Helicopter low-altitude flight path planning method based on motion state search and enhanced safety pipeline

Abstract

The application discloses a helicopter low-altitude flight path planning method based on motion state search and enhanced safety pipeline, which comprises the following steps: firstly, based on a grid map, segmented flight path planning is carried out to generate a helicopter motion state space considering dynamic constraints and maneuvering characteristics; then, after the state space is expanded and a cost function is established, an optimal path search is carried out in the expanded state space by using an improved hybrid A* algorithm; then, in a local map which is updated in real time according to dynamic threats, an enhanced safety pipeline is generated by expanding, pruning and interpolating based on the initial flight path according to the safety constraints of the flight path; finally, the B-spline curve is used to fit the flight path in the enhanced safety pipeline, and the curve parameters are optimized by using the minimum cost criterion, so that the helicopter low-altitude flight path planning optimization is realized. According to the application, the state space considering the dynamic constraints and maneuvering characteristics of the helicopter is used to generate the enhanced safety pipeline in real time and to serve as the optimization space of the fitting parameters of the rear-end flight path, so that the adaptability of the planned flight path to the execution process of the helicopter is improved, and the flight safety of the helicopter under the dynamic threats at low altitude is effectively improved.

Description

Technical Field

[0001] This invention relates to a technology in the field of flight path planning, specifically a method for planning low-altitude flight paths for helicopters based on motion state search and enhanced safety pipelines. Background Technology

[0002] Flight path planning is a core technology for guiding helicopters at low altitudes. It provides pilots with a safe and convenient reference path, assisting in the efficient execution of flight missions. This technology aims to autonomously plan an optimal path from the starting point to the target point, while taking into account constraints such as obstacle avoidance safety, energy consumption, the maneuverability of the planned target, and dynamic constraints.

[0003] One of the challenges in low-altitude helicopter trajectory planning is the weak maneuverability of the planned trajectory. Due to the highly discrete search space of the initial trajectory and insufficient consideration of helicopter dynamics constraints and maneuverability, the planned trajectory is prone to poor smoothness and high maneuver costs, resulting in weak executability. Furthermore, when helicopters perform low-altitude penetration, they often face new battlefield threats such as enemy radar reconnaissance and air defense fire, easily leading to weak dynamic safety of the planned trajectory. Summary of the Invention

[0004] This invention addresses the shortcomings of existing technologies, such as insufficient accuracy in motion modeling of the planned object, spatial overlap and nesting phenomena, and inability to be applied to complex terrain environments. It proposes a helicopter low-altitude trajectory planning method based on motion state search and enhanced safety pipeline. By considering the state space of helicopter dynamic constraints and maneuver characteristics, the method generates an enhanced safety pipeline in real time and uses it as the optimization space for back-end trajectory fitting parameters. This significantly improves the adaptability of the planned trajectory to the helicopter execution process and effectively enhances the flight safety of helicopters under low-altitude dynamic threats.

[0005] This invention is achieved through the following technical solution:

[0006] This invention relates to a helicopter low-altitude trajectory planning method based on motion state search and enhanced safety pipeline. First, segmented trajectory planning is performed based on a grid map to generate a helicopter motion state space considering dynamic constraints and maneuver characteristics. Then, by expanding the state space and establishing a cost function, an optimal path search is performed in the expanded state space using an improved hybrid A* algorithm. Next, in a local map with real-time updates of dynamic threats, an enhanced safety pipeline is generated based on the initial trajectory through expansion, pruning, and interpolation based on trajectory safety constraints. Finally, within the enhanced safety pipeline, the trajectory is fitted using a B-spline curve, and the curve parameters are optimized using a minimum cost criterion to achieve helicopter low-altitude trajectory planning optimization.

[0007] This invention relates to a system for implementing the above-mentioned method, comprising: a global map and task point input module, a front-end trajectory search module, an enhanced safety pipeline module, and a back-end trajectory optimization module, wherein: the global map and task point input module performs smoothing and interpolation processing on the original terrain data to generate a task space for trajectory planning; the front-end trajectory search module expands the task space for trajectory planning by establishing a kinematic model that considers the differences in three-axis control maneuvers of a rotorcraft, and searches for an initial trajectory using an improved hybrid A* algorithm that considers safety margins; the enhanced safety pipeline module expands the original safety pipeline centered on the track points of the initial trajectory, and obtains an enhanced safety pipeline with spatial continuity and concise representation through pruning and interpolation; the back-end trajectory optimization module applies positional constraints to the trajectory fitting control points according to the enhanced safety pipeline and generates a back-end optimized trajectory for execution with the helicopter, so as to ensure that the optimized trajectory is always in the safety space while taking into account the smoothness criteria of trajectory and maneuver.

[0008] Technical effect

[0009] In the state space expansion stage of the front-end trajectory search, this invention applies scale transformation and kinematic constraints to the three-axis control input of the planning object; based on the safety pipeline expansion, it introduces a detection and improvement mechanism for sub-pipeline redundancy merging and discontinuous interpolation; compared with the existing A* algorithm, this invention significantly shortens the trajectory length, improves the planning success rate, further improves the success rate of trajectory planning in complex scenarios, and reduces the trajectory execution cost. Attached Figure Description

[0010] Figure 1 This is an overall flowchart of the present invention;

[0011] Figure 2 Here is a flowchart of the Hybrid State A* algorithm for an example;

[0012] Figure 3 This is a flowchart illustrating the enhanced safety pipeline generation process used in this embodiment;

[0013] In the diagram: (a) is the safety sub-pipe expansion flowchart, and (b) is the safety sub-pipe trimming flowchart;

[0014] Figure 4 This is a comparison chart of the implementation effect of the present invention with that of traditional trajectory planning methods;

[0015] In the figure: (a) is the traditional method using A* as the search front end, and (b) is the proposed method;

[0016] Figure 5 A schematic diagram illustrating the kinematic constraints for planning flight paths;

[0017] In the figure: (a) and (b) are the velocity and acceleration curves output from the simulation experiment, respectively;

[0018] Figure 6 A schematic diagram illustrating the effect of flight path planning on the maneuverability of helicopters;

[0019] In the figure: (a) and (b) are the triaxial motion characteristic curves under different scale coefficients, respectively;

[0020] Figure 7 This is a schematic diagram illustrating the effect of enhancing the safety of the pipeline in an embodiment. Detailed Implementation

[0021] like Figure 1 As shown, this embodiment relates to a helicopter low-altitude trajectory planning method based on motion state search and enhanced safety ducts, including:

[0022] Step 1) Global Map and Task Point Input: After interpolation and filtering, the digital elevation map data is converted into a raster global map, and the user inputs the task track points. Specifically, this includes: original terrain smoothing, raster map conversion, and task point sequence input.

[0023] The original terrain smoothing is achieved by completing the terrain data using a cubic spline curve interpolation algorithm, followed by noise reduction of the terrain data using a two-dimensional Gaussian filter.

[0024] The raster map has a size of 40×40m and a raster resolution of 0.1m.

[0025] Step 2) Front-end trajectory search: Generate a helicopter motion state space that considers dynamic constraints and maneuver characteristics in the global map, and after expanding the state space and establishing a cost function, use an improved hybrid A* algorithm to search for the optimal path in the expanded state space.

[0026] The helicopter motion state space refers to the helicopter state transition equation with acceleration as the three-axis control input. Where: State transition matrix Control Matrix I3 is a 3x3 identity matrix, and 0 is a zero matrix filled in the corresponding dimension; the state node vector of the helicopter. Triaxial acceleration Corresponding three-axis control input Helicopter's initial state vector The rules for expanding the initial track are as follows:

[0027] The extended state space refers to: discretizing the three-axis control inputs into intervals, scaling them, expanding the trajectory into a tree structure in discrete form, and using the track points on this tree as the helicopter's state space, where: the feasible region u of the three-axis unit control input u0... d =[-u max u max The feasible region is uniformly divided using a scale n to generate 2n+1 discrete control input quantities, i.e. Constraints on acceleration a max By controlling the feasible boundary u of the input quantity max Apply, while the velocity constraint v max The acceleration constraint and the duration of the segmented trajectory expansion process are applied together. The three-axis unit control input u0(t) is scaled to approximate the difference in the helicopter's three-axis maneuver response to the unit control input, resulting in the three-axis control input u(t) = [u x (t), u y (t), u z (t)] T u(t) = Λ·u0(t), where: the transformation coefficient matrix Λ = diag(λ) x , λ y , λ z ), where T is the sampling period.

[0028] In this embodiment, n = 2.

[0029] In this embodiment, the scaling parameter of the control group is set to λ. x =1,λ y =1,λ z =1, the experimental group is θ x =1,λ y =1,λ z =0.5.

[0030] In this embodiment, the sampling period T = 0.001s.

[0031] The aforementioned cost function is the comprehensive cost function value f(n) = g(n) + h(n) for each node n in the expanded state space, where g(n) is the measure of the distance of the current state node from the starting point n. start The maneuver cost, h(n), is the distance n from the current state node to the target point. goal The cost of inspiration.

[0032] like Figure 2 As shown, the optimal path search includes:

[0033] 2.1) Determine the position and velocity of the starting point node_start and the target point node_goal as node state values, add the starting node to the list of nodes to be examined OpenList, and calculate the comprehensive cost function value f(n) of the starting point;

[0034] The OpenList of nodes to be examined includes a starting point node_start and a target point node_goal during the initialization phase. During the running phase, neighboring nodes generated by the expansion of the currently searched node are continuously added.

[0035] 2.2) When the list of nodes to be examined, OpenList, is not an empty set, select the node node_current corresponding to the smallest comprehensive cost function value from it;

[0036] 2.3) When the combined cost function values ​​of the starting point node_start and the target point node_goal are close, the target is considered to have been reached and the planning ends; otherwise, proceed to step 2.4.

[0037] 2.4) Starting from the current node_current, perform motion micro-element expansion and redundancy reduction on the expanded nodes so that each grid retains at most one node. The set of adjacent nodes generated by the previous node_current and the set of adjacent nodes after redundancy reduction are stored in the set of adjacent nodes_successor, which is expanded by the current traversed node_current according to the motion micro-element method.

[0038] 2.5) Traverse each node in the set of adjacent nodes node_successor and execute step 2.6 for each node;

[0039] 2.6) For nodes that are not yet included in CloseList and can pass collision detection, calculate the node transfer cost from the current extended node node to node_current, and add g(node_current) as the temporary cost g of node. temp ;

[0040] The collision detection refers to calculating the distance between the detected node and the boundary of the nearest obstacle in the map. If the distance is not positive, it means that the detected node has collided with the obstacle.

[0041] The node transfer cost refers to the control and time cost required for the transfer between nodes, which is characterized by the magnitude of the control input and the length of the movement.

[0042] The temporary cost g temp The formula is: g temp= g(node_current) + (||u(t)|| 2 +ρ)·τ, where u(t) represents the corresponding three-axis control input during the movement from node to node_current, τ represents the movement time, and ρ is the trade-off coefficient between control input and time cost.

[0043] 2.7) If the expanded node is not yet included in the OpenList, add it; otherwise, compare g(node) with the temporary cost g. temp The size of the node is determined, and g(node) is updated to the smaller of the two values.

[0044] 2.8) Calculate and update the total cost value f(node) of the node, and set node_current as the parent node of node.

[0045] 2.9) After the target node in the OpenList is visited, the final returned CloseList is obtained, which contains the point sequence in the optimal track.

[0046] Step 3) Enhanced security pipeline: In a local map where dynamic threats are updated in real time, an enhanced security pipeline is generated based on the tubular space that encloses the initial trajectory through expansion, trimming and interpolation based on trajectory security constraints.

[0047] The enhanced safety pipeline refers to a tubular space that encloses the initial trajectory with a set of cuboid-shaped sub-pipes, representing the safety range during the fitting and optimization of the subsequent trajectory.

[0048] like Figure 3 As shown in (a), the expansion specifically includes:

[0049] 3.1) The sub-pipes are represented by the six boundary faces of a cuboid. Initialize to discrete point P on the initial track i The corresponding x, y, z coordinates P ix P ix P iy P iy P iz P iz At the same time, initialize the collision detection variable Safe←True;

[0050] 3.2) Using the maximum number of expansions n as the upper bound of the iteration, when the expansion boundary and the obstacle point set O pass the collision detection, i.e., Safe = True, the sub-pipe boundary continues to expand along the six axes of the spatial coordinate system with a step size of ε, i.e., C μ ←C μ +ε,μ∈{x + x -y + y - , z + , z -};

[0051] 3.3) When the number of iterations reaches the upper limit or the boundary of the cuboid collides with an obstacle, the expansion of the sub-pipe terminates. The sub-pipe described by the six boundaries of the cuboid is denoted as C. i Then, the set of sub-pipes expanded by the algorithm for all initial waypoints is C = {C i}, i∈{1,2,...,N}.

[0052] The collision detection refers to calculating the distance between the boundary of the enhanced safety pipe and the nearest obstacle in the scene. If the distance is not positive, it means that the enhanced safety pipe has collided with the obstacle.

[0053] like Figure 3 As shown in (b), the trimming and interpolation specifically include:

[0054] 3.a) The set of sub-pipes obtained after expansion by traversing using a double loop: the set of sub-pipes currently being traversed in the outer loop {C i The sub-pipes in the} are checked for inclusion relationships. When an inclusion relationship occurs... In the case where i≠j, then from the initial enhanced safety pipeline set {C i Remove redundant sub-pipes C that are iterated within the inner loop. j ;

[0055] 3.b) For the traversed enhanced safety pipeline set {C} i The sub-pipes in} are subjected to continuity detection. When an adjacent sub-pipe C appears... i C j They do not intersect, that is In the case of P, then at its corresponding adjacent track point i P j Insert new waypoint P between ij And re-expand and generate new sub-pipes C ij ;

[0056] 3.c) When all sub-pipes in the initial enhanced safety pipe have been traversed and all sub-pipes have passed the redundancy inclusion and spatial continuity checks, the pruned and interpolated enhanced safety pipe is obtained.

[0057] The inclusion relationship detection refers to: traversing the expanded set of sub-pipes in a double-loop manner, comparing each sub-pipe pairwise, and detecting whether the space enclosed by the cuboid has an inclusion relationship.

[0058] The continuity detection refers to: after traversing the expanded set of sub-pipes in a double-loop manner, detecting whether there is a non-empty spatial intersection between the cuboids occupied by adjacent sub-pipes.

[0059] The aforementioned pruning and interpolation refer to the following operations: after traversing the expanded sub-pipe set in a double-loop manner, redundant sub-pipes are removed and discontinuous spaces are filled in through inclusion relationship detection and continuity detection, respectively.

[0060] Step 4) Back-end trajectory optimization: Within the enhanced safety pipeline obtained in Step 3, the trajectory is fitted using a B-spline curve, and the curve parameters are optimized using the minimum cost criterion to achieve helicopter low-altitude trajectory planning optimization.

[0061] The minimum cost criterion refers to applying inequality constraints to the B-spline control point parameters of the enhanced safety pipeline to ensure that the optimized trajectory in the back end lies within the space defined by the enhanced safety pipeline. Specifically, in the back end trajectory optimization process, the B-spline curve uses N+1 control points to fit M segments of trajectory, and the i-th control point in the μ-axis direction... satisfy: Where: i∈{0,1,...,N}, j∈{0,1,...,M-1}, and corresponds to the j-th segment track, μ∈{x,y,z}, and The sub-pipes C shown in the cuboid diagram are respectively j The positions of the two boundary surfaces in the μ-axis direction.

[0062] Based on specific practical experiments, the processor was an Intel(R) Core(TM) i5-7200U CPU@2.50GHz, and the memory was 16GB. The simulation environment was Ubuntu 18.04, using the ROS Melodic robot simulation framework and the Rviz data visualization tool. Note: To improve simulation efficiency, the physical scale of actual flight was scaled down by 10:1 (including velocity, acceleration, etc.).

[0063] In the embodiment using the A* algorithm as the front end, due to the high dispersion of the trajectory search space and the lack of consideration for the motion characteristics of the maneuvering object, the smoothness of the planned trajectory is not high and the maneuvering cost is large. The experimental results are as follows: Figure 4 As shown in (a). The front-end of this invention is based on an improved hybrid A* algorithm, whose search space is a more continuous state space generated by the basic helicopter dynamics model. Therefore, it can effectively improve the safety and convenience of flight path planning. In 20 and 50 simulations with random terrain, the planning success rate was improved by 10% and 8%, respectively. The experimental results are as follows. Figure 4 (b) and Table 1 are shown.

[0064] Table 1: Comparison Results of Improved Hybrid A* and A* Front-End Search

[0065]

[0066] This invention considers the helicopter's maneuverability during the front-end trajectory search process and can flexibly adjust the relative magnitude of the scaling transformation parameters according to the differences in the three-axis coupled control characteristics of different helicopters, thereby enhancing the adaptability of the planned trajectory to the control characteristics of the maneuvering object. In the embodiments, the parameter λ is set as the control group. x =1,λ y =1,λ z =1, the experimental group is λ x =1,λ y =1,λ z =0.5, and its corresponding triaxial motion characteristic curve is as follows Figure 5 As shown in (a) and 5(b), the numerical results are shown in Table 2.

[0067] Table 2: Effects of Triaxial Acceleration and Velocity Constraints

[0068]

[0069] Furthermore, this invention incorporates helicopter dynamic constraints during the front-end trajectory search process, which can be flexibly adjusted according to the actual situation of the maneuvering object, thereby making the planned trajectory highly adaptable to the dynamic characteristics of the maneuvering object. In the embodiments, v is applied to the helicopter in the simulation state respectively. max =2m / s and a max =3m / s 2 Constraints. From Figure 6 As shown in the velocity and acceleration simulation curves in (a) and 6(b), the motion state of the overall planned trajectory is effectively limited within the threshold range.

[0070] Finally, in the embodiment based on motion state search and enhanced safety pipeline, the expansion step size ε of the enhanced safety pipeline is set to 1m, and the maximum number of expansions n=4. The simulation results are as follows. Figure 7 As shown in Table 3, in 20 and 50 randomly generated simulation environments, the enhanced safety pipeline significantly improves the success rate of trajectory planning due to the additional consideration of safety constraints in dynamic environments.

[0071] Table 3: Comparative Experiment Results of Enhanced Safety Pipelines

[0072]

[0073] Compared with existing technologies, this method, based on an improved hybrid A* algorithm, applies a scaling transformation to the helicopter's three-axis control inputs during the state space expansion stage of the front-end trajectory search. This modeling of maneuver characteristics and imposing maximum maneuver limits achieves adaptability of the planned trajectory to the maneuver characteristics and constraints of the planned object, improving trajectory planning safety and reducing flight costs. Furthermore, the proposed enhanced safety pipeline's positional constraints on control points during the back-end trajectory fitting process effectively enhances the safety of trajectory planning in complex dynamic environments.

[0074] The above-described specific implementations can be partially adjusted by those skilled in the art in different ways without departing from the principles and purpose of the present invention. The scope of protection of the present invention is defined by the claims and is not limited to the above-described specific implementations. All implementation schemes within the scope of the claims are bound by the present invention.

Claims

1. A method for low-altitude flight path planning of a helicopter based on motion state search and enhanced safety tube, characterized in that, Firstly, the helicopter motion state space considering the dynamic constraints and maneuver characteristics is generated based on the segmented path planning of the grid map, and then the optimal path search is performed in the extended state space by using the improved hybrid A* algorithm after the state space is extended and the cost function is established; secondly, the enhanced safety pipeline is generated by expanding, pruning and interpolating based on the safety constraints of the path in the local map which is updated in real time by the dynamic threats, based on the initial path; Finally, the path is fitted by using the B-spline curve in the enhanced safety pipeline, and the curve parameters are optimized by the minimum cost criterion to realize the optimization of the low-altitude path planning of the helicopter, which specifically includes: Step 1) global map and task point input: after the digital elevation map data is interpolated and filtered, it is converted into a grid global map, and the task path points are input by the user, which specifically includes: original terrain smoothing, grid map conversion and task point sequence input; Step 2) front-end path search: the helicopter motion state space considering the dynamic constraints and maneuver characteristics is generated in the global map, and then the optimal path search is performed in the extended state space by using the improved hybrid A* algorithm after the state space is extended and the cost function is established; Step 3) enhanced safety pipeline: the enhanced safety pipeline is generated by expanding, pruning and interpolating based on the safety constraints of the path in the local map which is updated in real time by the dynamic threats, based on the tubular space enveloping the initial path; The enhanced safety pipeline refers to the tubular space enveloping the initial path by the collection of cuboid-shaped sub-pipelines, which represents the safety range during the fitting and optimization of the rear-end path; Step 4) rear-end path optimization: the path is fitted by using the B-spline curve in the enhanced safety pipeline obtained in step 3, and the curve parameters are optimized by the minimum cost criterion to realize the optimization of the low-altitude path planning of the helicopter.

2. The method according to claim 1, wherein, The helicopter motion state space is a state transition equation of the helicopter with acceleration as three-axis control input , wherein a state transition matrix , a control matrix , is a 3-order unit matrix, and 0 is a zero matrix filled in corresponding dimensions; a state node vector of the helicopter , three-axis acceleration corresponding three-axis control input , a starting state vector of the helicopter , and an extension rule of the initial flight path is ; The extended state space is that the three-axis control input quantity is interval discretized and scaled, the tree trajectory is extended in discrete form, and the track points thereon are taken as the helicopter state space, wherein the three-axis unit control input quantity The feasible region of ] is uniformly divided by the scale n to generate 2n+1 discrete control input quantities, that is The constraint on acceleration is applied through the feasible boundary of the control input quantity , and the velocity constraint is applied by the acceleration constraint and the time length of the piecewise trajectory extension process, so that the three-axis unit control input quantity is scaled to approximate the difference of the unit control input quantity to the three-axis helicopter maneuvering response, and the three-axis control input quantity is , , wherein the transformation coefficient matrix , is the sampling period.

3. The helicopter low-altitude trajectory planning method based on motion state search and enhanced safety pipeline as described in claim 1, characterized in that, The optimal path search includes: 2.1) Determine the position and velocity of the start node node_start and the goal node node_goal as node state values, add the start node to the list of nodes to be examined OpenList, and calculate the cost function value of the start node , the cost function value of each node n in the expanded state space where: is the cost of maneuvering the current state node from the start node , and is the heuristic cost of the current state node from the goal node . 2.2) when the list of nodes to be investigated OpenList is not empty, the node corresponding to the minimum comprehensive cost function value is selected from it; 2.3) when the comprehensive cost function values of the starting point node_start and the target point node_goal are close, it is considered that the target is reached, and the planning is ended, otherwise step 2.4 is specified; 2.4) the motion infinitesimal is extended from the current node node_current as the starting point, and the redundant nodes are removed, so that at most one node is retained in each grid, and the adjacent node set generated by the extension of the previous node node_current and the redundant removal is stored in the adjacent node set node_successor extended from the current traversed node node_current according to the motion infinitesimal method; 2.5) each node node in the adjacent node set node_successor is traversed, and step 2.6 is performed for each node; 2.6) For nodes not yet included in CloseList and able to be detected by collision, compute the node transfer cost of the current expanding node node to node_current, stack Post as node's temporary cost ; 2.7) If the extended node is not already included in the OpenList, add it; otherwise, compare... and temporary costs The size, and Update to the smaller of the two; 2.8) Calculate and update the integrated generation value of the node and node_current as the parent of node; 2.9) when the target node in OpenList is accessed, the CloseList containing the point sequence in the optimal path is obtained.

4. The method of claim 3, wherein the method further comprises: The node transition cost refers to: the size of the control input between nodes and the length of the motion time, representing the control and time cost required for node transition; The temporary price wherein: during movement from node node to node_current, denotes the corresponding three-axis control input quantity, denotes the movement duration, is a control input and time cost trade-off coefficient.

5. The helicopter low-altitude trajectory planning method based on motion state search and enhanced safety pipeline as described in claim 1, characterized in that, The expansion specifically includes: 3.1) Initialize the sub-pipes shown in the six boundary faces of the cuboid to discrete points on the initial track corresponding coordinates , while initializing the collision detection variables ;​ 3.2) with the maximum number of expansions n as the iteration upper bound, when the expansion boundary and the obstacle point set By collision detection, i.e. the sub-pipeline boundary is continued to expand along the six axes of the spatial coordinate system with a step size of , i.e. ; 3.3) When the number of iterations reaches the upper limit or the cuboid boundary collides with the obstacle, the sub-tube expansion terminates, and the sub-tube described by the 6 boundaries of the cuboid is Then the set of sub-tubes expanded by the algorithm for all initial track points is .

6. The method of claim 1, wherein the method is characterized by, The pruning and interpolation specifically includes: 3.a) The set of sub-pipes obtained after expansion by traversing using a double loop: the set of sub-pipes currently being traversed in the outer loop. The sub-pipes in the middle are used to check for inclusion relationships. When an inclusion relationship occurs... In this case, the initial enhanced safety pipeline set... Remove redundant sub-pipes that are iterated within the inner loop. ; 3.b) The enhanced safety pipeline set after traversal The sub-pipes in the system are continuously monitored, and when adjacent sub-pipes appear... They do not intersect, that is In the case of its corresponding adjacent trackpoint Insert new waypoints between And re-expand and generate new sub-pipes. ; 3.c) When all sub-pipes in the initial enhanced safety pipe are traversed, and the sub-pipes all pass the redundancy inclusion and spatial continuity detection, the enhanced safety pipe after pruning and interpolation is obtained.

7. The method according to claim 6, wherein, The inclusion detection refers to: traversing the expanded sub-pipe set through a double loop mode, comparing the sub-pipes two by two, and detecting whether the space surrounded by the cuboids has an inclusion relationship; The continuity detection refers to: traversing the expanded sub-pipe set through a double loop mode, and detecting whether there is a non-empty space intersection between the cuboids occupied by adjacent sub-pipes. The pruning and interpolation refers to: traversing the expanded sub-pipe set through a double loop mode, and performing redundancy sub-pipe reduction and discontinuous space completion operations through inclusion detection and continuity detection, respectively.

8. The method of claim 1, wherein the method further comprises: searching for a safe flight path of the helicopter based on the motion state of the helicopter. The minimum cost criterion refers to applying inequality constraints to the B-spline control point parameters to be optimized in the enhanced safety pipeline, ensuring that the optimized trajectory in the back end lies within the space defined by the enhanced safety pipeline. Specifically, the B-spline curves used in the back end trajectory optimization process share... Fit M segments of trajectory using control points. The i-th control point in the axial direction satisfy: ,in: , And corresponding to the j-th segment track, , and The sub-pipes shown are respectively cuboids exist The positions of the two boundary surfaces along the axial direction.

9. A system for implementing the helicopter low-altitude flight path planning method based on motion state search and enhanced safety tube according to any one of claims 1-8, characterized in that, It comprises: The global map and task point input module, the front-end track search module, the enhanced safety pipe module, and the back-end track optimization module, wherein: the global map and task point input module performs smoothing and interpolation processing on the original terrain data to generate a task space for track planning; the front-end track search module expands the task space for track planning by establishing a kinematic model considering the differences in three-axis control maneuvers of rotary-wing aircraft, and searches for an initial track using an improved hybrid A* algorithm considering safety margins; the enhanced safety pipe module expands an original safety pipe centered on the track points of the initial track, and obtains an enhanced safety pipe with spatial continuity and simple representation through pruning and interpolation; and the back-end track optimization module applies position constraints to the track fitting control points according to the enhanced safety pipe, and generates a back-end optimized track for execution with a helicopter, so as to ensure that the optimized track is always in a safety space while taking into account the smoothness criteria of the track and maneuver.

Images