Plateau multi-uav static path planning and dynamic rescheduling method and system

Abstract

This invention relates to the field of UAV collaborative patrol and path planning technology, and discloses a method and system for static trajectory planning and dynamic rescheduling of multiple UAVs in high-altitude areas. The method receives patrol mission instructions, acquires a set of patrol points, digital elevation model data, and environmental data such as air pressure and temperature, determines a comprehensive environmental correction factor, and forms corrected actual performance parameters. Based on these parameters, it performs terrain-aware static trajectory planning on candidate flight paths, forming a static trajectory scheme through comparison of steep ascent, descent, and turn modes. During patrol execution, it detects real-time environmental changes, dynamic no-fly zones, UAV malfunctions, and changes in communication coverage, making corresponding adjustments to affected trajectories, remaining tasks, and relay networks. A post-hoc check determines whether to continue patrolling or continue dynamic relay network adjustments. This invention is applicable to multi-UAV collaborative patrol missions in complex high-altitude terrain scenarios.

Description

Technical Field

[0001] This invention relates to the field of UAV collaborative patrol and path planning technology, and in particular to a method and system for static trajectory planning and dynamic rescheduling of multiple UAVs in high-altitude areas. Background Technology

[0002] Patrol missions in high-altitude areas are typically characterized by large patrol areas, long flight ranges, and extended durations, often traversing complex terrains such as mountains, canyons, and slopes. In these application scenarios, patrol missions not only involve patrol drones providing continuous coverage along a predetermined area, but also require task allocation among multiple drones, flight path organization, and maintaining communication coverage throughout the patrol process.

[0003] As patrol missions continue, high-altitude environmental conditions such as air pressure and temperature, as well as the status of communication links and drones, may constantly change, resulting in significant uncertainties in operational conditions. Existing multi-drone collaborative patrol technologies often struggle to simultaneously account for the ongoing impact of subsequent dynamic events on flight paths and mission organization when planning routes in complex high-altitude terrain. Furthermore, during patrol operations, existing methods often fail to maintain a continuous correspondence with the initial route planning results when faced with environmental changes, changes in no-fly zones, drone malfunctions, or changes in communication coverage.

[0004] Therefore, in the process of multi-UAV collaborative patrol in plateau scenarios, existing technologies generally suffer from problems such as insufficient connection between static trajectory planning and dynamic rescheduling, unstable mission succession, asynchronous trajectory adjustment and mission organization, and difficulty in maintaining communication coverage, which limit the continuous organization and stable execution of patrol missions. Summary of the Invention

[0005] To address the aforementioned issues, this invention proposes a method and system for static trajectory planning and dynamic rescheduling of multi-UAVs in high-altitude environments. This method, focusing on multi-UAV collaborative patrol missions under complex terrain and uncertain operating conditions in high-altitude areas, constructs a unified operational link consisting of environmental modeling and performance parameter correction, terrain-aware static trajectory planning, dynamic event-driven rescheduling processing, and communication coverage post-hoc checks. This ensures that the static trajectory planning results are continuously integrated with environmental changes, fault handling, mission continuity, and communication support processes during patrol execution, thereby improving the stability, continuity, and adaptability of multi-UAV collaborative patrol missions in high-altitude scenarios.

[0006] To achieve the above objectives, a first aspect of the present invention provides a method for static trajectory planning and dynamic rescheduling of multiple unmanned aerial vehicles (UAVs) in high-altitude areas, comprising:

[0007] Receive patrol mission instructions and acquire patrol point sets, digital elevation model data, and air pressure and temperature data corresponding to the plateau environment;

[0008] A comprehensive environmental correction factor is determined based on the air pressure data and the temperature data, and the corrected actual performance parameters are determined based on the comprehensive environmental correction factor.

[0009] Based on the digital elevation model data and the corrected actual performance parameters, terrain-aware static trajectory planning is performed on the candidate flight paths between the patrol points. Specifically, by comparing the candidate paths in steep ascent, steep descent, and steep turn modes, a static trajectory scheme that meets terrain safety constraints and is assigned to each UAV is formed.

[0010] The patrol mission is carried out according to the static flight path scheme, and environmental data, UAV status data and communication link status data are collected. Based on the environmental data, the UAV status data and the communication link status data, dynamic events are detected and the event types are determined.

[0011] Based on the event type, corresponding adjustments are made to the relevant items in the affected remaining flight paths, remaining patrol missions, and communication coverage status. These adjustments include at least: when the event type is an environmental change or a dynamic no-fly zone, replanning the remaining flight paths of the affected UAVs; when the event type is a UAV malfunction, generating a safe landing path for the malfunctioning UAV and reallocating its remaining patrol missions to other available UAVs; and when the event type is a relay UAV failure, or when communication coverage status is determined to be declining based on the communication link status data, dynamic adjustments to the relay network are performed.

[0012] A post-hoc check is performed on the corresponding adjusted communication coverage status, and the result of the post-hoc check determines whether to continue the patrol mission or continue the dynamic adjustment of the relay network.

[0013] A second aspect of the present invention provides a high-altitude multi-UAV static trajectory planning and dynamic rescheduling system, comprising an onboard processor, a memory, sensor components, and a communication component. The onboard processor is connected to the memory, the sensor components, and the communication component, respectively. The memory stores program instructions executable by the onboard processor, which, when executed, cause the onboard processor to call:

[0014] The plateau environment modeling and UAV performance correction module is used to determine a comprehensive environmental correction factor based on the air pressure and temperature data collected by the sensor components, and to determine the corrected actual performance parameters based on the comprehensive environmental correction factor.

[0015] The terrain-aware static trajectory planning module is used to receive the set of patrol points and digital elevation model data in the patrol mission instruction received by the communication component, and to perform terrain-aware static trajectory planning on the candidate flight paths between the patrol points based on the corrected actual performance parameters. In this module, by comparing the candidate paths in steep ascent mode, steep descent mode and steep turn mode, a static trajectory scheme that meets the terrain safety constraints and is assigned to each UAV is formed.

[0016] The sensor real-time monitoring module is used to receive environmental data and UAV status data collected by the sensor components and communication link status data obtained by the communication components, and to detect dynamic events and determine the event type based on the environmental data, the UAV status data and the communication link status data.

[0017] The dynamic rescheduling module is used to perform corresponding adjustments to the affected remaining flight paths, remaining patrol tasks, or communication coverage status according to the event type, and to perform a post-hoc check on the adjusted communication coverage status to determine whether to continue the patrol task or continue the relay network dynamic adjustment based on the post-hoc check result. The corresponding adjustments include at least: when the event type is an environmental change or a dynamic no-fly zone, replanning the remaining flight paths of the affected UAVs; when the event type is a UAV malfunction, generating a safe landing path for the malfunctioning UAV and reallocating the remaining patrol tasks of the malfunctioning UAV to other available UAVs; and when the event type is a relay UAV failure, or when the communication coverage status is determined to have decreased based on the communication link status data, performing relay network dynamic adjustment.

[0018] Compared with existing technologies, the present invention provides a method and system for static trajectory planning and dynamic rescheduling of multiple UAVs in high-altitude areas, which has the following technical advantages:

[0019] By incorporating environmental modeling and performance parameter correction, static trajectory planning, dynamic event handling, and post-hoc verification of communication coverage status into a unified operational process, multi-UAV collaborative patrol missions can maintain a continuous correspondence between static trajectory planning and dynamic rescheduling under complex terrain and uncertain operational conditions in high-altitude areas. By adjusting affected trajectories, mission organization relationships, and communication coverage status during the continuous execution of patrol missions, a consistent operational relationship can be maintained between patrol missions, trajectory changes, and communication support. Through the above overall operational approach, the mission succession relationship, trajectory adjustment relationship, and communication coverage relationship in multi-UAV collaborative patrols can be continuously maintained within the same mission cycle, thereby maintaining the stability and continuity of collaborative operations and providing corresponding support for the continuous organization and stable execution of patrol missions in high-altitude scenarios. Attached Figure Description

[0020] Figure 1This is a flowchart illustrating the overall processing of a method for static trajectory planning and dynamic rescheduling of multiple UAVs in high-altitude areas, provided in an embodiment of the present invention.

[0021] Figure 2 This invention provides a diagram illustrating the transmission relationship between key processing objects and post-hoc inspection results in each processing stage, as provided in an embodiment of the invention.

[0022] Figure 3 This is a schematic diagram of single-segment flight path formation and three-mode automatic comparison provided by an embodiment of the present invention;

[0023] Figure 4 This is a schematic diagram of dynamic event classification processing provided in an embodiment of the present invention;

[0024] Figure 5 This invention provides a schematic diagram of dynamic adjustment and post-hoc check backflow in a relay network.

[0025] Figure 6 This invention provides a structural block diagram of a plateau multi-UAV static trajectory planning and dynamic rescheduling system.

[0026] Figure label:

[0027] 31 - Straight Path Safety Determination; 32 - Steep Ascent Mode; 33 - Steep Descent Mode; 34 - Steep Turn Mode; 35 - Automatic Comparison; 36 - Path Safety Correction;

[0028] 41 - Event classification selection based on multi-source state data; 42 - Handling sudden environmental changes or dynamic no-fly zones; 43 - UAV failure handling; 44 - Safe landing path generation; 45 - Remaining mission continuation handling; 46 - Handling relay UAV failure or communication coverage degradation.

[0029] 51 – Prediction Layer; 52 – Response Layer; 53 – Relay Location Update; 54 – Adjustment Method Selection; 55 – Posterior Check; 56 – Communication Coverage Satisfaction Judgment;

[0030] 61-Airborne processor; 62-Memory; 63-Sensor components; 64-Communication components; 65-High-altitude environment modeling and UAV performance correction module; 66-Terrain perception static trajectory planning module; 67-Real-time sensor monitoring module; 68-Dynamic rescheduling module. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0032] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0033] Example 1

[0034] In one embodiment, the patrol mission is deployed in a complex plateau terrain area, including mountains, valleys, and slopes. Multiple patrol points are set up within the area, and multiple patrol drones coordinate to execute the patrol mission. A ground station issues patrol mission commands and receives patrol data, while a relay drone establishes a communication link between the patrol drones and the ground station. During patrol execution, environmental conditions, flight status, and communication link status continuously change. Therefore, this embodiment does not continuously execute along a predetermined route after the initial patrol track is formed. Instead, it sequentially completes environmental modeling and performance parameter correction, terrain-aware static track planning, patrol execution and dynamic event detection, dynamic event-driven rescheduling processing, and post-hoc checks of communication coverage status, based on the same set of task inputs. This ensures continuous connection between the static track planning results, the dynamic adjustment results during execution, and the communication support results. In existing materials, Figure 1 Used to express the overall processing flow described above. Figure 2 This is used to express the transitive relationship between key processing objects in each processing stage; this embodiment uses [example example]. Figure 1 and Figure 2 The main chain shown is used as the basis for the explanation.

[0035] See Figure 1 , Figure 1The overall processing flow of this embodiment is shown. The flow includes at least the following steps: Step S100, receiving mission instructions and acquiring basic data to obtain patrol mission instructions, patrol point sets, digital elevation model data, and air pressure and temperature data corresponding to the plateau environment; Step S200, performing environmental modeling and performance parameter correction to determine a comprehensive environmental correction factor based on the air pressure and temperature data and form corrected actual performance parameters; Step S300, performing terrain-aware static trajectory planning to form a static trajectory scheme based on the digital elevation model data and corrected actual performance parameters; Step S400, performing patrol, status monitoring, and event detection based on environmental data and unmanned aerial vehicle (UAV) systems. The system detects dynamic events and determines the event type based on the aircraft status data and communication link status data; in step S500, when the event type is a sudden environmental change or a dynamic no-fly zone, it performs replanning on the affected remaining flight paths; in step S600, when the event type is a UAV malfunction, it generates a safe landing path and performs continuation processing on the remaining patrol missions; in step S700, when the event type is a relay UAV failure or a decrease in communication coverage, it performs dynamic adjustment of the relay network, and after the processing of various dynamic events is completed, it performs a post-hoc check on the communication coverage status to determine whether to continue the patrol mission or continue the dynamic adjustment of the relay network based on the post-hoc check results.

[0036] See Figure 2 , Figure 2 This illustrates the continuous transfer relationship of key processing objects between processing stages in this embodiment. These key processing objects include at least: patrol mission instructions, patrol point set, digital elevation model data, integrated environmental correction factor, corrected actual performance parameters, single-segment flight path, static trajectory plan, event type, updated remaining trajectory, safe landing path, remaining mission continuation results, target location, location adjustment amount, communication coverage status, and post-hoc inspection results. These processing objects are not isolated from each other but flow continuously between the static trajectory planning stage, dynamic event processing stage, and communication support stage, ultimately converging at the post-hoc inspection node, thus forming... Figure 1 and Figure 2 A closed-loop processing link with common characteristics, wherein the posterior check node outputs the posterior check result to determine whether to continue the patrol mission or continue the relay network dynamic adjustment.

[0037] In one embodiment, the following example parameter set can be used to illustrate this embodiment: current flight altitude is 4500m, current ambient temperature is (-35℃), wind speed is 8m / s, and meteorological wind direction is 0°; the patrol drone is a fixed-wing aircraft, with a basic cruise speed of 25m / s, a basic rate of climb of 4m / s, a basic rate of descent of 3m / s, a nominal range of 80km, and a range safety margin of 10km; the normal cruise safety margin is 200m, and the landing safety margin is 50m. The above parameter set is only an example parameter set used to illustrate the relationship between environmental modeling and performance parameter correction, static trajectory planning and dynamic rescheduling, and does not constitute a limitation on the scope of protection of this invention.

[0038] In step S100, a patrol mission instruction issued by the ground station is received, and a patrol point set, digital elevation model data, and air pressure and temperature data corresponding to the plateau environment are acquired. The patrol mission instruction may include at least the patrol point set and base point information; the digital elevation model data is used to characterize the terrain undulations of the patrol area; and the air pressure and temperature data are used for subsequent flight performance correction. Thus, the patrol mission instruction, patrol point set, digital elevation model data, air pressure data, and temperature data together constitute the initial input for subsequent environmental modeling and performance parameter correction, static trajectory planning, and dynamic rescheduling during the execution period.

[0039] In step S200, a pressure correction factor is first determined based on the current flight altitude, then a temperature correction factor is determined based on the current ambient temperature, and a comprehensive environmental correction factor is determined from the pressure correction factor and the temperature correction factor. In one embodiment, the pressure correction factor k... p It can be written as:

[0040] ;

[0041] Where H represents the current flight altitude; H0 represents the reference altitude; λ H This indicates the high attenuation coefficient.

[0042] The temperature correction factor can be written as:

[0043] ;

[0044] Where T represents the current ambient temperature; T0 represents the reference temperature; λ T This represents the temperature decay coefficient.

[0045] The comprehensive environmental correction factor can then be expressed as:

[0046] K e =k p ·k t ;

[0047] Under the example parameters described above, a corresponding integrated environmental correction factor can be obtained, thereby generating the corrected actual cruise speed, corrected actual rate of climb, and corrected actual rate of descent. Specifically, the base cruise speed, base rate of climb, and base rate of descent are multiplied by the integrated environmental correction factor to form the corrected actual performance parameters.

[0048] In this embodiment, the corrected actual performance parameters are not only used for forming single-segment flight paths between patrol points, but also serve as a unified capability metric in the dynamic rescheduling phase of the execution period, participating in the remaining trajectory assessment, candidate landing point search range determination, safe landing path generation, and time scale estimation in relay network dynamic adjustment. Furthermore, the effective ground speed can be determined based on the projected components of the corrected actual cruise speed and wind speed in the flight direction, ensuring consistent calculation methods for flight capability characterization between the static planning phase and the dynamic handling phase of the execution period. The effective ground speed υ g It can be represented as:

[0049] υ g =max(1,υ+υ w );

[0050] Among them, υ g Indicates the effective ground speed; υ represents the corrected actual cruise speed; υ w This represents the projected component of wind speed along the flight direction. By unifying the effective ground speed for single-segment flight path time estimation, remaining range assessment, and dynamic rescheduling during the execution period, the static trajectory planning results and dynamic rescheduling results can be kept consistent on the flight capability scale.

[0051] In step S300, based on digital elevation model data and corrected actual performance parameters, terrain-aware static trajectory planning is performed on candidate flight paths between patrol points. Specifically, path sampling points are extracted along the line connecting any two patrol points, and the terrain height corresponding to each sampling point is obtained. First, a safety judgment is performed on straight paths. When a straight path meets the safety conditions, it is directly used as a single-segment flight path. When a straight path does not meet the safety conditions, paths corresponding to steep ascent, steep descent, and steep turn modes are formed respectively, and mode evaluation results are generated for the three modes. Based on the mode evaluation results, a comparison is performed to determine the optimal mode corresponding path. After obtaining the optimal mode corresponding path, path safety correction is performed on the optimal mode corresponding path to ensure that the entire path meets the terrain safety constraints. The single-segment flight paths formed between each patrol point are further spliced ​​to form a static trajectory scheme allocated to each UAV.

[0052] In one optional embodiment, the model evaluation result includes at least a flight time term, and in another embodiment, it further includes a path cost term, so that subsequent comparisons simultaneously reflect flight efficiency and path cost. In a further embodiment, the model comprehensive evaluation quantity can be expressed as:

[0053] J(Γ m )=λ t T m +λ c C m ;

[0054] Among them, J(Γ) m ) represents the pattern Γ m The overall evaluation quantity; T m Indicates the mode flight time term; C m Represents the path cost term; λ t and λ c These represent the weights of the time term and the cost term, respectively; the path corresponding to the optimal mode can be determined according to the principle of minimizing the comprehensive evaluation quantity.

[0055] In another alternative implementation, candidate path modes can be screened first based on terrain safety constraints, eliminating those that do not meet the safety conditions. Then, the remaining candidate path modes can be compared according to flight time and path cost, so as to achieve hierarchical selection that first meets safety constraints and then optimizes efficiency.

[0056] After comparing the three modes, a path safety correction is performed on the path corresponding to the optimal mode. If the flight altitude of any path point is lower than the sum of the terrain altitude and the safety margin, the flight altitude of that point is raised to a safe altitude, ensuring that the entire path meets terrain safety constraints. The corrected path corresponding to the optimal mode serves as the single-segment flight path between patrol points, and together with the single-segment flight paths corresponding to other flight segments, constitutes the static trajectory scheme. The static trajectory scheme serves as the initial scheme for subsequent patrol execution and dynamic rescheduling during the execution period.

[0057] In step S400, after forming a static flight path plan, each patrol drone initiates patrol execution according to its corresponding static flight path plan. During execution, the relay drone is located at a predetermined position capable of covering the patrol area and can establish a communication link between the patrol drone and the ground station. The patrol equipment carried by the patrol drone continuously monitors the target area and transmits mission data to the ground station via the relay link. Simultaneously, the sensor real-time monitoring process continuously collects environmental data, drone status data, and communication link status data. The environmental data includes at least real-time air pressure, temperature, wind speed, and wind direction; the drone status data includes at least remaining battery power, flight speed, attitude, and positioning status; and the communication link status data includes at least link quality and coverage status. Based on the above monitoring data, dynamic events are detected, and the event type is determined. The event types include at least sudden environmental changes, dynamic no-fly zones, drone malfunctions, and relay drone failures or decreased communication coverage.

[0058] In one embodiment, the detection of dynamic events is not based solely on a single state quantity, but rather on a joint determination of the event type based on environmental data, UAV state data, and communication link state data. The environmental data includes at least real-time air pressure, temperature, wind speed, and wind direction; the UAV state data includes at least remaining battery power, flight speed, attitude information, and location information; and the communication link state data includes at least link quality and coverage status. By performing fusion determination on the above multiple data types, execution-period state changes can be converted into corresponding event types such as environmental abrupt changes, dynamic no-fly zones, UAV malfunctions, and relay UAV failures or decreased communication coverage, thereby providing a unified determination entry point for subsequent branch processing.

[0059] In step S500, when the event type is an environmental change or a dynamic no-fly zone, the remaining flight paths of the affected UAVs are replanned. Specifically, the effective ground speed of each segment of the affected UAV is recalculated first, and then the relationship between the total distance of the remaining flight path and the remaining range is evaluated based on the recalculation results. If the evaluation result shows that the total distance of the remaining flight path is greater than the remaining range, it is determined that the affected remaining flight segments do not meet the requirements for continued execution, and local replanning is performed on the affected remaining flight segments. If the evaluation result shows that the total distance of the remaining flight path is not greater than the remaining range, the structure of the remaining flight path remains unchanged, and only the relevant speed parameters are updated. In the case of a sudden change in wind direction, the detour direction of the affected flight segments is updated synchronously, and the candidate path corresponding to the affected flight segments is re-formed accordingly. After the local update or parameter update is completed, the updated remaining flight path re-enters the patrol execution process.

[0060] In a further embodiment, the triggering of replanning can be determined based on changes in environmental data and / or changes in dynamic no-fly zones. When replanning is triggered, the effective ground speed of each segment of the affected UAV is recalculated, and the relationship between the remaining total track distance and the remaining range is evaluated based on the recalculation results. If the evaluation results indicate that the remaining total track distance is greater than the remaining range, local replanning is performed on the affected remaining segments. During local replanning, local path segments can be extracted according to a preset window size, and disturbance generation and fitness evaluation are performed on candidate paths within the window. The local paths are updated based on the evaluation results. In an optional embodiment, the fitness evaluation simultaneously considers the remaining path flight time, the remaining path distance, and the no-fly zone collision penalty, so that local replanning balances range, safety, and local optimization efficiency during execution.

[0061] In one optional implementation, the local replanning process can be further refined using a rolling window search and fitness evaluation method. In a further implementation, the local replanning process can be adjusted by combining window size, local path perturbation, and stopping conditions. In an even further implementation, a fitness function that simultaneously considers the remaining path flight time, remaining path distance, and no-fly zone collision penalty term can be used to evaluate the local update process, and the switching from global exploration to local convergence can be achieved by decreasing the contraction and expansion coefficients during the iteration process. The above further implementations are used to illustrate the optional refinement implementation of local replanning, without changing the main processing chain of "recalculating effective ground velocity—evaluating the relationship between remaining trajectory and remaining endurance—performing local replanning when necessary" in this embodiment.

[0062] In step S600, when the event type is UAV malfunction, the malfunctioning UAV is first identified, and then the process of safe landing and continuation of remaining tasks begins. For the malfunctioning UAV, the search range for candidate landing points is determined first based on the current flight status. A set of candidate landing points is generated within the search range, and each candidate point in the set is scored to determine the optimal landing point. After determining the optimal landing point, a safe landing path to the optimal landing point is generated. After the malfunctioning UAV exits the patrol mission, continuation processing is performed on its unfinished tasks to maintain the continuous execution of the patrol mission after the malfunction is resolved.

[0063] In one optional implementation, the triggering of a safe landing procedure can be further determined based on the severity of the fault. Specifically, a fault severity index can be generated based on the fault type and the degree of anomaly; when the fault severity index exceeds a preset threshold, the safe landing procedure is triggered; when the fault severity index does not exceed the preset threshold, a degraded patrol or continuous monitoring state can be entered. By introducing fault severity determination, all faults can be prevented from triggering a forced landing procedure indiscriminately, and the timeliness of handling high-risk faults can be improved.

[0064] In a further embodiment, the candidate landing point search range can be determined by combining the remaining battery percentage, full-charge range, current effective ground speed, and safety margin coefficient, so that the candidate landing point search range is consistent with the current execution capability, rather than being a rough estimate based solely on static remaining battery power. After determining the candidate landing point search range, a set of candidate landing points can be generated within this range, and the distance component, terrain slope flatness component, and terrain height variance component of each candidate landing point can be comprehensively scored to determine the optimal landing point.

[0065] In an alternative implementation, the decision to trigger a safe landing procedure can be further determined based on the severity of the fault. Specifically, the fault severity index S f It can be represented as:

[0066] S f =w f ·η b +w a ·δ a ;

[0067] in,

[0068] w f Indicates the weighting coefficient for the fault type;

[0069] w a Indicates the weighting coefficient for the degree of abnormality;

[0070] η b Indicates the current remaining battery percentage;

[0071] δ a This represents the normalized outlier deviation.

[0072] In one embodiment, the fault type weights corresponding to power failure, sensor failure, communication failure, and structural failure can be set to 1.0, 0.7, 0.5, and 0.9, respectively. When the fault severity index exceeds a preset threshold, a safe landing procedure is triggered; when the fault severity index does not exceed the preset threshold, a degraded patrol or continuous monitoring state can be entered.

[0073] In one optional implementation, the search range for candidate landing points can be determined by combining the remaining battery power percentage, full-charge range, current effective ground speed, and safety margin coefficient. In one implementation, the search radius can be expressed as:

[0074] ;

[0075] Among them, R S Represents the search radius; γ represents the safety margin factor; R full This indicates the driving range on a full charge; υ0 indicates the base cruising speed.

[0076] In one embodiment, the safety margin coefficient can be set to 0.5 to reserve a safety margin for landing maneuvers. After determining the search range, a set of candidate landing points can be generated according to a preset grid step size. In one embodiment, the grid step size can be set to 200 meters.

[0077] In a further embodiment, distance components, terrain slope flatness components, and terrain height variance components can be calculated for each candidate landing point, and a comprehensive score can be formed by pre-setting weights, so as to determine the optimal landing point based on the comprehensive score. The comprehensive score can be expressed as:

[0078] ;

[0079] Among them, Score i This represents the overall score of the i-th candidate point; Represents the distance component; Indicates the slope flatness component; Represents the variance component of terrain height; (λ) d ,λ s ,λ v The weights of each component are represented by ). The optimal landing point can be further determined according to the principle of maximizing the comprehensive score. After obtaining the optimal landing point, a controlled descent method is preferentially used to form the safe landing path, and terrain safety constraints, descent rate constraints, and landing endpoint constraints are imposed on the path.

[0080] Continuing in step S600, after the malfunctioning UAV exits its patrol mission, its unfinished tasks are processed for continuation. In one embodiment, the unfinished tasks of the malfunctioning UAV are reallocated using the task allocation matrix as a decision variable; during the reallocation process, endurance constraints, full task coverage constraints, and no-fly zone constraints are applied; the unfinished tasks are assigned to other available UAVs, and the other available UAVs are controlled to continue performing patrol missions according to the reallocation results; when only one available UAV remains, a downgraded allocation is performed according to the minimum incremental distance; tasks that cannot be allocated in time are recorded as pending continuation tasks and are re-included in the allocation process after the availability of available UAVs recovers.

[0081] In a further embodiment, candidate mission succession schemes can be compared simultaneously using total flight distance, maximum mission completion time, and load balancing as optimization objectives. Candidate schemes that do not meet the constraints can be eliminated or sorted in descending order based on the degree of constraint violation. Candidate schemes that meet the constraints can be further filtered based on a comprehensive score to determine the optimal mission succession scheme. Thus, fault identification, safe landing, mission succession, and mission recovery can be completed continuously within the same execution chain.

[0082] In step S700, when the event type is relay drone failure, or when communication coverage is determined to be declining based on communication link status data, dynamic adjustment of the relay network is performed. Specifically, the prediction layer first determines the target location of the relay drone based on forward path information; then the reaction layer generates a position adjustment amount based on the relative position between the patrol drone and the relay drone; and then updates the relay drone's position based on the position adjustment amount. Afterwards, depending on the degree of communication coverage decline or the relay drone failure status, a relay network adjustment method is selected between maintaining the original location, local adjustment, and global reconstruction, while maintaining a coverage continuity constraint that the communication coverage rate is not lower than a preset communication coverage rate threshold during the adjustment process.

[0083] In one optional implementation, the forward path information includes at least a sequence of positions formed by several future path points of the patrol drone. The prediction layer determines the target position of the relay drone based on the position sequence. The reaction layer generates a position adjustment amount based on the deviation between the target position and the current position of the relay drone, as well as the current position of the patrol drone and the current communication coverage status. In a further implementation, the relay network adjustment method can be selected from maintaining the original position, local adjustment, and global reconstruction based on the decline in communication coverage status and the failure status of the relay drone. Specifically, when the decline in communication coverage status is lower than a first threshold, the original position is maintained; when the decline in communication coverage status is between the first and second thresholds, a local adjustment is performed; and when the relay drone fails or the decline in communication coverage status exceeds the second threshold, a global reconstruction is performed. The above further implementations illustrate optional refined implementations of the adjustment method selection.

[0084] In one embodiment, the preset communication coverage threshold can be set to 0.95. After the relay network dynamic adjustment is completed, a posterior check is performed on the adjusted communication coverage status. The posterior check includes at least comparing the adjusted communication coverage with the preset communication coverage threshold; if the posterior check result indicates that the communication coverage status does not meet the communication coverage requirements, the relay network dynamic adjustment continues; if the posterior check result indicates that the communication coverage status meets the communication coverage requirements, the patrol task continues. In a further embodiment, in addition to comparing the communication coverage threshold, the posterior check can also be combined with a link quality threshold or a communication interruption duration threshold for comprehensive judgment, so as to improve the robustness of communication assurance judgment.

[0085] In one implementation, the relay network adjustment method is not fixed, but rather selected in stages between maintaining the original location, partial adjustment, and global reconstruction, depending on the degree of communication coverage degradation and / or relay drone failure. When the communication coverage degradation is only slight and the existing relay network still meets the coverage continuity constraint, the original location can be maintained or partial adjustments can be performed; when the communication coverage degradation further worsens, or relay drone failure occurs, global reconstruction can be performed. After each type of adjustment is completed, a post-hoc check is performed on the communication coverage status; if the post-hoc check result indicates that the communication coverage status still does not meet the communication coverage requirements, dynamic adjustment of the relay network continues.

[0086] In one implementation, after completing the corresponding processes in steps S500, S600, and S700, the updated remaining flight paths, safe landing paths, remaining mission continuation results, position adjustments, and communication coverage status are all uniformly entered into the same inspection node. The comprehensive environmental correction factor and its resulting corrected actual performance parameters serve as a unified capability scale throughout the static flight path planning and dynamic rescheduling processes. Although sudden environmental changes, dynamic no-fly zones, UAV malfunctions, and relay network status changes trigger different dynamic processing branches, the processing results of each branch are ultimately uniformly entered into the communication coverage status inspection node. The subsequent inspection results determine whether to continue the patrol mission or continue the relay network dynamic adjustment. Therefore, the updated results involving flight paths, missions, and communications during patrol execution are uniformly verified within the same processing chain, rather than being separated between different processing stages. This ensures that multi-UAV static flight path planning, dynamic event handling, and communication support in complex plateau terrain scenarios form a continuous connection within the same mission cycle.

[0087] Example 2

[0088] Based on the overall processing flow described in Example 1, this example further elaborates on the rules for forming single-segment flight paths in terrain-aware static trajectory planning, the rules for classifying and processing dynamic events during patrol execution, and the rules for dynamic adjustment and post-hoc check feedback of the relay network. In other words, Example 1 focuses on getting the main closed loop—environmental modeling and performance parameter correction, static trajectory planning, dynamic event-driven rescheduling, and post-hoc check of communication coverage status—run smoothly; while this example further explains how several key nodes within this main closed loop are specifically executed. To avoid repetition, this example will not repeat the task input, the formation of the comprehensive environmental correction factor, the acquisition of the corrected actual performance parameters, and the overall patrol execution framework, but will focus on... Figure 3 , Figure 4 and Figure 5 Expand.

[0089] See Figure 3 , Figure 3 This illustrates the process of forming a single-segment flight path and automatically comparing three modes. For any two candidate flight paths between patrol points, a straight line is not directly used as the execution path. Instead, a straight-line path safety determination 31 is performed first. If the determination result shows that the straight-line path meets the safety conditions, it can be directly used as a single-segment flight path. If the determination result shows that the straight-line path does not meet the safety conditions, paths corresponding to steep ascent mode 32, steep descent mode 33, and steep turn mode 34 are formed respectively, and mode evaluation results are generated for these three modes. Subsequently, based on the mode evaluation results, automatic comparison 35 is performed to determine the path corresponding to the optimal mode. After obtaining the path corresponding to the optimal mode, path safety correction 36 is performed to form a single-segment flight path that meets terrain safety constraints. Thus, Figure 3 The processing procedure shown has a clear sequential relationship: first, it is determined whether the straight path can be passed directly; then, an alternative path is constructed; then, mode evaluation is performed; and finally, the optimal output and safety correction are completed, rather than mechanically switching between the three modes.

[0090] In one embodiment, the model evaluation result includes at least a flight time term, reflecting the flight efficiency of the corresponding model under the current corrected actual performance parameters and current terrain conditions. In a further embodiment, the model evaluation result also includes a path cost term, reflecting path length, detour cost, climb / descent energy consumption impact, or terrain adaptability cost. To avoid simplifying model selection to a single shortest time criterion, this embodiment preferably uses a comprehensive evaluation quantity to compare the three models. In one embodiment, the comprehensive model evaluation quantity can be expressed as:

[0091] J(Γ m )=λ t T m +λ c C m ;

[0092] Among them, J(Γ) m ) represents the pattern Γ m The overall evaluation quantity; T m Indicates the flight time item for the corresponding mode; C m λ represents the path cost term for the corresponding pattern. t and λ c These represent the weight coefficients of the time term and the cost term, respectively. Therefore, the path corresponding to the optimal mode can be determined by the principle of minimizing the overall evaluation quantity:

[0093] ;

[0094] Among them, Γ up ,Γ down and Γ turn These represent the paths corresponding to steep ascent, steep descent, and steep turn modes, respectively.

[0095] This comprehensive comparison method avoids simply pursuing the shortest time while ignoring path costs, and also avoids simply reducing costs, which could lead to a significant decrease in flight efficiency.

[0096] In another embodiment, the selection can also employ a hierarchical rule of "safety screening first, efficiency optimization later." Specifically, based on terrain safety constraints, minimum safe altitude constraints, and path passability, the paths corresponding to the three modes are initially screened, eliminating candidate mode paths that do not meet the safety conditions. Then, only among the remaining candidate mode paths that meet the safety conditions are compared according to flight time and path cost to determine the optimal mode path. This hierarchical selection method is particularly suitable for scenarios with significant undulations in complex plateau terrain, severe obstruction of straight paths, or large differences in detour costs. By performing efficiency optimization after safety screening, a single flight path can be both feasible and maintain good overall efficiency.

[0097] For steep ascent mode 32, in one embodiment, the terrain slope gradient of the path sampling point sequence is first calculated, and a pre-climb is triggered based on the maximum gradient within the look-ahead window. When a pre-climb is triggered, the target pre-climb height, optimal climb angle, shortest pre-climb initiation distance, gradient urgency factor, and final pre-climb initiation distance are further determined, and a pre-climb height sequence is formed accordingly to generate the path corresponding to the steep ascent mode. This processing method is suitable for scenarios where the terrain rises rapidly ahead, and a straight path would result in insufficient safe height. By initiating the pre-climb before the path enters the terrain rise section, sudden large-slope climbs near obstacles can be reduced, improving the stability of path execution.

[0098] For steep descent mode 33, in one embodiment, the terrain gradient along the flight direction at each sampling point on the path is first calculated. Then, a target ground-hugging altitude sequence is formed based on the terrain gradient. Combining the descent energy consumption cost function with the segmented results of gentle slope, medium slope, and steep slope sections, the optimal descent altitude sequence is determined to generate the path corresponding to the steep descent mode. This processing method is suitable for scenarios where the terrain gradually descends ahead, and the descent trend can be used to shorten the path time or reduce overall energy consumption. By segmenting the descent path, more suitable descent strategies can be adopted for different slope sections, thereby achieving a balance between safe ground-hugging flight and energy consumption control.

[0099] For steep turn mode 34, in one embodiment, the projected width of the obstacle to be bypassed in the path normal vector direction is first extracted, and then the basic detour radius is determined accordingly. Subsequently, a wind direction correction coefficient is generated based on the wind direction and detour direction, thereby obtaining the dynamic detour radius. Then, the detour costs on the left and right sides are calculated separately, and the detour direction and its corresponding dynamic detour radius are determined according to the principle of lower cost, so as to generate the path corresponding to the steep turn mode. In a further embodiment, the detour costs on the left and right sides can be expressed as follows:

[0100] J L =αL L +βT L J R =αL R +βT R ;

[0101] Among them, J L and J R These represent the combined costs of leftward and rightward wrapping, respectively; L L and L R These represent the path lengths for left and right wrapping, respectively; T L and T R α and β represent the time terms for left and right turns, respectively; α and β represent the weights of the length and time terms, respectively. By incorporating the wind direction correction coefficient into the detour radius determination process, the steep turn mode can consider not only geometric obstacle avoidance relationships but also the actual impact of wind on detour costs in plateau scenarios.

[0102] After obtaining the path corresponding to the optimal mode, path safety correction is performed on the path. Specifically, when the flight altitude of any path point is lower than the sum of the terrain altitude and the safety margin at that point, the flight altitude at that point is raised to a safe altitude to ensure that the path meets the terrain safety constraints throughout. Thus, Figure 3The three-mode comparison process shown does not replace safety conditions with mode evaluation results. Instead, after the mode evaluation results determine the optimal mode's corresponding path, path safety correction is used to ensure the final output path is executable. The corrected optimal mode's corresponding path serves as a single-segment flight path between patrol points, which is further spliced ​​together to form a static track scheme. Through this processing chain, the single-segment flight path not only reflects terrain undulations and environmental constraints but also maintains consistency with the corrected actual performance parameters, forming the basis track for subsequent dynamic event processing.

[0103] See Figure 4 , Figure 4 The dynamic event classification and processing process is illustrated. During patrol execution, environmental data, UAV status data, and communication link status data are jointly input into the dynamic event detection process. After the event type is formed, it is classified at the event classification selection node 41 based on multi-source status data, and then respectively enters the environmental change or dynamic no-fly zone processing 42, UAV failure processing 43, and relay UAV failure or communication coverage degradation processing 46. Among them, UAV failure processing 43 is further connected to safe landing path generation 44 and remaining task continuation processing 45, so that the failure event processing forms a continuous processing chain of "fault identification - safe landing - task continuation". Figure 4 What is shown is not simply event tag sorting, but rather introducing different types of events into their corresponding subsequent execution chains.

[0104] For handling sudden environmental changes or dynamic no-fly zones 42, in one embodiment, the effective ground speed of each segment of the affected UAV is first recalculated, and the relationship between the total remaining track distance and the remaining range is evaluated based on the recalculation results. When the evaluation results indicate that the remaining track still meets the conditions for continued execution, the structure of the remaining track remains unchanged, and only the relevant speed parameters are updated. When the evaluation results indicate that the total remaining track distance is greater than the remaining range, it is determined that the affected remaining segments do not meet the requirements for continued execution, and local replanning is performed on the affected remaining segments. In the case of sudden wind changes, the detour direction of the affected segments can also be updated simultaneously, and corresponding candidate paths can be re-formed. Through this processing method, environmental changes do not directly cause the entire patrol mission to be recalculated from scratch, but instead prioritize the response to the affected local segments, thus taking into account both dynamic adaptability and the continuity of static planning results.

[0105] In a further embodiment, the local replanning can be refined using a rolling window search and fitness evaluation method. Specifically, local path segments can be extracted from the affected remaining flight segments according to a preset window size, and disturbance generation and fitness evaluation are performed on candidate paths within the window. The local paths are then updated based on the evaluation results. In a further embodiment, the fitness evaluation simultaneously considers the remaining path flight time, the remaining path distance, and the collision penalty term in the no-fly zone. The contraction and expansion coefficients can decrease with the iteration process, so that the search process gradually transitions from global exploration to local convergence. Through this further embodiment, local replanning can not only remain at the level of "re-bypassing obstacles" but also take into account endurance, safety, and local optimization efficiency during the execution period. This part is particularly suitable as a technical focus when further contraction is required in subsequent technical solutions.

[0106] For UAV fault handling 43, after the faulty UAV is identified, candidate landing point determination and safe landing path generation 44 are performed first. In an optional implementation, it can be further determined whether to trigger the safe landing procedure based on the fault severity index. The fault severity index can be expressed as:

[0107] S f =w f ·η b +w a ·δ a ;

[0108] in,

[0109] S f An index representing the severity of a fault;

[0110] w f Indicates the weighting coefficient for the fault type;

[0111] w a Indicates the weighting coefficient for the degree of abnormality;

[0112] η b Indicates the current remaining battery percentage;

[0113] δ a This represents the normalized outlier deviation.

[0114] In one embodiment, the fault type weights corresponding to power failure, sensor failure, communication failure, and structural failure can be set to 1.0, 0.7, 0.5, and 0.9, respectively. When S f When the preset threshold is exceeded, the safe landing procedure is triggered; when S f If the threshold is not exceeded, the system can enter a degraded patrol or continuous monitoring state. By introducing a differentiated judgment through a fault severity index, all faults can be prevented from entering a forced landing process, and high-risk faults can be handled in a timely manner.

[0115] After identifying the faulty drone, the search range for candidate landing sites is first determined based on its current flight status. This search range can be determined by considering the remaining battery percentage, full-charge range, current effective ground speed, and safety margin factor. In one embodiment, the search radius can be expressed as:

[0116] ;

[0117] Among them, R S Represents the search radius; γ represents the safety margin factor; R full This indicates the driving range on a full charge; υ0 indicates the base cruising speed.

[0118] In one embodiment, the safety margin coefficient can be set to 0.5 to preserve landing maneuverability. After determining the search range, a set of candidate landing points can be generated according to a preset grid step size; in one embodiment, the grid step size can be 200 meters. By incorporating the current effective ground velocity into the search range calculation, the candidate landing point search range can be kept consistent with the current execution capability, rather than being roughly estimated based solely on static remaining battery power.

[0119] In a further embodiment, distance components, terrain slope flatness components, and terrain height variance components can be calculated for each candidate landing point, and a comprehensive score is formed by preset weights to determine the optimal landing point. The distance component is used to characterize the flight cost from the faulty UAV to the candidate landing point, the terrain slope flatness component is used to characterize the smoothness of the landing area surface, and the terrain height variance component is used to characterize the terrain undulation within the neighborhood of the candidate point. After obtaining the optimal landing point, a safe landing path to the optimal landing point is preferably generated using a controlled descent method, and terrain safety constraints, descent rate constraints, and landing endpoint constraints are applied to the path. In one embodiment, the terrain safety margin during the landing phase can be taken as 50 meters, and the landing endpoint accuracy threshold can be taken as 5 meters. After the landing path planning is completed, the system sends a landing notification to the ground station, which includes at least the landing point coordinates, the estimated landing time, and the fault type information. Through the above method, Figure 4 The safe landing path generation in the process is not just about "selecting a point", but a complete sub-chain that includes "determining the search range - generating candidate points - comprehensive scoring - path generation - notification output".

[0120] Continue in Figure 4In the remaining task continuation processing 45, continuation processing is performed on the unfinished tasks of the faulty UAV. In one embodiment, under the constraints of endurance constraints, full task coverage constraints, and no-fly zone constraints, the task allocation matrix is ​​used as the decision variable to reallocate the unfinished tasks of the faulty UAV, and other available UAVs are controlled to continue to perform patrol tasks according to the reallocation results. In a further embodiment, the total flight distance, maximum task completion time, and load balancing degree can be used as optimization objectives to compare candidate continuation schemes; for candidate schemes that do not meet the constraints, they can be eliminated or sorted in descending order according to the degree of constraint violation; for candidate schemes that meet the constraints, they can be further filtered based on the comprehensive score to determine the optimal task continuation scheme. When only one available UAV remains, the system automatically switches to a degraded allocation method with the minimum incremental distance; tasks that cannot be allocated immediately are recorded as pending continuation tasks and are re-included in the allocation after the conditions are restored. Through this processing method, the failure event will not only stop at the level of the faulty UAV exiting, but will further extend to the core issue of "how to continue the task" in multi-aircraft collaboration.

[0121] See Figure 5 , Figure 5 The diagram illustrates the dynamic adjustment and posterior check feedback process of the relay network. This process includes at least a prediction layer 51, a response layer 52, a relay position update 53, an adjustment method selection 54, and a posterior check 55. In one embodiment, the prediction layer 51 determines the target position of the relay drone based at least on forward path information formed by several future waypoints of the patrol drone; the response layer 52 generates a position adjustment amount based on the deviation between the target position and the current position of the relay drone, as well as the current position of the patrol drone and the current communication coverage status; the position adjustment amount is used to drive the relay position update 53 to form the updated relay drone position. Thus, the prediction layer and the response layer are not abstract module names, but rather respectively undertake two different functions: "target position determination for future mission progress" and "position compensation for current deviation status."

[0122] After completing relay location update 53, the process proceeds to adjustment method selection 54. In a further implementation, the relay network adjustment method can be selected from among maintaining the original location, partial adjustment, and global reconstruction, based on the degree of communication coverage degradation and the failure status of the relay drone. Specifically, when the communication coverage degradation is below a first threshold, the original location is maintained; when the communication coverage degradation is between the first and second thresholds, partial adjustment is performed; and when the relay drone fails or the communication coverage degradation exceeds the second threshold, global reconstruction is performed. This hierarchical processing rule avoids triggering large-scale reconstruction with slight fluctuations in communication status and also avoids insufficient recovery due to only minor adjustments when the relay drone fails or communication coverage is severely degraded.

[0123] In one optional implementation, a coverage continuity constraint is maintained during the dynamic adjustment of the relay network, ensuring that the communication coverage rate is not lower than a preset communication coverage rate threshold. In one implementation, the preset communication coverage rate threshold can be 0.95. After the dynamic adjustment of the relay network is completed, a posterior check 55 is performed on the communication coverage status. The posterior check includes at least comparing the adjusted communication coverage rate with the preset communication coverage threshold; when the posterior check result indicates that the communication coverage status does not meet the communication coverage requirements, the posterior check result is fed back to the dynamic adjustment process of the relay network, driving the system to continue executing target location determination, location adjustment amount formation, relay location update, and adjustment method selection; when the posterior check result indicates that the communication coverage status meets the communication coverage requirements, the current patrol execution status is maintained, and the patrol task continues. Through this feedback mechanism, Figure 5 It is no longer just a relay location optimization diagram, but a closed-loop diagram of "adjustment - inspection - satisfaction judgment - continued adjustment or continued patrol".

[0124] In a further embodiment, the posterior check, in addition to comparing the communication coverage threshold, can further combine a link quality threshold or a communication interruption duration threshold for comprehensive judgment. The link quality threshold characterizes the availability of the adjusted relay link, and the communication interruption duration threshold characterizes the maximum allowable continuous communication interruption duration during patrol execution. The communication coverage status is determined to meet the communication coverage requirements only if the adjusted communication coverage status meets the preset communication coverage requirements and the link quality meets the requirements, or the continuous communication interruption duration does not exceed the preset interruption duration threshold. This further embodiment makes the posterior check no longer rely solely on a single coverage indicator, but possesses a comprehensive judgment capability oriented towards communication availability.

[0125] In summary, this embodiment revolves around Figure 3 , Figure 4 and Figure 5 The main closed loop established in Example 1 was further refined at the key rule level. For Figure 3 This embodiment realistically describes the three-mode path formation, the composition of mode evaluation results, the comparison of comprehensive evaluation quantities, and the path safety correction; for Figure 4 In this embodiment, the handling of sudden environmental changes or dynamic no-fly zones, fault handling, and mission continuity are written as a continuous execution chain; for Figure 5This embodiment presents a complete closed-loop process comprising the prediction layer, reaction layer, adjustment method selection, posterior check, and backflow rescheduling. Therefore, without altering the main flow of Embodiment 1, this embodiment further specifies the key processing rules for the static planning phase and the dynamic rescheduling phase during execution, and embeds backup protection layers in the specification, including mode evaluation results, comprehensive evaluation quantities, local replanning fitness functions, fault severity indices, candidate landing point scores, multi-objective optimization for mission succession, and posterior check conditions.

[0126] In one embodiment, see Figure 6 , Figure 6 This document illustrates the system structure and functional module settings for executing the aforementioned plateau multi-UAV static trajectory planning and dynamic rescheduling method. The system is suitable for deployment in plateau patrol scenarios with complex terrain, and is used to complete processes such as patrol mission reception, environmental modeling and performance parameter correction, terrain-aware static trajectory planning, patrol execution monitoring, dynamic event handling, and post-hoc checks of communication coverage status. Unlike Embodiments 1 and 2, which focus on method flow and key rules, this embodiment emphasizes how the above processing chain is implemented at the system structure level, that is, how hardware components support module calls, how processing objects are passed between modules, and how the system completes the same patrol mission processing chain under different deployment methods.

[0127] In one embodiment, the system includes at least an airborne processor 61, a memory 62, a sensor assembly 63, and a communication assembly 64. The airborne processor 61 is connected to the memory 62, the sensor assembly 63, and the communication assembly 64. The airborne processor 61 is used to execute core processing logic such as environmental modeling and performance parameter correction, static trajectory planning, event detection, and dynamic rescheduling. The memory 62 is used to store program instructions executable by the airborne processor 61 and data required for each stage. The sensor assembly 63 is used to collect environmental data and UAV status data. The communication assembly 64 is used to receive patrol mission instructions issued by the ground station, upload patrol mission execution data, acquire communication link status data, and establish a communication link between the patrol UAV, the relay UAV, and the ground station. Figure 6 61-64 in the diagram are not simply a list of hardware components, but rather they undertake four basic responsibilities: computing, storage, sensing, and communication, providing underlying support for the subsequent use of functional modules 65-68.

[0128] In one embodiment, the ground station sends patrol mission instructions to the airborne processor 61 via the communication component 64 and receives patrol data returned during the patrol mission execution. The patrol drone, as a mission execution platform, works collaboratively with the airborne processor 61, sensor component 63, and communication component 64. The relay drone, as a communication support platform, works with the communication component 64 to maintain the communication link between the patrol area and the ground station. Through this collaborative structure, the ground station mainly undertakes the functions of mission issuance, result reception, and collaborative computing when necessary; the patrol drone undertakes mission execution and local processing; the relay drone undertakes communication support; and the airborne processor 61, as the core processing unit of this system, is responsible for connecting the processing chains of environmental modeling and performance parameter correction, static planning, event detection, and dynamic rescheduling.

[0129] In one embodiment, the memory 62 stores program instructions executable by the airborne processor 61. When executed, the program instructions cause the airborne processor 61 to sequentially call the plateau environment modeling and UAV performance correction module 65, the terrain perception static trajectory planning module 66, the sensor real-time monitoring module 67, and the dynamic rescheduling module 68 to complete the aforementioned patrol mission processing chain. The program instructions can be stored in the airborne storage medium or sent from the ground station to the memory 62 via the communication component 64 during deployment or upgrade. In addition to program instructions, the memory 62 can also store task input data, parameter and environmental data, path and scheme data, and status and record data for the airborne processor 61 to access at different processing stages. Through this method of separating program instructions and data storage, the system can stably execute the preset processing chain and achieve deployment flexibility by replacing data or updating instructions in different mission scenarios.

[0130] In one embodiment, the plateau environment modeling and UAV performance correction module 65 is used to form a comprehensive environmental correction factor based on the air pressure and temperature data collected by the sensor component 63, and further form the corrected actual performance parameters. The sensor component 63 includes at least environmental perception sensors and state perception sensors; wherein, the environmental perception sensors provide air pressure data, temperature data, wind speed data, and wind direction data, and the state perception sensors provide remaining battery power, flight speed, attitude information, and position information. After receiving the air pressure and temperature data, the plateau environment modeling and UAV performance correction module 65 forms the comprehensive environmental correction factor, and further corrects the base cruise speed, base climb rate, and base descent rate to output the corrected actual cruise speed, corrected actual climb rate, and corrected actual descent rate. Thus, module 65 corresponds to step S200 in the method embodiment at the system level, and is responsible for providing a unified capability scale for all subsequent processing stages.

[0131] Furthermore, in one embodiment, the corrected actual performance parameters output by module 65 are not only used in the static planning phase, but can also be called upon in the dynamic rescheduling phase for remaining trajectory evaluation, fault landing search range determination, safe landing path generation, and time scale estimation in relay network dynamic adjustment. In other words, module 65 not only undertakes the function of "forming integrated environmental correction factors and corrected actual performance parameters" within the system, but also plays the role of providing a unified capability scale for subsequent modules. By maintaining this unified capability scale at the system level, modules 66, 67, and 68 can maintain consistency in flight capability representation, thereby reducing the scale gap between the static planning phase and the execution phase. Although this point has been explained in the method embodiment, it also needs to be clarified in the system embodiment, as it determines which core processing objects are shared among the functional modules.

[0132] In one embodiment, the terrain-aware static trajectory planning module 66 receives the patrol point set and digital elevation model data from the patrol mission instruction received by the communication component 64, and performs terrain-aware static trajectory planning on candidate flight paths between patrol points based on the corrected actual performance parameters output by module 65. Specifically, module 66 can first perform a safety judgment on straight paths; if the straight path meets the safety conditions, it directly outputs the corresponding single-segment flight path; if the straight path does not meet the safety conditions, it forms paths corresponding to steep ascent, steep descent, and steep turn modes, and performs a comparison based on the mode evaluation results to determine the optimal mode corresponding path; subsequently, module 66 performs path safety correction on the optimal mode corresponding path, thereby forming a single-segment flight path that meets the terrain safety constraints, and further splices them to form a static trajectory scheme. Thus, module 66 corresponds to step S300 in the method embodiment at the system level and directly undertakes... Figure 3 The static programming processing chain is shown.

[0133] In one optional implementation, the pattern evaluation result generated by module 66 includes at least a flight time term; in a further implementation, the pattern evaluation result may also include a path cost term, and a comprehensive evaluation quantity can be formed through weighted combination to comprehensively compare and select paths corresponding to different patterns. Thus, module 66 not only performs path generation but also undertakes the composite functions of pattern comparison, optimal output, and path safety correction. Unifying these subprocesses into module 66 at the system level helps to... Figure 6 The system structure and Figure 3 The static programming processing chain shown remains consistent, eliminating the need to break down the three modes into multiple finer system modules, thus achieving a balance. Figure 6 Structural simplicity and completeness of methodological rules.

[0134] In one embodiment, the real-time sensor monitoring module 67 receives environmental data and UAV status data collected by the sensor component 63, as well as communication link status data acquired by the communication component 64, to detect dynamic events and determine event types. The environmental data includes at least real-time air pressure, temperature, wind speed, and wind direction; the UAV status data includes at least remaining battery power, flight speed, attitude information, and location information; and the communication link status data includes at least link quality and coverage status. Module 67 performs fusion judgment on the above multiple data types and outputs the event type, which includes at least environmental changes, dynamic no-fly zones, UAV malfunctions, and relay UAV failures or decreased communication coverage. Therefore, module 67 corresponds to step S400 in the method embodiment at the system level and plays a role in the conversion from "state perception to event type" throughout the system.

[0135] In one embodiment, the dynamic rescheduling module 68 is used to perform corresponding adjustments to the relevant items in the affected remaining tracks, remaining patrol tasks, and communication coverage status according to the event type output by module 67, and to perform a post-hoc check on the communication coverage status after the corresponding adjustments, so as to determine whether to continue the patrol task or continue the relay network dynamic adjustment based on the post-hoc check result. Thus, module 68 undertakes the execution-period dynamic processing tasks corresponding to steps S500, S600, and S700 in the method embodiment at the system level. Compared with modules 65, 66, and 67, module 68 has a larger internal functional span, and therefore its input-output relationship is also richer: on the one hand, it receives static track schemes, event types, and various status data; on the other hand, it outputs updated remaining tracks, safe landing paths, remaining task continuation results, position adjustment amounts, and post-hoc check results.

[0136] Specifically, when the event type is an environmental abrupt change or a dynamic no-fly zone, module 68 processes the remaining flight paths of the affected UAVs. This processing includes at least: recalculating the effective ground speed of each segment of the affected UAV, and evaluating the relationship between the total remaining flight path distance and the remaining range based on the recalculation results; when the evaluation results indicate that the total remaining flight path distance is greater than the remaining range, performing local replanning on the affected remaining flight paths; when the evaluation results indicate that the remaining flight path still meets the requirements for continued execution, only updating the relevant speed parameters and keeping the remaining flight path structure unchanged. Through this processing method, module 68 can maintain the continuity of the static flight path scheme at the system level, while making local adaptations to sudden environmental changes during execution, without having to completely overturn the existing flight path planning results.

[0137] When the event type is UAV malfunction, module 68 further undertakes the functions of generating a safe landing path and handling the continuation of remaining tasks. In one embodiment, module 68 can further determine whether to trigger a safe landing procedure based on the severity of the malfunction; after triggering the safe landing procedure, it first determines the search range of candidate landing points, then generates a set of candidate landing points, and forms a comprehensive score based on the distance component, terrain slope flatness component, and terrain height variance component of the candidate landing points to determine the optimal landing point; after determining the optimal landing point, it generates a safe landing path to the optimal landing point. After the malfunctioning UAV exits the mission, module 68 performs continuation processing on the unfinished tasks. Under the constraints of endurance, full mission coverage, and no-fly zone, the unfinished tasks are reassigned to other available UAVs; when only one available UAV remains, it switches to a degraded allocation method with the minimum incremental distance; for tasks that cannot be immediately allocated, they are recorded as pending continuation tasks and are re-included in the allocation after the conditions are restored.

[0138] When the event type is relay drone failure, or when communication coverage is determined to be declining based on communication link status data, module 68 performs dynamic adjustment of the relay network. Specifically, module 68 can internally call the prediction layer and reaction layer processing logic, where the prediction layer determines the target position of the relay drone based on forward path information, and the reaction layer generates a position adjustment amount based on the relative position between the patrol drone and the relay drone; module 68 updates the relay drone position based on the position adjustment amount, and selects a relay network adjustment method among maintaining the original position, partial adjustment, and global reconstruction based on the communication coverage status and / or the relay drone failure status; in one embodiment, the relay network adjustment method can be selected in stages according to the degree of communication coverage decline and / or relay drone failure, so that the three methods of maintaining the original position, partial adjustment, and global reconstruction match the current communication assurance risk level. During the adjustment process, the coverage continuity constraint of maintaining the communication coverage rate not lower than a preset threshold is maintained; after the adjustment is completed, a post-hoc check is performed on the communication coverage status, and the decision on whether to continue the patrol mission or continue to perform dynamic adjustment of the relay network is made based on the post-hoc check result. By incorporating the dynamic adjustment of the relay network and the post-hoc check into module 68, the system architecture can be maintained as a four-module system, while retaining the closed-loop characteristics of "adjustment-check-back" in the implementation method.

[0139] In one embodiment, the communication component 64 is used not only to establish the basic communication link between the patrol drone, the relay drone, and the ground station, but also to transmit multiple types of processing objects at different processing stages. Specifically, the communication component 64 can at least be used to receive patrol mission instructions issued by the ground station, upload patrol mission execution data, obtain communication link status data, establish a communication link between the patrol drone and the relay drone, and establish a communication link between the relay drone and the ground station. In a further embodiment, the communication component 64 can also be used to transmit static trajectory schemes, updated remaining trajectories, remaining mission continuation results, and relay position update results. Thus, the communication component 64 not only undertakes the underlying link establishment function, but also undertakes the responsibility of transmitting multiple key processing objects between the ground station, the patrol drone, and the relay drone, thereby becoming an important foundation for the establishment of the "object flow relationship" within the system.

[0140] In one embodiment, the system can be deployed in multiple ways. In one deployment, the plateau environment modeling and UAV performance correction module 65, the terrain-aware static trajectory planning module 66, the sensor real-time monitoring module 67, and the dynamic rescheduling module 68 are all executed by the onboard processor 61 on the patrol UAV, thus forming a fully airborne deployment structure. In another deployment, the onboard processor 61 is responsible for environment modeling and performance parameter correction, real-time monitoring, and dynamic rescheduling. It collaborates with the ground station to perform some static trajectory formation and static task organization calculations, and then transmits the static trajectory scheme or update results back to the onboard processor 61 via the communication component 64. In yet another deployment, the dynamic adjustment of the relay network can be executed collaboratively by the processing unit on the relay UAV and the onboard processor 61 on the patrol UAV. The patrol UAV provides path and status information, while the relay UAV performs position updates based on communication support requirements. Thus, the system can operate entirely onboard, or it can be deployed in a ground station-assisted or relay-UAV-assisted manner depending on communication conditions and task complexity. By retaining this deployment flexibility in the system implementation, it is possible to better meet the actual application needs of high-altitude patrol scenarios, where communication conditions vary greatly and task complexity fluctuates significantly.

[0141] In one embodiment, the program instructions stored in the memory 62 can correspond to environmental modeling and performance parameter correction instructions, static track formation instructions, real-time monitoring instructions, and dynamic rescheduling instructions. When the airborne processor 61 executes the program instructions, it sequentially calls the environmental modeling and performance parameter correction, static track formation, event detection, and dynamic rescheduling processes. The functional modules in the system can be represented as logical functional modules, or they can be implemented in software on the airborne processor 61 using program instructions stored in the memory 62. In another embodiment, some or all module functions can be implemented through dedicated processing logic, programmable logic devices, or a combination of hardware and software. Although the above implementation methods differ in physical implementation, they do not affect the object flow relationship and module calling relationship of the aforementioned system. By retaining this description, the system embodiments can be avoided from being misunderstood as necessarily requiring a single, unique implementation form.

[0142] In this embodiment, through the hardware configuration of the onboard processor 61, memory 62, sensor components 63, and communication components 64, and the calling relationships of the plateau environment modeling and UAV performance correction module 65, the terrain-aware static trajectory planning module 66, the sensor real-time monitoring module 67, and the dynamic rescheduling module 68, the system can execute the aforementioned patrol task processing chain. Specifically, the system uses the plateau environment modeling and UAV performance correction module 65 to generate corrected actual performance parameters, the terrain-aware static trajectory planning module 66 to generate a static trajectory plan, the sensor real-time monitoring module 67 to generate event types, and then the dynamic rescheduling module 68 performs remaining trajectory processing, safe landing path generation, remaining task continuation processing, and relay network dynamic adjustment according to different event types, and decides whether to continue patrol execution or continue relay network dynamic adjustment based on the posterior check results. Thus, Figure 6 The system structure shown is not a separate scheme parallel to the method implementation, but a specific implementation of the method processing chain described in Embodiment 1 and Embodiment 2 at the system structure level.

[0143] In summary, this embodiment revolves around Figure 6 This paper provides a system-level mapping explanation of environmental modeling and performance parameter correction, static trajectory planning, event detection, dynamic rescheduling, and post-hoc check closed loop in the aforementioned method embodiments. By clarifying the hardware configuration relationships of 61-64, the functional module calling relationships of 65-68, the transmission responsibilities of multiple processing objects undertaken by communication component 64, and the deployment methods such as full airborne / ground station collaboration / relay aircraft collaboration, this embodiment enables the system solution to obtain structural support corresponding to the method embodiments. Meanwhile, since the dynamic rescheduling module 68 continues to retain key processing relationships such as remaining trajectory processing, safe landing path generation, remaining task continuation processing, relay network dynamic adjustment, and post-hoc check closed loop, therefore... Figure 6 and Figures 1 to 5 There is also a clear one-to-one correspondence between them.

[0144] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for static trajectory planning and dynamic rescheduling of multiple unmanned aerial vehicles (UAVs) in high-altitude areas, characterized in that, include: Receive patrol mission instructions and acquire patrol point sets, digital elevation model data, and air pressure and temperature data corresponding to the plateau environment; A comprehensive environmental correction factor is determined based on the air pressure data and the temperature data, and the corrected actual performance parameters are determined based on the comprehensive environmental correction factor. Based on the digital elevation model data and the corrected actual performance parameters, terrain-aware static trajectory planning is performed on the candidate flight paths between the patrol points. Specifically, by comparing the candidate paths in steep ascent, steep descent, and steep turn modes, a static trajectory scheme that meets terrain safety constraints and is assigned to each UAV is formed. The patrol mission is carried out according to the static flight path scheme, and environmental data, UAV status data and communication link status data are collected. Based on the environmental data, the UAV status data and the communication link status data, dynamic events are detected and the event types are determined. Based on the event type, corresponding adjustments are made to the affected remaining flight paths, remaining patrol missions, or communication coverage status. These adjustments include at least: when the event type is an environmental change or a dynamic no-fly zone, replanning the remaining flight paths of the affected UAVs; when the event type is a UAV malfunction, generating a safe landing path for the malfunctioning UAV and reallocating its remaining patrol missions to other available UAVs; and when the event type is a relay UAV failure, or when communication coverage status is determined to be declining based on the communication link status data, dynamic adjustments to the relay network are performed. A post-hoc check is performed on the corresponding adjusted communication coverage status, and the result of the post-hoc check determines whether to continue the patrol mission or continue the dynamic adjustment of the relay network.

2. The plateau multi-UAV static trajectory planning and dynamic rescheduling method according to claim 1, characterized in that, The terrain-aware static trajectory planning includes: Perform a safety assessment on the straight path; When the straight path does not meet the safety conditions, paths corresponding to the steep ascent mode, steep descent mode, and steep turn mode are generated respectively, and mode evaluation results are formed respectively. Based on the evaluation results of the aforementioned modes, the steep ascent mode, the steep descent mode, and the steep turn mode are compared and selected; The optimal path is determined based on the comparison results and used as a single-segment flight path in the static trajectory scheme. Path safety correction is then performed on the optimal path.

3. The plateau multi-UAV static trajectory planning and dynamic rescheduling method according to claim 2, characterized in that, The steep ascent mode includes: Calculate the terrain slope gradient of the path sampling point sequence; Whether to trigger pre-climb is determined based on the maximum gradient within the look-ahead window; When triggering a preclimb, determine the target preclimb height, optimal climb angle, shortest preclimb start distance, gradient urgency factor, and final preclimb start distance; Based on the pre-climb target height, the optimal climb angle, the shortest pre-climb start distance, the gradient urgency factor, and the final pre-climb start distance, a pre-climb height sequence is generated to form the path corresponding to the steep ascent mode.

4. The plateau multi-UAV static trajectory planning and dynamic rescheduling method according to claim 2, characterized in that, The steep descent mode includes: Calculate the terrain gradient along the flight direction at each sampling point on the path; Determine the target ground-hull height sequence based on the terrain gradient; Construct a cost function for reducing energy consumption; The descent path is divided into sections based on gentle slopes, medium slopes, and steep slopes; The optimal descent altitude sequence is determined based on the descent energy cost function and the segmentation results to form the path corresponding to the steep descent mode.

5. The plateau multi-UAV static trajectory planning and dynamic rescheduling method according to claim 2, characterized in that, The steep turn mode includes: Extract the projection width of the obstacle to be avoided along the path normal vector direction; The basic bypass radius is determined based on the projected width; The wind direction correction factor is determined based on the wind direction and the detour direction, and the dynamic detour radius is determined accordingly. Calculate the detour cost for both the left and right sides separately; The detour direction and its corresponding dynamic detour radius are determined based on the detour cost, and the path corresponding to the steep turn mode is generated.

6. The plateau multi-UAV static trajectory planning and dynamic rescheduling method according to claim 1, characterized in that, When the event type is an environmental change or a dynamic no-fly zone, the replanning of the remaining flight paths of the affected UAV includes: Determine whether to trigger a replanning based on changes in environmental data and / or dynamic changes in no-fly zones; The effective ground speed of each segment of the affected UAV was recalculated, and the relationship between the remaining total track distance and the remaining range was evaluated based on the recalculation results. If the evaluation results indicate that the total remaining flight path distance is greater than the remaining flight range, a local replanning is performed on the affected remaining flight segment. The evaluation of the local replanning considers at least the remaining path flight time, the remaining path distance, and the no-fly zone collision penalty. In the event of a sudden change in wind direction, the detour direction of the affected flight segment is updated synchronously, and the local replanning is performed accordingly.

7. The plateau multi-UAV static trajectory planning and dynamic rescheduling method according to claim 1, characterized in that, The drone malfunction includes at least one of power malfunction, sensor malfunction, and structural or aerodynamic malfunction. The process of generating a safe landing path for a faulty drone includes: The severity of the fault determines whether to trigger a safe landing procedure or enter a degraded patrol or continuous monitoring state. When the safe landing procedure is triggered, the search range of candidate landing points is determined based on the current flight status. The search range of candidate landing points is determined by combining the remaining battery percentage, the range on a full charge, the current effective ground speed, and the safety margin coefficient. A set of candidate landing points is generated within the search range of the candidate landing points; Each candidate landing point is scored based on distance component, terrain slope flatness component, and terrain height variance component. The optimal landing point is determined based on the scoring results, and a safe landing path to the optimal landing point is generated.

8. The plateau multi-UAV static trajectory planning and dynamic rescheduling method according to claim 7, characterized in that, The reassignment of the remaining patrol missions of the faulty drone to other available drones includes: Using the task allocation matrix as the decision variable, the unfinished tasks of the faulty UAVs are reallocated; During the reallocation process, endurance constraints, mission coverage constraints, and no-fly zone constraints are applied. The unfinished tasks are assigned to other available drones, and the other available drones are controlled to continue the patrol mission according to the reassignment result; When only one usable drone remains, downgraded allocation is performed based on the minimum incremental distance. Tasks that cannot be immediately assigned are recorded as pending tasks and will be reassigned once conditions are restored.

9. The plateau multi-UAV static trajectory planning and dynamic rescheduling method according to claim 1, characterized in that, The dynamic adjustment of the relay network includes: The target location of the relay drone is determined by the prediction layer based on the forward path information. The reaction layer generates position adjustment amounts based on the relative positions between patrol drones and relay drones. Update the relay drone's position based on the aforementioned position adjustment amount; Based on the degree of communication coverage degradation and / or relay drone failure, select the relay network adjustment method from among maintaining the original location, partial adjustment, and global reconstruction; During the adjustment process, the coverage continuity constraint is maintained to ensure that the communication coverage rate is not lower than the preset communication coverage threshold; After the relay network dynamic adjustment or other dynamic event processing is completed, a post-hoc check is performed on the communication coverage status, and the relay network dynamic adjustment continues to be performed if the post-hoc check result shows that the communication coverage status does not meet the communication coverage requirements.

10. A plateau multi-UAV static trajectory planning and dynamic rescheduling system, characterized in that, The system includes an onboard processor, a memory, sensor components, and a communication component. The onboard processor is connected to the memory, sensor components, and communication component, respectively. The memory stores program instructions executable by the onboard processor. When executed, these program instructions cause the onboard processor to call: The plateau environment modeling and UAV performance correction module is used to determine a comprehensive environmental correction factor based on the air pressure and temperature data collected by the sensor components, and to determine the corrected actual performance parameters based on the comprehensive environmental correction factor. The terrain-aware static trajectory planning module is used to receive the set of patrol points and digital elevation model data in the patrol mission instruction received by the communication component, and to perform terrain-aware static trajectory planning on the candidate flight paths between the patrol points based on the corrected actual performance parameters. In this module, by comparing the candidate paths in steep ascent mode, steep descent mode and steep turn mode, a static trajectory scheme that meets the terrain safety constraints and is assigned to each UAV is formed. The sensor real-time monitoring module is used to receive environmental data and UAV status data collected by the sensor components and communication link status data obtained by the communication components, and to detect dynamic events and determine the event type based on the environmental data, the UAV status data and the communication link status data. The dynamic rescheduling module is used to perform corresponding adjustments to the affected remaining flight paths, remaining patrol tasks, or communication coverage status according to the event type, and to perform a post-hoc check on the adjusted communication coverage status to determine whether to continue the patrol task or continue the relay network dynamic adjustment based on the post-hoc check result. The corresponding adjustments include at least: when the event type is an environmental change or a dynamic no-fly zone, replanning the remaining flight paths of the affected UAVs; when the event type is a UAV malfunction, generating a safe landing path for the malfunctioning UAV and reallocating the remaining patrol tasks of the malfunctioning UAV to other available UAVs; and when the event type is a relay UAV failure, or when the communication coverage status is determined to have decreased based on the communication link status data, performing relay network dynamic adjustment.

Images