An unmanned aerial vehicle emergency landing path planning method and system

Abstract

The application discloses an unmanned aerial vehicle emergency forced landing path planning method and system, comprising the following steps: collecting unmanned aerial vehicle emergency forced landing state data and environmental constraint data; generating a remaining control capability constraint set; constructing a three-dimensional flight space representation, and generating a candidate node set and a candidate connection edge set; performing controllability determination on the candidate connection edge set to generate an initial controllability topology graph; performing topology pruning on the initial controllability topology graph to generate a risk suppression topology graph; calculating an uncertainty entropy value set, performing entropy gating contraction on the risk suppression topology graph to generate a low-uncertainty topology graph; performing safety assertion verification on the low-uncertainty topology graph to generate a verification topology subgraph; extracting a forced landing connected domain set to generate an emergency forced landing path planning result. Through the topology planning method of controllability constraints, risk propagation and uncertainty entropy gating, the unmanned aerial vehicle emergency forced landing path is generated, and the method has the advantages of high safety, strong stability and high executability.

Description

Technical Field

[0001] This invention relates to the field of path planning, and in particular to a method and system for planning emergency landing paths for unmanned aerial vehicles (UAVs). Background Technology

[0002] Drones are widely used in tasks such as power line inspection, disaster monitoring, and emergency rescue. When there is insufficient energy, reduced control capability, or sudden changes in the external environment, emergency landing operations are often required. Existing emergency landing path planning methods for drones are usually based on preset flight rules, simplified environmental models, or a single cost function to search for feasible paths in three-dimensional space. Some methods introduce wind field or obstacle constraints to assist in the assessment of flight safety, but overall they are still mainly based on static or local constraints.

[0003] However, existing technologies generally lack systematic constraint modeling of the remaining control capabilities of UAVs, making it difficult to accurately determine the feasibility of connection paths under control-constrained conditions. Furthermore, the impact of disturbance environments is often assessed using local risk methods, failing to describe the propagation of risks within the flight topology, resulting in insufficient path stability. In addition, existing methods do not adequately consider the combined impact of uncertainties such as positioning, wind field, and control response, lack a unified uncertainty quantification and topology contraction mechanism, and fail to ensure the overall reachability and consistency of the emergency landing terminal domain through safety assertion verification and connectivity analysis, thus affecting the reliability and safety of emergency landing decisions. Summary of the Invention

[0004] One objective of this invention is to propose a method and system for planning emergency landing paths for unmanned aerial vehicles (UAVs). This invention achieves the generation of emergency landing paths for UAVs through a topology planning method based on controllability constraints, risk propagation, and uncertainty entropy gating, and has the advantages of high safety, strong stability, and high executability.

[0005] An emergency landing path planning method for unmanned aerial vehicles (UAVs) according to an embodiment of the present invention includes the following steps: Collect data on the emergency landing status and environmental constraints of drones; Generate a set of remaining control capability constraints based on UAV emergency landing status data; A three-dimensional flight space representation is constructed based on the space occupancy constraint data in the environmental constraint data, and a candidate node set and a candidate connection edge set are generated in the three-dimensional flight space representation; Perform controllability determination on the candidate connection edge set, and verify the consistency of the heading change requirements, climb or descent requirements and energy consumption requirements corresponding to the candidate connection edges based on the remaining control capability constraint set, and generate an initial controllable topology map. A risk propagation map is constructed based on the perturbation environmental data in the environmental constraint data, and the risk propagation intensity is calculated. This is then mapped to the risk constraint parameters of the initial controllability topology map. Based on the risk constraint parameters, topology pruning is performed on the initial controllability topology map to generate a risk suppression topology map. Based on the emergency landing status data of UAVs and the disturbance environment data, the set of uncertainty entropy values ​​is calculated, and entropy gating shrinkage processing is performed on the risk suppression topology to generate a low uncertainty topology. Perform security assertion checks on the low-uncertainty topology graph, remove nodes and edges that do not meet the security assertion checks, and generate a checked topology subgraph; Based on the verification topology subgraph, the set of connected components that can be forced to land is extracted, the target forced landing terminal domain is determined, and the emergency forced landing path planning result is generated.

[0006] Optionally, the emergency landing status data of the UAV includes remaining energy data, control capability data, and motion status data, and the environmental constraint data includes space occupancy constraint data and disturbance environment data.

[0007] Optionally, the generation of the remaining control capability constraint set specifically includes: Acquire remaining energy data, control capability data, and motion status data of the drone in emergency landing status; Calculate available energy budget constraints based on remaining energy data; The maximum controllable steering angular velocity constraint and the maximum stable descent rate constraint are calculated based on the control capability data. Calculate the minimum safe height constraint based on motion state data; The available energy budget constraint, maximum controllable steering angular velocity constraint, maximum stable descent rate constraint, and minimum safe height constraint are uniformly organized and structurally encapsulated to form a set of residual control capability constraints.

[0008] Optionally, the generation of the candidate node set and the candidate connection edge set specifically includes: The spatial boundary of the three-dimensional flight space is determined based on the space occupancy constraint data in the environmental constraint data. The spatial boundary includes the horizontal boundary range and the vertical height range. The spatial occupancy constraint data is uniformly mapped to the same spatial coordinate reference, and the occupancy area merging process is performed on the spatial occupancy constraint data under the spatial coordinate reference to form an occupancy area set; Construct a three-dimensional flight space representation within the spatial boundary, and write an occupancy identifier for the spatial unit in the three-dimensional flight space representation; A set of candidate nodes is generated in the three-dimensional flight space representation, and node coordinate information and node index information are written for each candidate node; A set of candidate connection edges is generated based on the set of candidate nodes, and the starting node index, ending node index, and edge length information are written for each candidate connection edge. The 3D flight space representation, candidate node set, and candidate edge set are encapsulated in a structured manner.

[0009] Optionally, the generation of the initial controllability topology graph specifically includes: Based on the coordinates of the starting node and the ending node of each connecting edge in the candidate connecting edge set, determine the spatial span characteristics of each candidate connecting edge; By combining the horizontal displacement and connecting edge length information in the spatial span characteristics, the heading change requirements corresponding to the candidate connecting edges are derived. By combining the vertical displacement in the spatial span characteristics, the climbing or descending requirements corresponding to the candidate connecting edges are determined; Based on the length information of the candidate connecting edges and combined with the current motion state data of the UAV, determine the energy consumption requirements of the UAV when flying along the candidate connecting edges. The heading change requirement of the candidate connecting edge is compared with the maximum controllable turning angular velocity constraint in the set of remaining controllability constraints. When the heading change requirement exceeds the maximum controllable turning angular velocity constraint, the corresponding candidate connecting edge is marked as an uncontrollable connecting edge. The climbing or descending requirements of candidate connecting edges are compared with the maximum stable descent rate constraint in the set of remaining control capabilities constraints. When the climbing or descending requirements exceed the maximum stable descent rate constraint, the corresponding candidate connecting edge is marked as an uncontrollable connecting edge. The energy consumption demand of the candidate connection edge is compared with the available energy budget constraint in the set of remaining control capability constraints. When the energy consumption demand exceeds the available energy budget constraint, the corresponding candidate connection edge is marked as an uncontrollable connection edge. Based on the spatial path corresponding to the candidate connection edge, determine the minimum altitude trajectory of the candidate connection edge during flight, and check the minimum altitude value of the minimum altitude trajectory against the minimum safe altitude constraint. When the minimum altitude value is lower than the minimum safe altitude constraint, mark the corresponding candidate connection edge as an uncontrollable connection edge. The candidate connection edges marked as uncontrollable are removed from the candidate connection edge set, and the unmarked candidate connection edges are retained to form a controllable connection edge set. An initial controllable topology graph is constructed based on the candidate node set and the controllable connection edge set.

[0010] Optionally, the generation of the risk suppression topology map specifically includes: Based on the disturbance environment data, a set of disturbance-affected areas is determined, and the corresponding spatial location range is marked for each disturbance-affected area; Map the set of disturbance-affected regions to the initial controllability topology graph to determine the spatial relationship between each disturbance-affected region and the nodes and candidate connecting edges in the topology graph; Based on the topological connection relationships between nodes and connecting edges, the state of the disturbance-affected area is propagated step by step along the topological structure to generate a set of disturbance state at the node level and the connecting edge level. Calculate the risk propagation intensity of each node and each connecting edge based on the set of disturbance effect states; The risk propagation intensity is normalized to generate risk constraint parameters for the initial controllability topology graph, and the risk constraint parameters are bound to the corresponding nodes and connecting edges respectively. Based on the risk constraint parameters, the initial controllable topology graph is pruned to remove nodes and connecting edges whose risk constraint parameters are greater than or equal to the preset risk tolerance threshold, thereby generating a risk-suppressed topology graph.

[0011] Optionally, the generation of the low-uncertainty topology graph specifically includes: Based on emergency landing status data and disturbance environment data, the location uncertainty entropy value, wind field uncertainty entropy value and control response uncertainty entropy value are determined and aligned to form a set of uncertainty entropy values; The set of uncertainty entropy values ​​includes location uncertainty entropy values, wind field uncertainty entropy values, and control response uncertainty entropy values; Normalize the set of uncertain entropy values ​​to generate normalized entropy values; Uncertainty entropy structure parameters are constructed based on normalized entropy values. These parameters are used to characterize the synergistic amplification effect and consistency deviation effect of multi-source uncertainties. The topological shrinkage radius is obtained by linearly mapping the structural parameters of the uncertainty entropy. Based on the uncertainty entropy structure parameter and the topology shrinkage radius, entropy-gated shrinkage processing is performed. When the uncertainty entropy structure parameter is greater than the preset entropy gating threshold, spatial domain shrinkage is performed on the risk suppression topology map with the geometric center of the target forced landing area as the center. Nodes and connecting edges whose spatial positions exceed the topology shrinkage radius are removed to generate a low-uncertainty topology map, while maintaining the spatial consistency and connection continuity of nodes in the low-uncertainty topology map.

[0012] Optionally, the generation of the verification topology subgraph specifically includes: In a low-uncertainty topology graph, security assertion checks are performed on the set of nodes and the set of connecting edges, including: When there is no continuous node sequence formed by connecting edges between the current location node of the UAV and the corresponding node of the candidate forced landing area in a low uncertainty topology graph, it is determined that the nodes and connecting edges constituting the low uncertainty topology graph do not satisfy the safety assertion check. When the spatial coordinates of any node fall within the spatial range defined by the boundary information of the occupied area in the spatial occupancy constraint data, it is determined that the node does not meet the security assertion check. When the spatial trajectory range value corresponding to any connecting edge overlaps with the boundary information of the occupied area in the spatial occupancy constraint data, the connecting edge is determined not to satisfy the security assertion check. When the spatial trajectory corresponding to any connecting edge exceeds the spatial domain range determined by the entropy gated contraction process, it is determined that the connecting edge does not satisfy the security assertion check. If the start or end node of any connecting edge is determined to not satisfy the security assertion check, then the connecting edge is determined to not satisfy the security assertion check. Nodes and edges that are determined not to meet the security assertion check are removed, while nodes and edges that are not determined not to meet the security assertion check are retained, and a check topology subgraph is generated.

[0013] Optionally, the generation of the emergency landing path planning results specifically includes: In the verification topology subgraph, the current position node of the UAV is taken as the starting node. Based on the connectivity relationship between the nodes formed by the connecting edges, the verification topology subgraph is subjected to connectivity traversal processing to extract the set of all nodes that are connected to the current position node of the UAV, forming a set of connectable regions that can be forced to land. In the set of connectable regions that can be forced to land, nodes corresponding to candidate forced landing areas are selected as the set of candidate nodes for forced landing terminals; For each node in the set of candidate nodes for forced landing terminals, the corresponding forced landing terminal domain is determined based on its spatial coordinate information and the relationship between the connecting edges. When multiple forced landing terminal domains exist simultaneously, the forced landing terminal domain with the smallest connection path length is selected as the target forced landing terminal domain based on the connection path length information from the current position node of the UAV to each forced landing terminal domain in the verification topology subgraph. After the target forced landing terminal domain is determined, the emergency landing path planning result is generated based on the connection path from the current location node of the UAV to the target forced landing terminal domain in the verification topology subgraph.

[0014] An emergency landing path planning system for unmanned aerial vehicles (UAVs) according to an embodiment of the present invention includes: The emergency landing data acquisition module is used to collect emergency landing status data and environmental constraint data of the UAV; The control capability constraint generation module is used to generate the set of remaining control capability constraints; The 3D flight space construction module is used to construct a 3D flight space representation based on space occupancy constraint data, and generate a set of candidate nodes and a set of candidate connecting edges in the 3D flight space representation; The controllability determination module is used to perform controllability determination on the candidate connection edge set. Based on the remaining control capability constraint set, it performs consistency verification on the heading change requirements, climb or descent requirements and energy consumption requirements corresponding to the candidate connection edges, and generates an initial controllability topology map. The risk propagation modeling module is used to construct a risk propagation map based on disturbed environment data and calculate the risk propagation intensity. It maps the risk propagation intensity to the risk constraint parameters of the initial controllability topology map and performs topology pruning on the initial controllability topology map based on the risk constraint parameters to generate a risk suppression topology map. The uncertainty entropy modeling module is used to calculate the set of uncertainty entropy values ​​and perform entropy gating shrinkage on the risk suppression topology to generate a low-uncertainty topology. The security assertion verification module is used to perform security assertion verification on low-uncertainty topology graphs, remove nodes and connecting edges that do not meet the security assertion verification, and generate a verified topology subgraph. The emergency landing path generation module is used to extract the set of connectable regions that can be landed based on the verified topology subgraph, determine the target emergency landing terminal domain, and generate emergency landing path planning results.

[0015] The beneficial effects of this invention are: This invention unifies the modeling of UAV emergency landing status data with environmental constraint data, introduces a set of residual control capability constraints, and performs consistency verification on the heading change requirements, climb or descent requirements, and energy consumption requirements of candidate connection edges. This ensures that the path planning process is controlled from the source by the actual executable capabilities of the UAV in an emergency state, thereby avoiding the generation of flight paths that are impossible under power, attitude, or energy conditions. Compared to existing methods based solely on geometric reachability or a single cost function, this invention eliminates uncontrollable paths at the topology level in advance, improving the executability and reliability of emergency landing path planning results in actual flight control.

[0016] Furthermore, this invention constructs a risk propagation graph and calculates the risk propagation intensity, mapping the impact of disturbances to the topology as risk constraint parameters. It also performs entropy-gated contraction of the topology using an uncertainty entropy value set, effectively suppressing interference from high-risk, high-uncertainty regions in path planning. Based on this, a safety assertion verification mechanism is used to uniformly constrain nodes and connecting edges, and the set of connectable regions suitable for forced landing is extracted using the verified topology subgraph. This achieves stable determination of the forced landing terminal domain and path generation under complex disturbance environments. This method can make reasonable selections based on topology information when multiple forced landing terminal domains coexist, comprehensively improving the safety, consistency, and stability of the emergency landing decision-making process, and providing more reliable path planning results for UAVs performing forced landing missions under extreme or adverse conditions. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of an emergency landing path planning method for unmanned aerial vehicles (UAVs) proposed in this invention; Figure 2 This is a schematic diagram illustrating the execution entropy gating contraction and generation of a low-uncertainty topology graph in the UAV emergency landing path planning method proposed in this invention. Detailed Implementation

[0018] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0019] refer to Figures 1-2 A method for planning emergency landing paths for unmanned aerial vehicles (UAVs) includes the following steps: Collect data on the emergency landing status and environmental constraints of drones; Generate a set of remaining control capability constraints based on UAV emergency landing status data; A three-dimensional flight space representation is constructed based on the space occupancy constraint data in the environmental constraint data, and a candidate node set and a candidate connection edge set are generated in the three-dimensional flight space representation; Perform controllability determination on the candidate connection edge set, and verify the consistency of the heading change requirements, climb or descent requirements and energy consumption requirements corresponding to the candidate connection edges based on the remaining control capability constraint set, and generate an initial controllable topology map. A risk propagation map is constructed based on the perturbation environmental data in the environmental constraint data, and the risk propagation intensity is calculated. This is then mapped to the risk constraint parameters of the initial controllability topology map. Based on the risk constraint parameters, topology pruning is performed on the initial controllability topology map to generate a risk suppression topology map. Based on the emergency landing status data of UAVs and the disturbance environment data, the set of uncertainty entropy values ​​is calculated, and entropy gating shrinkage processing is performed on the risk suppression topology to generate a low uncertainty topology. Perform security assertion checks on the low-uncertainty topology graph, remove nodes and edges that do not meet the security assertion checks, and generate a checked topology subgraph; Based on the verification topology subgraph, the set of connected components that can be forced to land is extracted, the target forced landing terminal domain is determined, and the emergency forced landing path planning result is generated.

[0020] In this embodiment, the emergency landing status data of the UAV includes remaining energy data, control capability data, and motion status data. The remaining energy data is used to characterize the energy reserve information of the UAV that can be used for continuous flight and descent operations in the emergency landing state. The control capability data is used to characterize the control capability information of the UAV that can actually perform flight actions in the emergency landing state. The motion status data is used to characterize the status information of the UAV's real-time flight state in the emergency landing state, including the current position coordinates and the current speed status. The environmental constraint data includes space occupancy constraint data and disturbance environment data. The space occupancy constraint data is used to characterize the spatial constraint information of the distribution of inaccessible or restricted areas in the UAV's flight space, including the boundary information of the occupied area in three-dimensional space. The disturbance environment data is used to characterize the external environmental disturbance information that affects the flight stability and control feasibility of the UAV, including wind speed information and wind direction information.

[0021] In this embodiment, the generation of the residual control capability constraint set specifically includes: Acquire remaining energy data, control capability data, and motion status data of the drone in emergency landing status; Based on the remaining energy data, the available energy budget constraint is calculated, and the remaining energy data is quantified as the energy budget limit used for flight maintenance, steering adjustment and descent operations during emergency landing, which serves as the available energy budget constraint. The maximum controllable steering angular velocity constraint and the maximum stable descent rate constraint are calculated based on the control capability data. Based on the steering control capability reflected in the control capability data, the upper limit of the steering angle velocity allowed for the UAV in emergency landing is determined as the maximum controllable steering angle velocity constraint. Based on the vertical control capability and attitude stability capability reflected in the control capability data, the upper limit of the descent rate allowed for the UAV in emergency landing is determined as the maximum stable descent rate constraint. The minimum safe altitude constraint is calculated based on motion state data. Based on flight altitude status and descent trend information, the minimum flight altitude limit that the UAV needs to maintain during emergency landing is determined as the minimum safe altitude constraint. The available energy budget constraint, maximum controllable steering angular velocity constraint, maximum stable descent rate constraint, and minimum safe height constraint are uniformly organized and structurally encapsulated to form a set of residual control capability constraints.

[0022] In this embodiment, the generation of the candidate node set and the candidate connection edge set specifically includes: The spatial boundary of the three-dimensional flight space is determined based on the space occupancy constraint data in the environmental constraint data. The spatial boundary includes the horizontal boundary range and the vertical height range, and the vertical height range is consistent with the minimum safe height constraint. The spatial occupancy constraint data is uniformly mapped to the same spatial coordinate reference. Under the spatial coordinate reference, the spatial occupancy constraint data is processed to merge the occupancy regions, forming an occupancy region set, which includes the boundary information of the occupancy regions. Performing the occupancy area merging process refers to merging occupancy areas in the spatial occupancy constraint data that have overlapping, intersecting, or spacing smaller than the preset spatial resolution under a unified spatial coordinate reference, forming a set of non-overlapping occupancy areas, and representing each occupancy area with the merged spatial boundary information; A three-dimensional flight space representation is constructed within the spatial boundary, and an occupation identifier is written for the spatial unit in the three-dimensional flight space representation. The occupation identifier is used to distinguish between passable spatial units and occupied spatial units. The three-dimensional flight space representation adopts a three-dimensional node index structure. A candidate node set is generated in the three-dimensional flight space representation, and node coordinate information and node index information are written for each candidate node. The candidate node set consists of spatial nodes that satisfy the occupancy identifier as passable space unit and satisfy the vertical height range constraint. A candidate edge set is generated based on the candidate node set, and the starting node index, ending node index and edge length information are written for each candidate edge. The candidate edge set consists of the edge that satisfies the spatial connectivity condition between any two candidate nodes. The spatial connectivity condition includes that the connecting line segment does not cross the boundary of the occupied area and the vertical height value of the connecting line segment is within the vertical height range. The 3D flight space representation, candidate node set, and candidate edge set are encapsulated in a structured manner.

[0023] In this embodiment, the generation of the initial controllability topology graph specifically includes: Based on the coordinates of the starting node and the ending node of each connecting edge in the candidate connecting edge set, the spatial span characteristics of each candidate connecting edge are determined. The spatial span characteristics include horizontal displacement, vertical displacement and connecting edge length information. By combining the horizontal displacement and connecting edge length information in the spatial span characteristics, the heading change requirements corresponding to the candidate connecting edge are derived, which characterizes the heading adjustment range that the UAV needs to complete when flying along the candidate connecting edge. By combining the vertical displacement in the spatial span characteristics, the climb or descent requirements corresponding to the candidate connecting edge are determined. This is used to characterize the longitudinal height adjustment action required for the UAV to perform under a given spatial span when flying along the candidate connecting edge. It is determined by the vertical displacement of the candidate connecting edge and the length of the connecting edge, and is used to characterize the longitudinal adjustment intensity required to complete the flight along the connecting edge. Based on the length information of the candidate connecting edges and combined with the current motion state data of the UAV, determine the energy consumption requirements of the UAV when flying along the candidate connecting edges. The heading change requirement of the candidate connecting edge is compared with the maximum controllable turning angular velocity constraint in the set of remaining controllability constraints. When the heading change requirement exceeds the maximum controllable turning angular velocity constraint, the corresponding candidate connecting edge is marked as an uncontrollable connecting edge. The climbing or descending requirements of candidate connecting edges are compared with the maximum stable descent rate constraint in the set of remaining control capabilities constraints. When the climbing or descending requirements exceed the maximum stable descent rate constraint, the corresponding candidate connecting edge is marked as an uncontrollable connecting edge. The energy consumption demand of the candidate connection edge is compared with the available energy budget constraint in the set of remaining control capability constraints. When the energy consumption demand exceeds the available energy budget constraint, the corresponding candidate connection edge is marked as an uncontrollable connection edge. Based on the spatial path corresponding to the candidate connection edge, determine the minimum altitude trajectory of the candidate connection edge during flight, and check the minimum altitude value of the minimum altitude trajectory against the minimum safe altitude constraint. When the minimum altitude value is lower than the minimum safe altitude constraint, mark the corresponding candidate connection edge as an uncontrollable connection edge. The candidate connection edges marked as uncontrollable are removed from the candidate connection edge set, and the unmarked candidate connection edges are retained to form a controllable connection edge set. An initial controllable topology graph is constructed based on the candidate node set and the controllable connection edge set.

[0024] In this embodiment, the generation of the risk suppression topology map specifically includes: Based on the disturbance environment data, a set of disturbance-affected areas is determined, and the corresponding spatial location range is marked for each disturbance-affected area; Under a unified spatial coordinate reference, based on wind speed and direction information from the disturbance environment data, the direction and intensity values ​​of the external disturbances experienced by the UAV at its current flight altitude and position are calculated. The main influence direction of the disturbance is determined based on the direction value, and the disturbance influence distance is determined based on the intensity value. Starting from the UAV's current position, the disturbance coverage area is delineated in the three-dimensional flight space representation along the main influence direction of the disturbance and according to the disturbance influence distance value, forming a single disturbance influence area. The above processing is performed on the disturbance coverage areas corresponding to different spatial positions and different altitude layers to generate multiple disturbance influence areas, and all disturbance influence areas are aggregated to form a disturbance influence area set. Map the set of disturbance-affected regions to the initial controllability topology graph to determine the spatial relationship between each disturbance-affected region and the nodes and candidate connecting edges in the topology graph; The process involves sequentially obtaining the spatial location range values ​​corresponding to each disturbance-affected region in the disturbance-affected region set; for each node in the initial controllable topology graph, obtaining the spatial coordinate values ​​corresponding to the node, and determining whether the node's spatial coordinate values ​​fall within the spatial location range values ​​to generate a node disturbance-affected identifier value; for each candidate connection edge in the initial controllable topology graph, obtaining the spatial trajectory range values ​​corresponding to the connection edge, and determining whether the spatial trajectory range values ​​overlap with the spatial location range values ​​to generate a connection edge disturbance-affected identifier value; based on the node disturbance-affected identifier value and the connection edge disturbance-affected identifier value, determining the spatial association relationship between each disturbance-affected region and the nodes and candidate connection edges in the topology graph, and forming a disturbance-affected region-node association set and a disturbance-affected region-connection edge association set. When the node disturbance-affected identifier value or the connection edge disturbance-affected identifier value is 1, it indicates that the corresponding node or candidate connection edge is located within the disturbance-affected region and is affected by the disturbance; when the identifier value is 0, it indicates that the corresponding node or candidate connection edge is not affected by the disturbance. Based on the topological connection relationships between nodes and connecting edges, the state of the disturbance-affected area is propagated step by step along the topological structure to generate a set of disturbance state at the node level and the connecting edge level. Nodes with a disturbance-affected flag value of 1 and candidate connecting edges with a disturbance-affected flag value of 1 in the disturbance-affected region-node association set and the disturbance-affected region-connecting edge association set are used as initial propagation sources, and initial disturbance effect state values ​​are generated for each initial propagation source. For any node in the initial controllability topology graph, the topology level value of the node is determined, which is the minimum number of connecting edges traversed between the node and any initial propagation source. For this node, the cumulative disturbance effect state value of the node is calculated, which is the sum of the disturbance effect state values ​​contributed by all initial propagation sources to this node. The disturbance state value contributed by the node is the initial disturbance state value of the initial propagation source divided by the node's topology level value. The cumulative disturbance state value of the node is taken as the node disturbance state value, and the above calculation is performed on all nodes in the initial controllable topology graph to form a node-level disturbance state set. For any candidate connection edge in the initial controllable topology graph, the node disturbance state values ​​of the nodes at both ends of the candidate connection edge are determined, and the average value of the node disturbance state values ​​of the two ends of the node is taken as the connection edge disturbance state value. The above calculation is performed on all candidate connection edges to form a connection edge-level disturbance state set. When obtaining the initial disturbance effect state value for each node and each candidate connection edge, a corresponding effect intensity value is generated for each disturbance influence region. The effect intensity value is a numerical value calculated based on the wind speed and wind direction information within the disturbance influence region, used to characterize the degree of influence of the disturbance influence region on the UAV's flight state. For any node, it is determined whether it falls within the spatial location range of a disturbance influence region based on its spatial coordinates. If it does, the effect intensity value of that disturbance influence region is assigned as the initial disturbance effect state value for that node. If it falls within multiple disturbance influence regions simultaneously, the effect intensity values ​​of the multiple disturbance influence regions are averaged to obtain the initial disturbance effect state value for that node. If the action state value does not fall within any disturbance influence area, then 0 is assigned as the initial disturbance action state value of the node. For any candidate connection edge, it is determined whether it spatially overlaps with the spatial location range value of a disturbance influence area based on the spatial trajectory range value corresponding to the connection edge. If spatial overlap occurs, then the effect intensity value of the disturbance influence area is assigned as the initial disturbance action state value of the candidate connection edge. If it spatially overlaps with multiple disturbance influence areas at the same time, then the effect intensity values ​​of the multiple disturbance influence areas are averaged to obtain the initial disturbance action state value of the candidate connection edge. If no spatial overlap occurs, then 0 is assigned as the initial disturbance action state value of the candidate connection edge. The state of the disturbance-affected region is the disturbance-affected state of the disturbance-affected region on the nodes and candidate connecting edges in the initial controllable topology graph. It is based on mapping the disturbance-affected region to the topology graph and determining whether the spatial coordinates of the nodes fall within the spatial location range of the disturbance-affected region and whether the spatial trajectory of the connecting edge overlaps with the disturbance-affected region, thereby obtaining the disturbance-affected value of the nodes and the disturbance-affected value of the connecting edges. The disturbance effect state is the node-level disturbance effect state and the connection-edge-level disturbance effect state formed after the above disturbance effects propagate step by step along the connection relationship of nodes and connecting edges in the topology. The disturbance effect state value is the numerical value after quantifying the disturbance effect state; The risk propagation intensity of each node and each connecting edge is calculated based on the set of disturbance effect states. The risk propagation intensity is used to characterize the degree of impact of the disturbance environment on the reachability of nodes and the executability of connecting edges. Obtain the node perturbation state value corresponding to each node in the node-level perturbation state set, and use the node perturbation state value as the node risk propagation intensity of the corresponding node; obtain the connection edge perturbation state value corresponding to each connection edge in the connection edge perturbation state set, and use the connection edge perturbation state value as the connection edge risk propagation intensity of the corresponding connection edge; generate a unique risk propagation intensity for each node and each connection edge; The risk propagation intensity is normalized to generate risk constraint parameters for the initial controllability topology graph, and the risk constraint parameters are bound to the corresponding nodes and connecting edges respectively. The risk constraint parameter value is used to characterize the relative risk level of the node or the connecting edge in the initial controllable topology graph. The risk constraint parameter value is bound to the corresponding node and the corresponding connecting edge to form an initial controllable topology graph with risk constraint parameters. Based on the risk constraint parameters, the initial controllable topology graph is pruned to remove nodes and connecting edges whose risk constraint parameters are greater than or equal to the preset risk tolerance threshold, thereby generating a risk-suppressed topology graph.

[0025] In this embodiment, the generation of the low-uncertainty topology graph specifically includes: Based on emergency landing status data and disturbance environment data, the location uncertainty entropy value, wind field uncertainty entropy value and control response uncertainty entropy value are determined and aligned to form a set of uncertainty entropy values; Under a unified time reference, a preset time window is selected. The current position coordinate sequence within this time window is extracted from the motion state data. A position change sequence is generated based on the changes in the current position coordinates at adjacent times. The position change sequence is divided into preset intervals, and the frequency of occurrence of each interval is counted. A positioning uncertainty entropy value is calculated based on the frequency of occurrence. The wind speed and wind direction information sequences within this time window are extracted from the disturbance environment data. The wind speed change sequence and wind direction change sequence are calculated and normalized. The two are weighted and summed to obtain the disturbance state sequence. The disturbance state sequence is divided into several disturbance state intervals, and the wind field uncertainty entropy value is calculated based on the frequency of occurrence of each interval. The maximum controllable steering angular velocity constraint and the maximum stable descent rate constraint are extracted from the control capability data. The current speed state sequence within the corresponding time window is extracted from the motion state data. A control response deviation sequence is generated based on the deviation between the current speed state sequence and the corresponding control capability constraint. The control response deviation sequence is divided into preset intervals, and the frequency of occurrence of each interval is counted. A control response uncertainty entropy value is calculated based on the frequency of occurrence. When generating the control response deviation sequence, the maximum controllable steering angular velocity constraint and the maximum stable descent rate constraint are extracted from the control capability data, and the current speed state sequence within the time window is extracted from the motion state data. Based on the speed direction change amplitude and the speed vertical component change amplitude at each moment in the current speed state sequence, the actual steering change and actual descent change at the corresponding moment are determined respectively. The actual steering change is compared with the maximum controllable steering angular velocity constraint moment by moment to obtain the steering deviation value, and the actual descent change is compared with the maximum stable descent rate constraint moment by moment to obtain the descent deviation value. Then, the steering deviation value and descent deviation value at the same moment are normalized and weighted to generate the control response deviation value at the corresponding moment, and arranged in chronological order to form the control response deviation sequence. The set of uncertainty entropy values ​​includes location uncertainty entropy values, wind field uncertainty entropy values, and control response uncertainty entropy values; Normalize the set of uncertain entropy values ​​to generate normalized entropy values; Uncertainty entropy structure parameters are constructed based on normalized entropy values. These parameters are used to characterize the synergistic amplification effect and consistency deviation effect of multi-source uncertainties.

[0026] in, Represents the structural parameters of uncertainty entropy. This represents the normalized positioning uncertainty entropy value. This represents the normalized wind field uncertainty entropy value. This represents the normalized control response uncertainty entropy value. and Represents the linear fusion coefficient. Indicates the synergistic amplification factor. and Indicates the coefficient of consistency deviation; The linear fusion coefficient, the cooperative amplification coefficient, and the consistency deviation coefficient are all preset parameters. The linear fusion coefficient is used to characterize the basic weight ratio of the positioning uncertainty entropy value, the wind field uncertainty entropy value, and the control response uncertainty entropy value in the overall uncertainty entropy structure parameter. The cooperative amplification coefficient is used to characterize the amplification effect of multiple uncertainties being at a high level at the same time on the degree of influence of the overall uncertainty. The consistency deviation coefficient is used to characterize the degree of influence of the difference in values ​​among multiple uncertainties on the stability of the overall uncertainty. The above coefficients are set according to the flight control capability and safety strategy requirements during the initialization phase and remain unchanged during the emergency landing path planning process to ensure the determinism and repeatability of the calculation process of the overall uncertainty entropy structure parameter. By using a linear fusion coefficient to differentiate the basic contribution ratios of positioning uncertainty entropy, wind field uncertainty entropy, and control response uncertainty entropy, the overall uncertainty entropy structure parameter can reflect the different impact weights of different uncertainty sources on the feasibility of UAV emergency landing. By using a synergistic amplification coefficient to characterize the coupling effect when multiple uncertainties are at a high level simultaneously, the overall uncertainty entropy structure parameter can reflect the nonlinear risk amplification characteristics of the landing path feasibility when multiple uncertainties are superimposed. By using a consistency deviation coefficient to constrain the stability reduction effect caused by the differences in values ​​among multiple uncertainties, the overall uncertainty entropy structure parameter can reflect the adverse effects of inconsistent changes in various uncertainties on flight control stability and decision reliability, thereby improving the reliability and safety of the overall uncertainty assessment results in emergency landing scenarios. The topological shrinkage radius is obtained by linearly mapping the structural parameters of the uncertainty entropy. Based on the uncertainty entropy structure parameters and the topology shrinkage radius, entropy-gated shrinkage processing is performed. When the uncertainty entropy structure parameters are greater than the preset entropy gating threshold, spatial domain shrinkage is performed on the risk suppression topology map with the geometric center of the target landing area as the center. Nodes and connecting edges whose spatial positions exceed the topology shrinkage radius are removed to generate a low-uncertainty topology map, while maintaining the spatial consistency and connection continuity of nodes in the low-uncertainty topology map. The target forced landing area refers to the candidate forced landing area that meets the forced landing spatial conditions in the risk suppression topology map. The preset entropy gate threshold is the upper limit of uncertainty entropy that is set in advance based on the test calibration results of historical flight data, under the premise of meeting the safety requirements of emergency forced landing, and is used to determine the acceptable range of uncertainty under the current emergency forced landing state.

[0027] In this embodiment, the generation of the verification topology subgraph specifically includes: In a low-uncertainty topology graph, security assertion checks are performed on the set of nodes and the set of connecting edges, including: When there is no continuous node sequence formed by connecting edges between the current location node of the UAV and the corresponding node of the candidate forced landing area in a low uncertainty topology graph, it is determined that the nodes and connecting edges constituting the low uncertainty topology graph do not satisfy the safety assertion check. When the spatial coordinates of any node fall within the spatial range defined by the boundary information of the occupied area in the spatial occupancy constraint data, it is determined that the node does not meet the security assertion check. When the spatial trajectory range value corresponding to any connecting edge overlaps with the boundary information of the occupied area in the spatial occupancy constraint data, the connecting edge is determined not to satisfy the security assertion check. When the spatial trajectory corresponding to any connecting edge exceeds the spatial domain range determined by the entropy gated contraction process, it is determined that the connecting edge does not satisfy the security assertion check. If the start or end node of any connecting edge is determined to not satisfy the security assertion check, then the connecting edge is determined to not satisfy the security assertion check. Nodes and edges that are determined not to meet the security assertion check are removed, while nodes and edges that are not determined not to meet the security assertion check are retained, and a check topology subgraph is generated.

[0028] In this embodiment, the generation of emergency landing path planning results specifically includes: In the verification topology subgraph, the current position node of the UAV is taken as the starting node. Based on the connectivity relationship between the nodes formed by the connecting edges, the verification topology subgraph is subjected to connectivity traversal processing to extract the set of all nodes that are connected to the current position node of the UAV, forming a set of connectable regions that can be forced to land. In the set of connectable regions that can be forced to land, nodes corresponding to candidate forced landing areas are selected as the set of candidate nodes for forced landing terminals; For each node in the candidate node set for forced landing terminals, the corresponding forced landing terminal domain is determined based on its spatial coordinate information and connection edge relationship. The forced landing terminal domain consists of nodes that correspond to the candidate forced landing area and are connected in the verification topology subgraph. When multiple forced landing terminal domains exist simultaneously, the forced landing terminal domain with the smallest connection path length is selected as the target forced landing terminal domain based on the connection path length information from the current position node of the UAV to each forced landing terminal domain in the verification topology subgraph. After the target forced landing terminal domain is determined, the emergency landing path planning result is generated based on the connection path from the current location node of the UAV to the target forced landing terminal domain in the verification topology subgraph.

[0029] An emergency landing path planning system for unmanned aerial vehicles (UAVs) includes: The emergency landing data acquisition module is used to collect emergency landing status data and environmental constraint data of the UAV; The control capability constraint generation module is used to generate the set of remaining control capability constraints; The 3D flight space construction module is used to construct a 3D flight space representation based on space occupancy constraint data, and generate a set of candidate nodes and a set of candidate connecting edges in the 3D flight space representation; The controllability determination module is used to perform controllability determination on the candidate connection edge set. Based on the remaining control capability constraint set, it performs consistency verification on the heading change requirements, climb or descent requirements and energy consumption requirements corresponding to the candidate connection edges, and generates an initial controllability topology map. The risk propagation modeling module is used to construct a risk propagation map based on disturbed environment data and calculate the risk propagation intensity. It maps the risk propagation intensity to the risk constraint parameters of the initial controllability topology map and performs topology pruning on the initial controllability topology map based on the risk constraint parameters to generate a risk suppression topology map. The uncertainty entropy modeling module is used to calculate the set of uncertainty entropy values ​​and perform entropy gating shrinkage on the risk suppression topology to generate a low-uncertainty topology. The security assertion verification module is used to perform security assertion verification on low-uncertainty topology graphs, remove nodes and connecting edges that do not meet the security assertion verification, and generate a verified topology subgraph. The emergency landing path generation module is used to extract the set of connectable regions that can be landed based on the verified topology subgraph, determine the target emergency landing terminal domain, and generate emergency landing path planning results.

[0030] Example 1: To verify the feasibility of the present invention in practice, it was applied to an emergency landing support mission for a multi-rotor UAV in a complex urban scenario in a coastal city. In this scenario, the UAV is responsible for inspecting urban infrastructure and emergency situation awareness. The flight area has a variety of complex environmental constraints, such as high-density building clusters, overhead lines, temporary construction areas, and unstable sea wind disturbances. When the UAV experiences sudden energy decay and control performance degradation during the mission, the traditional landing method that relies on fixed alternative landing points or single rule judgment is difficult to provide a safe, executable, and risk-controllable landing path in a timely manner. This can easily lead to path interruption, risk accumulation, or unreachable landing areas, which seriously affects the safety of the UAV and the ground environment.

[0031] In this scenario, the method of the present invention is deployed on an onboard computing platform of an unmanned aerial vehicle (UAV) and operates in real time in conjunction with the flight control system and the environmental perception system. When the UAV enters an emergency landing state, the system first synchronously acquires the UAV's current remaining energy state, executable control capabilities, and real-time motion status information. Simultaneously, it accesses urban 3D spatial constraint data and real-time wind field disturbance information. Based on this data, the system automatically generates a set of remaining control capability constraints to limit the executable range of various heading changes, descent behaviors, and energy consumption during subsequent path reasoning. Subsequently, the system constructs a 3D flight space representation under a unified spatial coordinate reference and automatically generates candidate nodes and candidate connecting edges that meet the spatial traversability conditions. Based on this, controllability is determined, forming an initial controllable topology structure containing only executable flight paths.

[0032] To address the challenges of frequent wind direction changes and significant local disturbances in urban environments, the system further constructs a disturbance environmental risk propagation map. This model models the propagation of disturbance impacts along the topology, enabling path risks to reflect the cumulative risk trend across the overall spatial connectivity rather than relying solely on local judgments. Based on this, the system combines UAV positioning stability, wind field variation characteristics, and control response deviations to calculate multi-source uncertainty entropy structural parameters. These parameters are then used to perform entropy-gated contraction processing on the topology, proactively eliminating spatial regions with excessive uncertainty and insufficient executability. This results in a structurally stable and risk-controllable low-uncertainty topology. Subsequently, a security assertion verification mechanism checks the integrity and security of nodes and connections within the topology, ultimately generating a verified topology subgraph that satisfies all constraints.

[0033] In this verification topology subgraph, the system starts from the UAV's current position, traverses and analyzes the spatial regions that maintain connectivity, automatically extracts the set of connectable regions for emergency landing, and, when multiple candidate emergency landing terminal regions coexist, determines the optimal target emergency landing terminal region by combining path connectivity length and spatial reachability, ultimately outputting a complete emergency landing path planning result. Through continuous application verification in this urban scenario, the method of this invention can effectively solve problems such as unreachable emergency landing paths, fragmented risk assessment, and unstable decision-making in complex environments. It completes path planning and updating within a limited time, improving the reliability, continuity, and environmental adaptability of emergency landing decisions, and providing stable and executable technical support for the safe emergency landing of UAVs in real complex scenarios.

[0034] To verify the performance of the present invention, it was compared with the traditional method. The comparison results are shown in Table 1.

[0035] Table 1. Performance Comparison of Different Path Planning Methods in Emergency Landing Scenarios

[0036] As shown in Table 1, the traditional rule-driven forced landing method has an average path planning time of 186 milliseconds, which is relatively low. However, its path planning success rate is only 71.4%, the forced landing path interruption rate reaches 18.3%, and the high-risk area mis-entry rate is 21.5%. These data show that although the method has a simple calculation process, it relies solely on fixed rules to filter paths and does not model the overall connectivity structure of the flight space and the propagation impact of environmental disturbances. This leads to situations where the path is locally feasible but overall unreachable during actual forced landings, resulting in a low success rate and a high interruption rate.

[0037] The forced landing method based on single risk assessment shows improvements in safety indicators compared to traditional rule-based methods. Its success rate of single-path planning increases to 83.9%, and the false entry rate into high-risk areas decreases to 12.1%, indicating that the ability of the path to avoid high-risk spaces is enhanced after the introduction of risk assessment. However, the average path planning time of this method increases to 241 milliseconds. This indicates that although the method assesses the risks, the risk information mainly participates in path selection in the form of local weights, lacking a characterization of the risk propagation effect in the topology. As a result, once environmental disturbances change during flight, the original path still needs to be frequently adjusted.

[0038] In comparison, the emergency landing path planning method of this invention achieves a 92.6% success rate in one-time path planning, higher than the other two methods. Simultaneously, the path interruption rate is reduced to 5.2%, and the high-risk area mis-entry rate is reduced to 7.8%. These data indicate that this invention effectively eliminates structurally unsustainable nodes and connecting edges with high risk accumulation during the path generation stage, ensuring high continuity and safety of the final path during execution. Its average path planning time is 185 milliseconds, roughly on par with traditional rule-based methods and significantly lower than methods based on single risk assessments. This demonstrates that by performing risk trimming and uncertainty reduction at the topology level, the computational overhead required for subsequent path revisions is reduced.

[0039] Further analysis of the reachability of the emergency landing terminal area and the path stability index shows that the reachability of the emergency landing terminal area of ​​the present invention reaches 94.1%, and the number of path stability fluctuations is 1.2 times, both of which are better than the comparison method. This indicates that after extracting the connected components in the verification topology subgraph, the selected emergency landing terminal area maintains a stable connection relationship in space, thereby reducing the risk of path failure caused by topological breaks during flight.

[0040] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for planning emergency landing paths for unmanned aerial vehicles (UAVs), characterized in that, Includes the following steps: Collect data on the emergency landing status and environmental constraints of drones; Generate a set of remaining control capability constraints based on UAV emergency landing status data; A three-dimensional flight space representation is constructed based on the space occupancy constraint data in the environmental constraint data, and a candidate node set and a candidate connection edge set are generated in the three-dimensional flight space representation; Perform controllability determination on the candidate connection edge set, and verify the consistency of the heading change requirements, climb or descent requirements and energy consumption requirements corresponding to the candidate connection edges based on the remaining control capability constraint set, and generate an initial controllable topology map. A risk propagation map is constructed based on the perturbation environmental data in the environmental constraint data, and the risk propagation intensity is calculated. This is then mapped to the risk constraint parameters of the initial controllability topology map. Based on the risk constraint parameters, topology pruning is performed on the initial controllability topology map to generate a risk suppression topology map. Based on the emergency landing status data of UAVs and the disturbance environment data, the set of uncertainty entropy values ​​is calculated, and entropy gating shrinkage processing is performed on the risk suppression topology to generate a low uncertainty topology. Perform security assertion checks on the low-uncertainty topology graph, remove nodes and edges that do not meet the security assertion checks, and generate a checked topology subgraph; Based on the verification topology subgraph, the set of connected components that can be forced to land is extracted, the target forced landing terminal domain is determined, and the emergency forced landing path planning results are generated. The generation of the risk suppression topology specifically includes: Based on the disturbance environment data, a set of disturbance-affected areas is determined, and the corresponding spatial location range is marked for each disturbance-affected area; Map the set of disturbance-affected regions to the initial controllability topology graph to determine the spatial relationship between each disturbance-affected region and the nodes and candidate connecting edges in the topology graph; Based on the topological connection relationships between nodes and connecting edges, the state of the disturbance-affected area is propagated step by step along the topological structure to generate a set of disturbance state at the node level and the connecting edge level. Calculate the risk propagation intensity of each node and each connecting edge based on the set of disturbance effect states; The risk propagation intensity is normalized to generate risk constraint parameters for the initial controllability topology graph, and the risk constraint parameters are bound to the corresponding nodes and connecting edges respectively. Based on the risk constraint parameters, the initial controllable topology graph is pruned to remove nodes and connecting edges whose risk constraint parameters are greater than or equal to the preset risk tolerance threshold, thereby generating a risk-suppressed topology graph.

2. The method for planning an emergency landing path for a drone according to claim 1, characterized in that, The emergency landing status data of the UAV includes remaining energy data, control capability data, and motion status data, while the environmental constraint data includes space occupancy constraint data and disturbance environment data.

3. The method for planning an emergency landing path for a drone according to claim 1, characterized in that, The generation of the remaining control capability constraint set specifically includes: Acquire remaining energy data, control capability data, and motion status data of the drone in emergency landing status; Calculate available energy budget constraints based on remaining energy data; The maximum controllable steering angular velocity constraint and the maximum stable descent rate constraint are calculated based on the control capability data. Calculate the minimum safe height constraint based on motion state data; The available energy budget constraint, maximum controllable steering angular velocity constraint, maximum stable descent rate constraint, and minimum safe height constraint are uniformly organized and structurally encapsulated to form a set of residual control capability constraints.

4. The method for planning an emergency landing path for a drone according to claim 1, characterized in that, The generation of the candidate node set and the candidate connection edge set specifically includes: The spatial boundary of the three-dimensional flight space is determined based on the space occupancy constraint data in the environmental constraint data. The spatial boundary includes the horizontal boundary range and the vertical height range. The spatial occupancy constraint data is uniformly mapped to the same spatial coordinate reference, and the occupancy area merging process is performed on the spatial occupancy constraint data under the spatial coordinate reference to form an occupancy area set; Construct a three-dimensional flight space representation within the spatial boundary, and write an occupancy identifier for the spatial unit in the three-dimensional flight space representation; A set of candidate nodes is generated in the three-dimensional flight space representation, and node coordinate information and node index information are written for each candidate node; A set of candidate connection edges is generated based on the set of candidate nodes, and the starting node index, ending node index, and edge length information are written for each candidate connection edge. The 3D flight space representation, candidate node set, and candidate edge set are encapsulated in a structured manner.

5. The method for planning an emergency landing path for a drone according to claim 1, characterized in that, The generation of the initial controllability topology specifically includes: Based on the coordinates of the starting node and the ending node of each connecting edge in the candidate connecting edge set, determine the spatial span characteristics of each candidate connecting edge; By combining the horizontal displacement and connecting edge length information in the spatial span characteristics, the heading change requirements corresponding to the candidate connecting edges are derived. By combining the vertical displacement in the spatial span characteristics, the climbing or descending requirements corresponding to the candidate connecting edges are determined; Based on the length information of the candidate connecting edges and combined with the current motion state data of the UAV, determine the energy consumption requirements of the UAV when flying along the candidate connecting edges. The heading change requirement of the candidate connecting edge is compared with the maximum controllable turning angular velocity constraint in the set of remaining controllability constraints. When the heading change requirement exceeds the maximum controllable turning angular velocity constraint, the corresponding candidate connecting edge is marked as an uncontrollable connecting edge. The climbing or descending requirements of candidate connecting edges are compared with the maximum stable descent rate constraint in the set of remaining control capabilities constraints. When the climbing or descending requirements exceed the maximum stable descent rate constraint, the corresponding candidate connecting edge is marked as an uncontrollable connecting edge. The energy consumption demand of the candidate connection edge is compared with the available energy budget constraint in the set of remaining control capability constraints. When the energy consumption demand exceeds the available energy budget constraint, the corresponding candidate connection edge is marked as an uncontrollable connection edge. Based on the spatial path corresponding to the candidate connection edge, determine the minimum altitude trajectory of the candidate connection edge during flight, and check the minimum altitude value of the minimum altitude trajectory against the minimum safe altitude constraint. When the minimum altitude value is lower than the minimum safe altitude constraint, mark the corresponding candidate connection edge as an uncontrollable connection edge. The candidate connection edges marked as uncontrollable are removed from the candidate connection edge set, and the unmarked candidate connection edges are retained to form a controllable connection edge set. An initial controllable topology graph is constructed based on the candidate node set and the controllable connection edge set.

6. The method for planning an emergency landing path for a drone according to claim 1, characterized in that, The generation of the low-uncertainty topology graph specifically includes: Based on emergency landing status data and disturbance environment data, the location uncertainty entropy value, wind field uncertainty entropy value and control response uncertainty entropy value are determined and aligned to form a set of uncertainty entropy values; The set of uncertainty entropy values ​​includes location uncertainty entropy values, wind field uncertainty entropy values, and control response uncertainty entropy values; Normalize the set of uncertain entropy values ​​to generate normalized entropy values; Uncertainty entropy structure parameters are constructed based on normalized entropy values. These parameters are used to characterize the synergistic amplification effect and consistency deviation effect of multi-source uncertainties. The topological shrinkage radius is obtained by linearly mapping the structural parameters of the uncertainty entropy. Based on the uncertainty entropy structure parameter and the topology shrinkage radius, entropy-gated shrinkage processing is performed. When the uncertainty entropy structure parameter is greater than the preset entropy gating threshold, spatial domain shrinkage is performed on the risk suppression topology map with the geometric center of the target forced landing area as the center. Nodes and connecting edges whose spatial positions exceed the topology shrinkage radius are removed to generate a low-uncertainty topology map, while maintaining the spatial consistency and connection continuity of nodes in the low-uncertainty topology map.

7. The method for planning an emergency landing path for a drone according to claim 1, characterized in that, The generation of the verification topology subgraph specifically includes: In a low-uncertainty topology graph, security assertion checks are performed on the set of nodes and the set of connecting edges, including: When there is no continuous node sequence formed by connecting edges between the current location node of the UAV and the corresponding node of the candidate forced landing area in a low uncertainty topology graph, it is determined that the nodes and connecting edges constituting the low uncertainty topology graph do not satisfy the safety assertion check. When the spatial coordinates of any node fall within the spatial range defined by the boundary information of the occupied area in the spatial occupancy constraint data, it is determined that the node does not meet the security assertion check. When the spatial trajectory range value corresponding to any connecting edge overlaps with the boundary information of the occupied area in the spatial occupancy constraint data, the connecting edge is determined not to satisfy the security assertion check. When the spatial trajectory corresponding to any connecting edge exceeds the spatial domain range determined by the entropy gated contraction process, it is determined that the connecting edge does not satisfy the security assertion check. If the start or end node of any connecting edge is determined to not satisfy the security assertion check, then the connecting edge is determined to not satisfy the security assertion check. Nodes and edges that are determined not to meet the security assertion check are removed, while nodes and edges that are not determined not to meet the security assertion check are retained, and a check topology subgraph is generated.

8. The method for planning an emergency landing path for a drone according to claim 1, characterized in that, The generation of the emergency landing path planning results specifically includes: In the verification topology subgraph, the current position node of the UAV is taken as the starting node. Based on the connectivity relationship between the nodes formed by the connecting edges, the verification topology subgraph is subjected to connectivity traversal processing to extract the set of all nodes that are connected to the current position node of the UAV, forming a set of connectable regions that can be forced to land. In the set of connectable regions that can be forced to land, nodes corresponding to candidate forced landing areas are selected as the set of candidate nodes for forced landing terminals; For each node in the set of candidate nodes for forced landing terminals, the corresponding forced landing terminal domain is determined based on its spatial coordinate information and the relationship between the connecting edges. When multiple forced landing terminal domains exist simultaneously, based on the connection path length information from the current position node of the UAV to each forced landing terminal domain in the verification topology subgraph, the forced landing terminal domain with the smallest connection path length is selected as the target forced landing terminal domain. After the target forced landing terminal domain is determined, the emergency landing path planning result is generated based on the connection path from the current location node of the UAV to the target forced landing terminal domain in the verification topology subgraph.

9. A drone emergency landing path planning system, executing the drone emergency landing path planning method according to any one of claims 1 to 8, characterized in that, include: The emergency landing data acquisition module is used to collect emergency landing status data and environmental constraint data of the UAV; The control capability constraint generation module is used to generate the set of remaining control capability constraints; The 3D flight space construction module is used to construct a 3D flight space representation based on space occupancy constraint data, and generate a set of candidate nodes and a set of candidate connecting edges in the 3D flight space representation; The controllability determination module is used to perform controllability determination on the candidate connection edge set. Based on the remaining control capability constraint set, it performs consistency verification on the heading change requirements, climb or descent requirements and energy consumption requirements corresponding to the candidate connection edges, and generates an initial controllability topology map. The risk propagation modeling module is used to construct a risk propagation map based on disturbed environment data and calculate the risk propagation intensity. It maps the risk propagation intensity to the risk constraint parameters of the initial controllability topology map and performs topology pruning on the initial controllability topology map based on the risk constraint parameters to generate a risk suppression topology map. The uncertainty entropy modeling module is used to calculate the set of uncertainty entropy values ​​and perform entropy gating shrinkage on the risk suppression topology to generate a low-uncertainty topology. The security assertion verification module is used to perform security assertion verification on low-uncertainty topology graphs, remove nodes and connecting edges that do not meet the security assertion verification, and generate a verified topology subgraph. The forced landing path generation module is used to extract the set of connectable regions that can be forced landing based on the verified topology subgraph, determine the target forced landing terminal domain, and generate emergency forced landing path planning results. The generation of the risk suppression topology specifically includes: Based on the disturbance environment data, a set of disturbance-affected areas is determined, and the corresponding spatial location range is marked for each disturbance-affected area; Map the set of disturbance-affected regions to the initial controllability topology graph to determine the spatial relationship between each disturbance-affected region and the nodes and candidate connecting edges in the topology graph; Based on the topological connection relationships between nodes and connecting edges, the state of the disturbance-affected area is propagated step by step along the topological structure to generate a set of disturbance state at the node level and the connecting edge level. Calculate the risk propagation intensity of each node and each connecting edge based on the set of disturbance effect states; The risk propagation intensity is normalized to generate risk constraint parameters for the initial controllability topology graph, and the risk constraint parameters are bound to the corresponding nodes and connecting edges respectively. Based on the risk constraint parameters, the initial controllable topology graph is pruned to remove nodes and connecting edges whose risk constraint parameters are greater than or equal to the preset risk tolerance threshold, thereby generating a risk-suppressed topology graph.

Images