

[0128] Simplified blind spot The definition is as follows.

[0129]

[0130] According to blind spots The definition of can be used to solve for the coordinates of each point on the far plane of the blind zone at any distance, as shown in the following formula.

[0131]

[0132] In the formula, This represents the unit vector indicating the direction of blind zone extension. Indicates the location of the detection point. This represents the distance from the far plane of the blind zone to the detection point. The far plane of the blind zone is represented by the first The coordinates of the points.

[0133] After obtaining the blind zone in the reference coordinate system, it is necessary to transform the local coordinates of each point in the blind zone definition to global coordinates to obtain the blind zone in the global coordinate system. The definition of is as follows.

[0134]

[0135] in , The formula is as follows.

[0136]

[0137] In the formula, This is the transformation matrix for converting local coordinates to global coordinates. This is the transformation matrix for converting global coordinates to local coordinates. It's worth noting that all vectors in the equation are column vectors. When a vector interacts with the transformation matrix... When multiplying vectors, they are in homogeneous form. This homogeneous form means that for coordinate vectors, one dimension is added and its value is 1; for direction vectors, one dimension is added and its value is 0.

[0138] After constructing the navigation grid and assigning the cost function, the algorithm is guided to find a globally optimal path by minimizing the cost function. Here, the exposure function is used as the cost function, as follows.

[0139]

[0140] In the formula, Indicates via node The total cost of reaching the destination. Indicates the distance from the starting point to the node. The cumulative cost, Indicates from node The estimated cost of traveling from the starting point to the destination. Indicates from node To the node The cost per step, when A value of 1 indicates the single-step cost from the starting point to node 1. Indicates from node To the node The initial cost per step, when A value of 1 indicates the initial cost of a single step from the starting point to node 1. Indicates from node The estimated initial cost of the blind spot from departure to destination. Indicates from node The initial cost of the estimated exposure zone from departure to destination. Indicates the cost coefficient. Indicates the blind zone cost coefficient. This represents the cost coefficient for the exposure area.

[0141] A simulated force field is then constructed, and the second path is calculated based on the resultant force at each point in the simulated force field. The force field formula is as follows:

[0142]

[0143] In the formula, Indicates the position of an object in a force field ,time The resultant force function, Indicates the first The inherent gravitational gain coefficient of a gravitational source, Indicates the first The inherent repulsive gain coefficient of a repulsive source, Indicates the position of an object in a force field ,time The gravitational function of a certain force source. Indicates the position of an object in a force field ,time The repulsive force function of a certain force source. Represents the set of locations of a certain force source. Indicates the first A set of locations of gravitational sources Indicates the first The set of locations of repulsive force sources This represents a function to determine the effectiveness of the gravitational force exerted by a certain force source on an object. This function represents the determination of the effectiveness of the repulsive force exerted by a force source on an object. Indicates position The distance function to a certain force source, Indicates position The gravitational direction function of a certain force source is given, and the function value is a unit vector. It is worth noting that the summation operation in the formula uses vector addition.

[0144] In practical applications, , This is a hyperparameter, set manually, and the ideal value can be found through multiple experiments based on the actual situation. and The definition of validity and invalidity also needs to be based on the actual situation.

[0145] According to the force field formula, the net force on an object in a force field is also related to the geometry of the force source. Here, force sources are classified according to their geometry into point-like, line-like, and surface-like force sources. Linear force sources include line segment-type, ray-type, and straight-line-type force sources, while surface-like force sources include local-type, half-plane-type, and planar-type force sources. For more complex geometries, they can be composed of combinations of basic point-like, line-like, and surface-like force sources. For example, a polygonal force source can be composed of multiple point-like and line-like force sources, and a cubic force source can be composed of multiple point-like, line-like, and surface-like force sources.

[0146] The following are the distance functions corresponding to force sources with different geometries given in this specific embodiment. and gravitational direction function Definition.

[0147] The distance function and gravitational direction function of a point force source are defined as follows.

[0148]

[0149] In the formula, Indicates position To location set The projection, represents zero vector. Since it is a point force source, its position set contains only one point, i.e. the projection point .

[0150] The distance function and the direction function of the line force source are defined as follows.

[0151]

[0152] where, represents the vertical projection of position to the straight line on which the position set is located, represents zero vector. For the linear force source, there is always , for the ray type and line segment type force source, there are both and cases.

[0153] The distance function and the direction function of the surface force source are defined as follows.

[0154]

[0155] where, represents the vertical projection of position to the plane on which the position set is located, represents zero vector. For the plane type force source, there is always , for the half plane type and local type force source, there are both and cases.

[0156] In the process of maneuvering, the agent calculates the maneuvering based on the force field. In a maneuvering iteration cycle, first, the direction of the resultant force suffered by the agent is calculated according to the final position of the agent in the last iteration cycle, i.e. the maneuvering direction of the agent, then the displacement of the agent is calculated combined with the speed of the agent and the length of the iteration cycle, and finally the final position of the agent is calculated. Then gradually iterate down until the end point is reached.

[0157] The position iteration formula can be as follows.

[0158]

[0159] where, represents the final position of the iteration cycle, represents the final position of the iteration cycle, represents the speed of the agent, represents the length of the iteration cycle, represents the agent in the first the resultant force received by the iteration cycle.

[0160] Please refer to Figure 3 As shown in the figure, based on the same inventive concept as the aforementioned agent concealment maneuvering method, an embodiment of the present application provides an agent concealment maneuvering system 300, which comprises a first blind area solving module 301, a second blind area solving module 302, a path planning module 303, and a maneuvering implementation module 304.

[0161] It should be noted that the agent concealment maneuvering method of the embodiment of the present application can be applied to the agent concealment maneuvering system 300 of the embodiment of the present application. The agent concealment maneuvering system 300 can be configured in a terminal. The terminal can include, but is not limited to, a PC (Personal Computer), a PDA (tablet computer), a smart phone, a smart wearable device, and the like.

[0162] Specifically, the first blind area solving module 301 is configured to determine a tangent plane of an obstacle, the tangent plane being perpendicular to a line connecting a geometric center of the obstacle and a detection point; the second blind area solving module 302 is configured to determine a first projection of the obstacle on the tangent plane with the detection point as a projection center, and determine a blind area corresponding to the obstacle based on a boundary of the first projection and a straight line connecting the detection point and the boundary; the path planning module 303 is configured to determine a cost function of a blind area, an obstacle area, and an exposure area respectively, and determine a reference path based on the cost function; and the maneuvering implementation module 304 is configured to determine the reference path as a gravitational source, determine the obstacle as a repulsive source, construct a simulated force field, and generate a second path based on the simulated force field.

[0163] Please refer to Figure 4 As shown in the figure, the embodiment of the present application further provides an electronic device 400, which comprises at least one processor 401, a memory 402 (for example, a non-volatile memory), an internal memory 403, and a communication interface 404, and the at least one processor 401, the memory 402, the internal memory 403, and the communication interface 404 are connected together via an internal bus 405. The at least one processor 401 is configured to invoke at least one program instruction stored or encoded in the memory 402, so as to enable the at least one processor 401 to perform various operations and functions of the agent concealment maneuvering method described in various embodiments of the present specification.

[0164] In embodiments of the present specification, the electronic device 400 can include, but not limited to, a personal computer, a server computer, a workstation, a desktop computer, a laptop computer, a notebook computer, a mobile electronic device, a smart phone, a tablet computer, a cellular phone, a personal digital assistant (PDA), a handheld device, a messaging device, a wearable electronic device, a consumer electronic device, and the like.

[0165] The embodiments of the present application also provide a computer readable medium, which carries computer execution instructions, and the computer execution instructions can be used to implement various operations and functions of the agent concealment maneuver method described in the embodiments of the present specification when executed by a processor.

[0166] The computer readable medium in the present application can be a computer readable signal medium or a computer readable storage medium or any combination of the two. The computer readable storage medium may, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared or semiconductor system, device or apparatus, or any combination of the above. More specific examples of the computer readable storage medium can include, but not limited to, an electrical connection with one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, the computer readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device or apparatus.

[0167] In the present application, the computer readable signal medium can include a data signal propagating in a baseband or as a carrier wave in a propagated data signal, which carries computer readable program code. Such a propagated data signal can take many forms, including but not limited to, an electromagnetic signal, an optical signal or any suitable combination thereof. The computer readable signal medium can also be any computer readable medium that can send, propagate or transmit the program for use by or in connection with an instruction execution system, device or apparatus. The program code contained on the computer readable medium can be transmitted by any suitable medium, including but not limited to wireless, wire, optical cable, RF or the like, or any suitable combination thereof.

[0168] Those skilled in the art will appreciate that embodiments of the present application can be devised for a variety of applications. It is intended that the present application covers all such applications of the embodiments disclosed herein, whether or not the specific data is disclosed. It is intended to embrace all alternatives, modifications and sub-combinations of the embodiments disclosed herein and to cover all changes, alterations and modifications to the embodiments disclosed herein that are within the spirit and scope of the present application. It is intended to cover all such changes, alterations and modifications as fall within the scope of the appended claims and their equivalents. It is intended that changes in, or additions of, parts and / or sub-combinations thereof can be made to the embodiments disclosed herein and still be encompassed by the present application. It is intended that the present application covers all such applications of the embodiments disclosed herein, whether or not the specific data is disclosed. It is intended to embrace all alternatives, modifications and sub-combinations of the embodiments disclosed herein and to cover all changes, alterations and modifications to the embodiments disclosed herein that are within the spirit and scope of the present application. It is intended to cover all such changes, alterations and modifications as fall within the scope of the appended claims and their equivalents. It is intended that changes in, or additions of, parts and / or sub-combinations thereof can be made to the embodiments disclosed herein and still be encompassed by the present application.

[0169] The present application is described herein with reference to the drawings, in which are shown embodiments in accordance with the present application. Aspects of the present application are illustrated by way of example, and not limitation, in the accompanying drawings, in which like references indicate similar elements, and in which: Figure 1 a method or a flow Figure 1 a means for carrying out any one of the functions described in a method or a flow

[0170] The foregoing description of specific exemplary embodiments of the application has been presented for the purposes of illustration and description. It is not intended to be a limitation on the broad concepts of the application. Obviously, many modifications and variations of the present application are possible in light of the above teachings. It is intended that the scope of the application encompass all such modifications and variations as long as they are within the spirit and scope of the application. It is intended that the application encompass all changes and modifications of the specific exemplary embodiments which fall within the scope of the appended claims along with their full scope of equivalents.

[0171] It will be apparent to those skilled in the art that the present application is not limited to the specific embodiments described herein and that many modifications, additions, substitutions and changes can be made to the specific embodiments without departing from the spirit and scope of the present application. Accordingly, the scope of the present application should not be limited by the description of the specific embodiments, but should be defined by the appended claims and their equivalents.

[0172] Furthermore, it should be understood that although the specification is described in terms of embodiments, not every embodiment includes every feature described. The specification can include implicit combinations of explicitly mentioned features and / or explicit combinations of implicitely mentioned features. Each embodiment depends on the explicit combinations of features and / or the implicit combinations of features made specifically within that embodiment, and each such embodiment can be combined with every other such embodiment to create further embodiments.

Claims

1. An agent cloaking maneuver method, comprising: The method comprises: determining a tangent plane of the obstacle in the target region, the tangent plane being perpendicular to a line connecting a geometric center of the obstacle and a detection point; determining a first projection of the obstacle on the tangent plane with the detection point as a projection center, and determining a blind area corresponding to the obstacle based on a boundary of the first projection and a straight line connecting the detection point and the boundary; determining a reference path in the target region based on a path cost function, the path cost function including costs of the path passing through the blind area, the obstacle and an exposure area, the exposure area being an area in the target region excluding the blind area and the obstacle; constructing a simulated force field in the target region to generate a stealthy maneuver path of an agent with the reference path as an attractive force source and the obstacle as a repulsive force source; wherein the determining of the reference path in the target region based on the path cost function comprises: generating a navigation grid in the target region, the navigation grid including a plurality of nodes and edges connecting the nodes; determining a cost corresponding to each edge based on a type of area where the edge is located, the type of area including the blind area, the obstacle area and the exposure area; and determining the reference path with a minimum cumulative cost from a start point to an end point based on the costs corresponding to the edges. The cost corresponding to the edge connecting the nodes is: wherein, is a point is the Euclidean distance to the point of detection; is a distance decay function; is a decay coefficient; is a reference cost coefficient; , are the start and end points of an edge in the navigation mesh, respectively; is a parametric representation of a point on an edge in the navigation mesh; is a linear interpolation parameter; , are the cost of the blind zone and the exposure zone, respectively; The cost function corresponding to the edge connecting the nodes is: wherein, , are the start and end points of an edge in the navigation mesh, respectively; is a parametric representation of a point on the edge in the navigation mesh; is a linear interpolation parameter; is a point is the instantaneous probability of being discovered by the th probe point ; is the Euclidean distance from the point to the th probe point ; are coefficients and ; is the cost density of an arbitrary point on the edge affected by multiple probe points; is a risk cost scaling coefficient, is a base movement cost constant, is the total number of probe points; is the cost function corresponding to the edge of the connection node.

2. The method of claim 1, wherein, The constructing of the simulated force field in the target region to generate the stealthy maneuver path comprises: determining a resultant force acting on the current position of the agent in the simulated force field; determining a position of the agent at a next time based on a magnitude of the resultant force, a direction of the resultant force, a preset speed and an iteration period length; updating the position at the next time as the current position until the current position is a preset end point.

3. The method of claim 1, wherein, The method further comprises: determining a nearest point of the current position of the agent on the reference path; determining that an attractive force of an interval of the reference path from the start point to the nearest point is invalid; determining the resultant force acting on the current position in the simulated force field based on the repulsive force source and the remaining valid attractive force sources.

4. The method of claim 1, wherein, The determining of the tangent plane of the obstacle in the target region comprises: determining a plane being perpendicular to a line connecting a detection point and a geometric center of an obstacle and tangent to the obstacle, to form a candidate plane set; determining a plane farthest from the detection point in the candidate plane set as the tangent plane of the obstacle.

5. The method of claim 1, wherein, The determining of the reference path with a minimum cumulative cost from a start point to an end point based on the costs corresponding to the edges comprises: in the navigation grid, taking the start point as an expansion node to traverse neighbor nodes connected to the expansion node, to determine a first path cost and a second path cost, wherein the first path cost represents a cost of reaching the neighbor nodes from the start point via the expansion node, and the second path cost represents an expected cost of reaching the end point from the neighbor nodes; determining whether to update an optimal path cost of the neighbor nodes based on the first path cost; if yes, determining the expansion node as a parent node of the neighbor nodes. Based on the first path cost and the second path cost, in the backup topology nodes with parent nodes and not as expansion nodes, the expansion node with the minimum sum of the first path cost and the second path cost is selected for traversal until the terminal point determines its parent node; based on the parent node relationship, the reference path in the navigation grid is determined.

6. The method of claim 5, wherein, The second path cost is: wherein, , are the neighbor nodes and the end point, respectively; is a parametric representation of any point between the neighbor nodes and the end point; is a linear interpolation parameter; is a point discovered by the first probe point ; is the Euclidean distance from the point to the first probe point ; is a coefficient and ; is the cost density of any point on the edge affected by multiple probe points; is a risk cost scaling coefficient, is a base movement cost constant, is the total number of probe points; is the second path cost.

7. The method of claim 1, wherein, The blind area corresponding to the obstacle is determined, including: The second projection of the obstacle on the region boundary is determined with the detection point as the projection center; The region surrounded by the side surface, the first projection and the second projection is determined as the blind area of the obstacle; The side surface is a curved surface or a plane formed by the straight line connecting the detection point and the boundary of the first projection sweeping between the first projection and the second projection.

8. An agent cloaking maneuvering system applying the agent cloaking maneuvering method according to any one of claims 1-7, characterized in that, Including: The first blind area solving module is configured to determine the tangent plane of the obstacle in the target region, and the tangent plane is perpendicular to the line connecting the geometric center of the obstacle and the detection point; The second blind area solving module is configured to determine the first projection of the obstacle on the tangent plane with the detection point as the projection center, and determine the blind area corresponding to the obstacle based on the boundary of the first projection and the straight line connecting the detection point and the boundary; The path planning module is configured to determine a reference path in the target region based on a path cost function, the path cost function including the cost of the path passing through the blind area, the obstacle and an exposure area, and the exposure area is an area in the target region excluding the blind area and the obstacle; The maneuvering implementation module is configured to construct an analog force field in the target region to generate a concealed maneuvering path of an agent with the reference path as an attractive force source and the obstacle as a repulsive force source.

9. An electronic device, comprising: Including: At least one processor; And A memory storing instructions, when the instructions are executed by the at least one processor, the at least one processor is caused to execute the agent concealment maneuvering method as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions for causing a computer to execute the agent concealment maneuvering method of any one of claims 1-7. The computer readable storage medium stores computer instructions for causing a computer to execute the agent concealment maneuvering method of any one of claims 1-7.

Images