An unmanned aerial vehicle cluster low-altitude path intelligent planning and scheduling method and system

Abstract

The application discloses a kind of unmanned aerial vehicle cluster low-altitude path intelligent planning and scheduling method and system, it is related to unmanned aerial vehicle intelligent control technical field, including task route analysis module, flight state record analysis module, low-altitude disturbance memory atlas construction module, path dynamic correction module and cluster scheduling execution module.Task route analysis module reads historical task record, generates task semantic draft according to task action;Flight state record analysis module generates instruction execution rebound trace, hover energy consumption tail texture and communication posture entrainment fragment based on task semantic draft;Low-altitude disturbance memory atlas construction module carries out individual factor stripping, passage segment co-occurrence judgment and action chain compression to the above passage features, generates low-altitude disturbance memory core;Path dynamic correction module generates low-altitude memory trial fitting route according to low-altitude disturbance memory core;The present application is used to deal with the implicit passage anomaly in low-altitude passable area.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for unmanned aerial vehicles (UAVs), specifically to a method and system for intelligent planning and scheduling of low-altitude paths for UAV swarms. Background Technology

[0002] In existing low-altitude path planning for UAV swarms, path generation is typically based on mission waypoints, map obstacle boundaries, no-fly zones, and flight distance consumption. While the system can determine whether a flight path is spatially passable, it struggles to identify hidden passage problems in locations such as gaps between buildings, under bridges, near billboards, and at the edges of tree canopies. These locations are usually represented as passable areas on the map, but during actual UAV flight, phenomena such as deviating from planned actions and repeated corrections, continued changes in energy consumption per unit segment after hovering operations, and communication link changes accompanied by attitude corrections often occur. Existing technologies often categorize these phenomena separately under flight control, energy management, or communication link management, lacking a processing method that integrates mission actions, flight control execution, hovering control, energy consumption per unit segment, and communication links into a single mission semantic passable segment for analysis. This makes it difficult to transform past flight anomalies into path scheduling data for subsequent missions.

[0003] The aforementioned shortcomings are particularly pronounced when drone swarms perform low-altitude tasks such as inspection, photography, security, and search. When multiple drones enter the same low-altitude passageway, if there are airflow disturbances, hovering energy trails, changes in communication links, or space constraints within the current passageway, the single-drone path planning result may still cause multiple drones to enter the current passageway along the original candidate routes. This leads to abnormal results such as repeated route corrections, repeated adjustments to hovering positions, disordered mission data transmission order, and mismatched formation entry order. These abnormal results disrupt the mission operation sequence of the drone swarm, resulting in multiple correction processes between mission routes and flight control actions. Furthermore, passageway characteristics already observed in historical missions cannot be inherited and used in subsequent missions. Therefore, it is necessary to construct an intelligent low-altitude path planning and scheduling system for drone swarms that can generate a low-altitude disturbance memory kernel based on historical mission semantic passageways and use this kernel to generate low-altitude memory trial routes and low-altitude disturbance memory scheduling route packages. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method and system for intelligent planning and scheduling of low-altitude paths for unmanned aerial vehicle (UAV) swarms, solving the problems mentioned in the background section.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent planning and scheduling system for low-altitude paths of unmanned aerial vehicle (UAV) swarms, comprising a mission route analysis module, a flight status record parsing module, a low-altitude disturbance memory map construction module, a path dynamic correction module, and a swarm scheduling execution module; The mission route analysis module reads mission route records, path planning instruction records, flight control execution records, hovering control records, unit segment energy consumption records, and communication link records from historical missions. It divides historical missions into mission semantic passage segments according to mission actions and generates mission semantic passage drafts. The flight status recording and analysis module generates command execution rebound traces, hovering energy consumption trail textures, and communication attitude fragments based on the mission semantic communication draft. The low-altitude disturbance memory map construction module performs individual UAV peeling, passage segment co-occurrence judgment and action chain compression on command execution bounce traces, hovering energy consumption trail textures and communication attitude clips to generate a low-altitude disturbance memory kernel. The path dynamic correction module divides the candidate routes of the new mission into new mission semantic passage segments, and generates low-altitude memory trial routes based on the matching results of the new mission semantic passage segments and the low-altitude disturbance memory core. The cluster scheduling and execution module generates low-altitude disturbance memory scheduling route packages based on low-altitude memory trial flight routes, combined with individual UAV operation profiles and task operation sequences.

[0006] Preferably, the mission route analysis module includes a multi-source mission record arrangement unit and a mission semantic passage segment encapsulation unit; The multi-source mission record arrangement unit reads mission route records, path planning instruction records, flight control execution records, hovering control records, unit segment energy consumption records, and communication link records from historical missions, and extracts historical mission number, UAV number, record time field, waypoint number, and mission action field from the records; Among them, the historical mission number comes from the mission flight path record; the UAV number comes from the flight control execution record; the record time field comes from the time field of each type of record; the waypoint number comes from the mission flight path record and the path planning instruction record; and the mission action field comes from the path planning instruction record and the mission flight path record. Based on the historical mission number and UAV number, the mission route records, path planning instruction records, flight control execution records, hovering control records, unit segment energy consumption records, and communication link records formed by the same UAV in the same historical mission are arranged according to the record time field to generate a single-aircraft mission execution time chain. In the single-machine mission execution time chain, each record position retains its source record type, enabling subsequent processing to distinguish whether the current record content comes from the mission flight path record, path planning instruction record, flight control execution record, hovering control record, unit segment energy consumption record, or communication link record. After generating the single-machine task execution time chain, the path planning instruction content in the path planning instruction record, the flight control execution content in the flight control execution record, the hovering control content in the hovering control record, the unit segment energy consumption content in the unit segment energy consumption record, and the communication link content in the communication link record under the same waypoint number are grouped into the same waypoint execution segment. The task semantic passage segment encapsulation unit receives the single-machine task execution time chain and waypoint execution fragments, and marks the task action category for each waypoint execution fragment according to the task action fields in the path planning instruction record and the task route record; The mission action categories include takeoff, straight-through, turning, hovering, photography, inspection, and return. The sources of these mission action categories are the action commands in the path planning instruction record and the mission node descriptions in the mission route record. Flight control execution records, hovering control records, unit segment energy consumption records, and communication link records are only used as sources of execution content under the current mission action category. Locate the record position where the task action category changes in the single-machine task execution time chain, and use the task action category change position as the boundary of the passage segment; Waypoint execution segments that are located within the same passageway boundary and have the same task action category are merged into a single task semantic passageway. The mission semantic passage retains historical mission numbers, UAV numbers, mission action categories, entry actions, departure actions, path planning instruction fragments, flight control execution fragments, hovering control fragments, unit segment energy consumption fragments, and communication link fragments. Specifically, the entry action originates from the path planning instruction content of the first waypoint execution fragment within the current mission semantic passage, and the departure action originates from the path planning instruction content of the last waypoint execution fragment within the current mission semantic passage. The path planning instruction fragments, flight control execution fragments, hovering control fragments, unit segment energy consumption fragments, and communication link fragments all originate from merged waypoint execution fragments within the current mission semantic passage. Multiple task semantic access segments formed by the same historical mission and the same drone are arranged according to the record time field to generate a task semantic access draft.

[0007] Preferably, the flight status recording and parsing module includes a segment action sequence construction unit and a passage feature generation unit; The same action sequence construction unit reads multiple task semantic passage segments from the task semantic passage draft. Each task semantic passage segment contains historical task number, UAV number, task action category, entry action, departure action, path planning instruction segment, flight control execution segment, hovering control segment, unit segment energy consumption segment, and communication link segment. Using a single task semantic passage as the processing object, the path planning instruction fragment in the current task semantic passage is read, and the planning heading action, planning speed action, planning altitude action, planning hovering action and planning departure action in the path planning instruction fragment are arranged according to the recorded time field to form a planning action sequence; Read the flight control execution segment in the current task semantic passage, and arrange the actual heading action, actual speed action, actual altitude action, attitude correction action, heading recovery action and speed correction action in the flight control execution segment according to the recorded time field to form an execution action sequence; Read the hover control segment in the semantic passage of the current task, and arrange the hover entry action, hover hold action and hover exit action according to the recorded time field to form a hover action sequence; Read the unit segment energy consumption fragment in the current task semantic passage segment, arrange the unit segment energy consumption change records according to the record time field, and form an energy consumption change sequence; Read the communication link segments in the current task semantic passage segment, and arrange the communication link connection record, communication link interruption record, communication link recovery record and communication link jitter record according to the record time field to form a communication action sequence; The planning action sequence, execution action sequence, hovering action sequence, energy consumption change sequence, and communication action sequence are collectively encapsulated into a single action sequence.

[0008] Preferably, the passage feature generation unit generates passage features based on the same action sequence within the same task semantic passage segment, namely instruction execution bounce trace, hovering energy consumption trail texture and communication posture clipping fragment; The method for obtaining the bounce trace of command execution is as follows: take the planned heading action and planned speed action in the planned action sequence as reference actions, and then find the action chain after the actual heading action or actual speed action is separated from the reference action in the executed action sequence. The action chain, in chronological order, includes: the action to detach from the reference action, the attitude correction action generated by the flight control execution segment, the heading correction action generated by the flight control execution segment, the speed correction action generated by the flight control execution segment, and the action to re-enter the reference action; the action chain, along with the historical mission number, UAV number, and mission action category of the mission semantic passage segment, is encapsulated as a command execution bounce trace; The method for obtaining hovering energy consumption trail texture is as follows: determine the hovering operation action chain based on the hovering entry action, hovering hold action and hovering departure action in the hovering action sequence, and then read the unit segment energy consumption change record located after the hovering departure action in the energy consumption change sequence. The energy consumption changes per unit segment of the hovering hold action, hovering departure action, and hovering departure action that continues with the passage action, as well as the speed recovery action in the sequence of actions, are encapsulated as hovering energy consumption trail textures. The method for obtaining communication attitude clips is as follows: search for communication link jitter records, communication link interruption records, or communication link recovery records in the communication action sequence, then read the execution action sequence during the communication action, and extract the attitude correction action, heading correction action, and speed correction action that occur in the same period. The communication link jitter record, communication link interruption record, or communication link recovery record, along with the attitude correction action, heading righting action, and speed correction action within the same period, are encapsulated into a communication attitude clip.

[0009] Preferably, the low-altitude disturbance memory map construction module includes an individual factor stripping and passage segment co-occurrence unit and a disturbance action chain compression and memory core encapsulation unit; Individual factor stripping and the co-occurrence unit of the passage segment encapsulates instruction execution bounce traces, hovering energy consumption trail textures, and communication posture clipping fragments into passage feature groups; The access feature groups are arranged individually according to the drone number. Under the same drone number, the same access features that appear in at least two different mission semantic access segment identifiers are listed as individual drone operation factors. Among them, the task semantic passage segment identifier is the segment number generated by the task flight path analysis module for the task semantic passage segment. The task semantic passage segment identifier includes the historical task number, UAV number, task action category and segment number. Specifically, when the command execution bounce traces under the same UAV number appear in the mission semantic passages of the straight-through action, turning action and return-home action respectively, the current command execution bounce traces are listed as individual UAV operation factors of the current UAV number. When hovering energy consumption trail textures under the same UAV number appear in the task semantic passages of multiple hovering operation actions, the current hovering energy consumption trail texture is listed as an individual UAV operation factor for the current UAV number. When communication attitude clips under the same UAV number appear in the task semantic passages of multiple task action categories, the current communication attitude clip is listed as an individual UAV operation factor for the current UAV number. After separating the individual operational factors of the UAV, the remaining traffic features are analyzed for co-occurrence according to the traffic segment identifier of the mission semantics. For the same semantic passage segment identifier, read the instruction execution rebound traces, hovering energy consumption trail textures, and communication attitude fragments under different UAV numbers, and arrange them according to the recorded time field; When there are at least two different UAV numbers with the same type of passage feature under the same task semantic passage segment identifier, or when there are at least two different types of passage features with different UAV numbers arranged consecutively according to the record time field, the current task semantic passage segment identifier is listed as a passage segment environmental disturbance factor, and a passage segment co-occurrence record is generated.

[0010] The passage segment co-occurrence record includes the task semantic passage segment identifier, the set of UAV numbers participating in the co-occurrence, the names of unconventional passage features participating in the co-occurrence, the task action categories participating in the co-occurrence, the record time field participating in the co-occurrence, and the action composition content participating in the co-occurrence; The co-occurrence records of the passage segment are derived from three types of unconventional passage features after separation of individual UAV operational factors. They are used to indicate that the flight anomalies in the current mission semantic passage segment are not caused by the execution characteristics of a single UAV, but by the low-altitude passage conditions of the current mission semantic passage segment. Preferably, the perturbation action chain compression and memory kernel encapsulation unit is based on the co-occurrence record of the passage segment and takes the semantic passage segment identifier of each task as the processing object to perform action chain compression processing on the passage features that participate in co-occurrence. First, read the instruction execution bounce traces, hovering energy consumption trail textures, and communication posture clips from the co-occurrence records of the passage segment, and then arrange the action composition content of the three types of unconventional passage features according to the recording time field. In the action chain compression process, the action of leaving the reference, attitude correction, heading correction, speed correction and re-entering the reference in the instruction execution rebound trace are regarded as the first type of action node. The hovering hold action, hovering departure action, unit segment energy consumption change record and speed recovery action in the hovering energy consumption trail texture are regarded as the second type of action nodes; The communication link jitter record, communication link interruption record, communication link recovery record, attitude correction action, heading correction action, and speed correction action in the communication attitude clip are regarded as the third type of action node; The action nodes are arranged into a perturbation action chain according to the recording time field, and the source unconventional common feature name of each type of action node is retained in the semantic common segment identifier of the same task. Based on the disturbance action chain, generate passage handling action indicators; When the disturbance action chain contains instruction execution bounce traces but does not contain hovering energy consumption trail textures and communication attitude fragments, the passage handling action is identified as speed smooth passage. When the disturbance action chain includes hover energy consumption trail texture, the passage action identifier is hover position replacement; When the disturbance action chain contains communication attitude clippings, the passage handling action is identified as a reordering of the transmission sequence; When the disturbance action chain contains both command execution bounce traces and communication attitude clips, and the passage segment co-occurrence record contains multiple UAV numbers, the passage handling action is identified as formation convergence passage; Finally, the task semantic passage segment identifier, passage segment co-occurrence record, and passage handling action identifier are encapsulated into a low-altitude disturbance memory kernel.

[0011] Preferably, the path dynamic correction module includes a new task passage segment matching unit and a passage handling route trial matching unit; The new mission passage matching unit receives new UAV swarm missions and low-altitude disturbance memory cores; The new drone swarm missions include mission waypoints, mission actions, and mission sequence. First, a set of candidate routes is generated based on the task waypoints, task actions, and task operation sequence in the new UAV swarm mission. Each candidate route includes a candidate route number, a sequence of task waypoints, a sequence of task actions, and a task operation sequence. The candidate route set is divided according to the sequence of task actions, and the waypoints continuously covered by the same task action are encapsulated into a new task semantic passage segment; each new task semantic passage segment contains a candidate route number, a new task action category, an entry waypoint, a departure waypoint, a sequence of waypoints within the segment, and a task operation order. The new task semantic passage is matched with the low-altitude perturbation memory kernel, as follows: Simultaneously, read the new mission action category, waypoint entry, waypoint exit, intra-segment waypoint sequence, and mission operation sequence from the new mission semantic passage segment, and read the mission semantic passage segment identifier and passage segment co-occurrence record from the low-altitude disturbance memory kernel; The read content is compared with similar fields to obtain the memory kernel matching results. The memory kernel matching results include candidate route number, new task semantic passage segment number, matched low-altitude disturbance memory kernel, passage segment co-occurrence record and passage handling action identifier.

[0012] The specific comparison content for similar fields is as follows: Task action category matching: The new task passage segment matching unit reads the new task action category in the new task semantic passage segment and reads the historical task action category in the historical task semantic passage segment associated with the low-altitude disturbance memory kernel. When both are the same task action category among straight-through action, turning action, hovering operation action, shooting action, inspection action, or return action, a task action category matching result is generated; Entry waypoint matching: The new mission passage segment matching unit reads the entry waypoint from the new mission semantic passage segment and reads the historical entry waypoint from the historical mission semantic passage segment. When both belong to the same low-altitude passage entry waypoint code, an entry waypoint matching result is generated; Departure waypoint matching: The new mission passage segment matching unit reads the departure waypoints from the new mission semantic passage segment and reads the historical departure waypoints from the historical mission semantic passage segments. When both belong to the same low-altitude passage exit waypoint code, a departure waypoint matching result is generated; Intra-segment waypoint sequence matching: The new mission passage segment matching unit reads the intra-segment waypoint sequence in the new mission semantic passage segment and reads the historical intra-segment waypoint sequence in the historical mission semantic passage segment. When the new intra-segment waypoint sequence and the historical intra-segment waypoint sequence contain the same low-altitude passage entrance waypoint code, low-altitude passage exit waypoint code, and consecutive passage waypoint codes between them, an intra-segment waypoint sequence matching result is generated; Task operation sequence matching: The new task passage segment matching unit reads the task operation sequence in the new task semantic passage segment and reads the historical task operation sequence in the historical task semantic passage segment. When both are in the same type of task operation stage, such as passing before entering the operation point, hovering at the operation point, leaving the operation point, or passing before returning, a task operation sequence matching result is generated; Preferably, the passage handling route matching unit performs route matching processing on the candidate route set based on the candidate route number, the new task semantic passage segment number and the passage handling action identifier in the memory kernel matching result; When the passage action is marked as smooth passage, the entry waypoint, the intra-segment waypoint sequence, and the departure waypoint in the current new task semantic passage segment are read, and the speed switching action between the entry waypoint and the departure waypoint is rewritten as a speed execution action that is continuously connected according to the waypoint sequence; at the same time, the new task action category and task operation sequence of the current new task semantic passage segment are retained, so that the task content of the current candidate route is not changed due to the speed execution action rewriting; When the passage action identifier is hover position replacement, the hover operation waypoint and task operation sequence in the current new task semantic passage segment are read, and within the same task operation sequence, the hover operation waypoint that does not match the hover position replacement identifier in the candidate route set is selected as the replacement waypoint; after replacement, the current unit retains the new task action category of the current new task semantic passage segment, and replaces the original hover operation waypoint with the new hover operation waypoint, so that the task operation sequence continues to follow the sequence in the new UAV swarm task; When the passage action is marked as a reordering of the return sequence, the candidate route number to which the current new task semantic passage segment belongs, the UAV number participating in the same task operation sequence, and the task data return mission marker are read, and the new task semantic passage segment that hits the reordering of the return sequence is placed into the time-sharing passage sequence; the time-sharing passage sequence includes the UAV number, candidate route number, new task semantic passage segment number, and passage order. When the passage action is identified as formation convergence passage, the passage route trial unit reads the intra-segment waypoint sequence and the UAV numbers participating in the same segment from the current new mission semantic passage segment. It then rewrites the passage method where multiple UAVs simultaneously enter the same intra-segment waypoint sequence into a longitudinal continuous passage method arranged by UAV number. The longitudinal continuous passage method includes UAV number, entry waypoint, intra-segment waypoint sequence, exit waypoint, and passage order. The current processing retains the mission waypoint sequence and mission operation order in the candidate routes, only changing the passage organization method when multiple UAVs enter the same new mission semantic passage segment. Finally, the candidate route number, the new mission semantic passage segment number, the invoked low-altitude disturbance memory core, the passage handling action identifier, the rewritten waypoint sequence, the rewritten speed execution action, the replaced hovering operation waypoint, the time-sharing passage sequence, and the longitudinal continuous passage mode are encapsulated into a low-altitude memory trial route.

[0013] The source chain of the empty memory trial route is as follows: a new UAV cluster task generates a set of candidate routes, the set of candidate routes generates a new task semantic passage segment, the new task semantic passage segment matches the low-altitude disturbance memory core, the passage handling action identifier in the low-altitude disturbance memory core triggers the route trial processing, and the route trial processing generates the low-altitude memory trial route. Preferably, the cluster scheduling execution module includes an individual UAV operation profile arrangement unit and a low-altitude disturbance memory scheduling route package generation unit; The drone individual operation profile arrangement unit uses the drone number as an index to count the occurrence frequency of command execution bounce traces, hovering energy consumption trail textures, and communication attitude clippings under the same drone number, and generates a drone individual operation profile; the drone individual operation profile includes the drone number, the number of occurrences of command execution bounce traces, the number of occurrences of hovering energy consumption trail textures, and the number of occurrences of communication attitude clippings. Four types of drone execution sequences are generated based on the type of passage handling action: The first category is a speed-smooth execution sequence, with the UAV numbers arranged in ascending order of the number of bounce marks appearing during instruction execution; The second category is the hover position replacement execution sequence, with the drone numbers arranged in ascending order of the number of times the hover energy consumption trail texture appears; The third category is the return sequence reordering execution sequence, where the UAV numbers are arranged in ascending order according to the frequency of occurrence of the communication attitude-carried fragments. The fourth category is the formation convergence and passage execution sequence. First, the UAV numbers are arranged in ascending order of the number of times the bounce marks appear during command execution, and then the UAV numbers are arranged in ascending order of the number of times the communication attitude-carried fragments appear. When two drone IDs appear the same number of times in the same sorting field, they are sorted according to the character order of the drone IDs recorded in the new drone cluster task, and the drone execution sequence is output. Among them, the types of passage handling actions include smooth speed passage, hovering position replacement, return sequence rearrangement, and formation convergence passage; The low-altitude disturbance memory scheduling route package generation unit is based on the low-altitude memory trial route and the UAV execution sequence. It reads the passage handling action identifier in the low-altitude memory trial route and calls the UAV execution sequence according to the passage handling action identifier. When the passage handling action is marked as speed smooth passage, the speed smooth passage execution sequence is invoked, the drone number at the top of the sequence is written into the current new task semantic passage segment, and the rewritten speed execution action is bound. When the passage handling action is identified as hover position replacement, the hover position replacement execution sequence is invoked, the UAV number at the top of the sequence is written into the current new task semantic passage segment, and the replaced hover operation waypoint is bound; When the passage handling action is marked as "passing through order reordering", the return order reordering execution sequence is invoked, the drone number at the top of the sequence is written to the task data return location, and written to the time-sharing passage sequence according to the order; When the passage action is marked as "formation convergence passage", the formation convergence passage execution sequence is invoked, and the numbers of each UAV are written sequentially into the longitudinal continuous passage mode. The UAV number, candidate route number, new mission semantic passage segment number, invoked low-altitude disturbance memory core, passage handling action identifier, rewritten waypoint sequence, rewritten speed execution action, replaced hovering operation waypoint, time-sharing passage sequence, longitudinal continuous passage mode, and mission operation order are encapsulated into a low-altitude disturbance memory scheduling route package.

[0014] A method for intelligent planning and scheduling of low-altitude paths for unmanned aerial vehicle (UAV) swarms includes the following steps: Step 1: The mission route analysis module reads the mission route records, path planning instruction records, flight control execution records, hovering control records, unit segment energy consumption records, and communication link records from historical missions. It divides historical missions into mission semantic passage segments according to mission actions and generates mission semantic passage drafts. Step 2: The flight status recording and analysis module generates command execution rebound traces, hovering energy consumption trail textures, and communication attitude fragments based on the mission semantic communication draft. Step 3: The low-altitude disturbance memory map construction module performs individual UAV peeling, passage segment co-occurrence judgment, and action chain compression on command execution bounce traces, hovering energy consumption trail textures, and communication attitude-related fragments to generate a low-altitude disturbance memory kernel. Step 4: The path dynamic correction module divides the candidate routes of the new mission into new mission semantic passage segments, and generates low-altitude memory trial routes based on the matching results of the new mission semantic passage segments and the low-altitude disturbance memory core. Step 5: The cluster scheduling and execution module generates low-altitude disturbance memory scheduling route packages based on the low-altitude memory trial route and combined with the individual UAV operation profile and task operation sequence.

[0015] This invention provides a method and system for intelligent planning and scheduling of low-altitude paths for unmanned aerial vehicle (UAV) swarms, which has the following beneficial effects: (1) During system operation, the path dynamic correction module generates low-altitude memory trial routes based on the low-altitude disturbance memory kernel, and the cluster scheduling execution module generates low-altitude disturbance memory scheduling route packages by combining the individual UAV operation profile and task operation sequence. The low-altitude disturbance memory scheduling route package encapsulates the candidate route number, the new task semantic passage segment number, the passage handling action identifier, the rewritten waypoint sequence, the rewritten speed execution action, the replaced hovering operation waypoint, the time-sharing passage sequence, the longitudinal continuous passage method, and the task operation sequence as the basis for cluster execution. Through the current processing method, before a new task enters a low-altitude passage segment with historical disturbance memory, it can be scheduled according to smooth speed passage, hovering position replacement, return sequence rearrangement, or formation convergence passage. This allows the UAV cluster to be constrained by the low-altitude disturbance memory kernel when performing low-altitude inspection, shooting, security, and search tasks, improving the problems of repeated route correction, repeated hovering position adjustment, chaotic return sequence, and mismatched formation entry sequence in existing cluster tasks.

[0016] (2) By constructing a sequence of actions within the same segment, a sequence of planned actions, a sequence of executed actions, a sequence of hovering actions, a sequence of energy consumption changes, and a sequence of communication actions are generated within each mission semantic passage segment, and these sequences are encapsulated together into a sequence of actions within the same segment. Through the above processing, the planned heading actions, planned speed actions, actual heading actions, actual speed actions, attitude correction actions, hovering actions, energy consumption change records per unit segment, and communication link jitter records can be arranged according to the recording time field within the same mission semantic passage segment, enabling the subsequent passage feature generation unit to analyze the sequential process between planned actions, flight control executed actions, hovering actions, energy consumption changes, and communication actions, thus solving the problem that existing technologies only view track deviations, energy consumption changes, or communication link changes separately, making it difficult to identify combined low-altitude passage anomalies.

[0017] (3) The candidate route set is processed by the route matching unit based on the route matching action identifier in the memory kernel matching result. For smooth speed passage, the current unit rewrites the speed switching action between the waypoint and the waypoint as a continuous speed execution action in the waypoint sequence; for hover position replacement, the current unit selects the hover operation waypoint that did not hit the hover position replacement identifier as the replacement waypoint within the same task operation sequence; for return sequence rearrangement, the current unit arranges the new task semantic passage segment that hits the return sequence rearrangement identifier into the time-sharing passage sequence; for formation convergence passage, the current unit rewrites the passage mode of multiple UAVs entering the same inner waypoint sequence at the same time as a longitudinal continuous passage mode arranged by UAV number. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the block diagram of an intelligent low-altitude path planning and scheduling system for unmanned aerial vehicle (UAV) swarms according to the present invention. Figure 2 This is a schematic diagram illustrating the steps of an intelligent low-altitude path planning and scheduling method for a drone swarm according to the present invention. Figure 3 This is a schematic diagram of the task semantic passage segment encapsulation and judgment process of the present invention; Figure 4 This is a schematic diagram of the passage handling action identifier generation and judgment process of the present invention. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0020] Example 1 This invention provides an intelligent low-altitude path planning and scheduling system for unmanned aerial vehicle (UAV) swarms. Please refer to [link / reference]. Figures 1 to 4 It includes a mission route analysis module, a flight status record parsing module, a low-altitude disturbance memory map construction module, a path dynamic correction module, and a cluster scheduling execution module; The mission route analysis module reads mission route records, path planning instruction records, flight control execution records, hovering control records, unit segment energy consumption records, and communication link records from historical missions. It divides historical missions into mission semantic passage segments according to mission actions and generates mission semantic passage drafts. The flight status recording and analysis module generates command execution rebound traces, hovering energy consumption trail textures, and communication attitude fragments based on the mission semantic communication draft. The low-altitude disturbance memory map construction module performs individual UAV peeling, passage segment co-occurrence judgment and action chain compression on command execution bounce traces, hovering energy consumption trail textures and communication attitude clips to generate a low-altitude disturbance memory kernel. The path dynamic correction module divides the candidate routes of the new mission into new mission semantic passage segments, and generates low-altitude memory trial routes based on the matching results of the new mission semantic passage segments and the low-altitude disturbance memory core. The cluster scheduling and execution module generates low-altitude disturbance memory scheduling route packages based on low-altitude memory trial flight routes, combined with individual UAV operation profiles and task operation sequences.

[0021] In this embodiment, historical mission flight path records, path planning instruction records, flight control execution records, hovering control records, unit segment energy consumption records, and communication link records are uniformly incorporated into the mission flight path analysis module for processing. These are then divided into mission semantic passage segments according to mission actions, generating mission semantic passage drafts. Through this current processing method, historical missions are no longer analyzed as entire flight paths, but rather as passage segments with mission action meanings. This allows flight paths, flight control executions, hovering operations, energy consumption changes, and communication links to be invoked by subsequent modules within the same mission semantic passage segment, solving the problem of scattered historical mission records in existing low-altitude path planning, which are difficult to incorporate into subsequent scheduling and judgment.

[0022] The flight status record analysis module generates command execution rebound traces, hovering energy consumption trail textures, and communication attitude fragments from the mission semantic passage draft. These three types of passage features are not ordinary track deviations, energy consumption values, or communication status records, but are formed by the sequential processes of planned actions, executed actions, hovering actions, energy consumption changes, and communication actions within the same mission semantic passage segment. Therefore, implicit passage phenomena such as repeated flight control corrections, continued energy consumption after hovering, and attitude corrections accompanying communication changes, which exist in passable locations such as gaps between buildings, under bridges, near billboards, and at the edges of tree canopies, can be converted into passage features that can be used for subsequent path planning.

[0023] This embodiment uses a low-altitude disturbance memory map construction module to perform individual UAV extraction, passage segment co-occurrence judgment, and action chain compression on command execution bounce traces, hovering energy consumption trail textures, and communication attitude-related fragments, generating a low-altitude disturbance memory kernel. Through this current processing method, passage characteristics repeatedly appearing in multiple mission semantic passage segments for the same UAV are categorized as individual UAV operational factors, while passage characteristics commonly appearing in the same mission semantic passage segment for multiple UAVs are categorized as passage segment environmental disturbance factors. This allows for the differentiation between the individual UAV's execution characteristics and low-altitude passage segment environmental disturbance factors before generating the low-altitude disturbance memory kernel, addressing the shortcomings of existing technologies that easily treat single-aircraft flight deviations directly as low-altitude area passage problems.

[0024] Example 2 Please refer to Figure 3 Specifically: the mission route analysis module includes a multi-source mission record arrangement unit and a mission semantic passage segment encapsulation unit; The multi-source mission record arrangement unit reads mission route records, path planning instruction records, flight control execution records, hovering control records, unit segment energy consumption records, and communication link records from historical missions, and extracts historical mission number, UAV number, record time field, waypoint number, and mission action field from the records; Based on the historical mission number and UAV number, the mission route records, path planning instruction records, flight control execution records, hovering control records, unit segment energy consumption records, and communication link records formed by the same UAV in the same historical mission are arranged according to the record time field to generate a single-aircraft mission execution time chain. After generating the single-machine task execution time chain, the path planning instruction content in the path planning instruction record, the flight control execution content in the flight control execution record, the hovering control content in the hovering control record, the unit segment energy consumption content in the unit segment energy consumption record, and the communication link content in the communication link record under the same waypoint number are grouped into the same waypoint execution segment. The task semantic passage segment encapsulation unit receives the single-machine task execution time chain and waypoint execution fragments, and marks the task action category for each waypoint execution fragment according to the task action fields in the path planning instruction record and the task route record; Locate the record position where the task action category changes in the single-machine task execution time chain, and use the task action category change position as the boundary of the passage segment; Waypoint execution segments that are located within the same passageway boundary and have the same task action category are merged into a single task semantic passageway. Multiple task semantic access segments formed by the same historical mission and the same drone are arranged according to the record time field to generate a task semantic access draft.

[0025] The flight status recording and parsing module includes a segment action sequence construction unit and a passage feature generation unit; The same action sequence construction unit reads multiple task semantic passage segments from the task semantic passage draft. Each task semantic passage segment contains historical task number, UAV number, task action category, entry action, departure action, path planning instruction segment, flight control execution segment, hovering control segment, unit segment energy consumption segment, and communication link segment. Using a single task semantic passage as the processing object, the path planning instruction fragment in the current task semantic passage is read, and the planning heading action, planning speed action, planning altitude action, planning hovering action and planning departure action in the path planning instruction fragment are arranged according to the recorded time field to form a planning action sequence; Read the flight control execution segment in the current task semantic passage, and arrange the actual heading action, actual speed action, actual altitude action, attitude correction action, heading recovery action and speed correction action in the flight control execution segment according to the recorded time field to form an execution action sequence; Read the hover control segment in the semantic passage of the current task, and arrange the hover entry action, hover hold action and hover exit action according to the recorded time field to form a hover action sequence; Read the unit segment energy consumption fragment in the current task semantic passage segment, arrange the unit segment energy consumption change records according to the record time field, and form an energy consumption change sequence; Read the communication link segments in the current task semantic passage segment, and arrange the communication link connection record, communication link interruption record, communication link recovery record and communication link jitter record according to the record time field to form a communication action sequence; The planning action sequence, execution action sequence, hovering action sequence, energy consumption change sequence, and communication action sequence are collectively encapsulated into a single action sequence.

[0026] The passage feature generation unit generates passage features within the same task semantic passage segment based on the same action sequence, namely instruction execution bounce trace, hovering energy consumption trail texture and communication posture clipping fragment. The method for obtaining the bounce trace of command execution is as follows: take the planned heading action and planned speed action in the planned action sequence as reference actions, and then find the action chain after the actual heading action or actual speed action is separated from the reference action in the executed action sequence. The action chain, in chronological order, includes: the action to detach from the reference action, the attitude correction action generated by the flight control execution segment, the heading correction action generated by the flight control execution segment, the speed correction action generated by the flight control execution segment, and the action to re-enter the reference action; the action chain, along with the historical mission number, UAV number, and mission action category of the mission semantic passage segment, is encapsulated as a command execution bounce trace; The method for obtaining hovering energy consumption trail texture is as follows: determine the hovering operation action chain based on the hovering entry action, hovering hold action and hovering departure action in the hovering action sequence, and then read the unit segment energy consumption change record located after the hovering departure action in the energy consumption change sequence. The energy consumption changes per unit segment of the hovering hold action, hovering departure action, and hovering departure action that continues with the passage action, as well as the speed recovery action in the sequence of actions, are encapsulated as hovering energy consumption trail textures. The method for obtaining communication attitude clips is as follows: search for communication link jitter records, communication link interruption records, or communication link recovery records in the communication action sequence, then read the execution action sequence during the communication action, and extract the attitude correction action, heading correction action, and speed correction action that occur in the same period. The communication link jitter record, communication link interruption record, or communication link recovery record, along with the attitude correction action, heading righting action, and speed correction action within the same period, are encapsulated into a communication attitude clip.

[0027] In this embodiment, a multi-source mission record organization unit organizes historical mission flight path records, path planning instruction records, flight control execution records, hovering control records, unit segment energy consumption records, and communication link records into a single-aircraft mission execution time chain according to the historical mission number, UAV number, and recording time field. Furthermore, path planning instruction content, flight control execution content, hovering control content, unit segment energy consumption content, and communication link content under the same waypoint number are grouped into the same waypoint execution segment. Through this processing, the mission planning content, flight control execution content, hovering control content, energy consumption change content, and communication link content, which were originally scattered in different records, are placed under the same historical mission, the same UAV, and the same waypoint number for subsequent analysis. This prevents historical mission records from remaining separately in their respective data systems for flight path, flight control, energy consumption, and communication, solving the problem that existing low-altitude path planning methods struggle to reuse integrated historical mission flight records.

[0028] Through the task semantic passage segment encapsulation unit, based on the task action fields in the path planning instruction record and task route record, the task action categories of waypoint execution segments are marked. The location where the task action category switches is used as the passage segment boundary. Waypoint execution segments with the same task action category and located within the same passage segment boundary are merged into task semantic passage segments. Through this processing, historical tasks are no longer analyzed as entire routes or individual waypoints, but are transformed into task semantic passage segments with task action meaning. This allows task actions such as straight-through, turning, hovering, photography, inspection, and return to base to be individually included in the subsequent flight status analysis process, solving the problem in existing technologies where low-altitude passage issues are difficult to bind to specific task actions.

[0029] This embodiment uses a traffic feature generation unit to generate command execution bounce traces, hovering energy consumption trail textures, and communication attitude entrainment fragments based on the same action sequence. The command execution bounce traces characterize the process of re-entering the planned action after the actual heading or speed action deviates from the planned action, through attitude correction, heading recovery, and speed correction actions. The hovering energy consumption trail texture characterizes the process where, after hovering hold and hover departure actions, the energy consumption change record per unit segment continues with the traffic action. The communication attitude entrainment fragments characterize the process where communication link jitter records, communication link interruption records, or communication link recovery records occur simultaneously with attitude correction, heading recovery, and speed correction actions.

[0030] Example 3 Please refer to Figure 4 Specifically: the low-altitude disturbance memory map construction module includes an individual factor stripping and passage segment co-occurrence unit and a disturbance action chain compression and memory core encapsulation unit; Individual factor stripping and the co-occurrence unit of the passage segment encapsulates instruction execution bounce traces, hovering energy consumption trail textures, and communication posture clipping fragments into passage feature groups; The access feature groups are arranged individually according to the drone number. Under the same drone number, the same access features that appear in at least two different mission semantic access segment identifiers are listed as individual drone operation factors. After separating the individual operational factors of the UAV, the remaining traffic features are analyzed for co-occurrence according to the traffic segment identifier of the mission semantics. For the same semantic passage segment identifier, read the instruction execution rebound traces, hovering energy consumption trail textures, and communication attitude fragments under different UAV numbers, and arrange them according to the recorded time field; When there are at least two different UAV numbers with the same type of passage feature under the same task semantic passage segment identifier, or when there are at least two different types of passage features with different UAV numbers arranged consecutively according to the record time field, the current task semantic passage segment identifier is listed as a passage segment environmental disturbance factor, and a passage segment co-occurrence record is generated.

[0031] The perturbation action chain compression and memory kernel encapsulation unit is based on the co-occurrence record of the passage segment and takes the semantic passage segment identifier of each task as the processing object to perform action chain compression processing on the passage features that participate in co-occurrence. In the action chain compression process, the action of leaving the reference, attitude correction, heading correction, speed correction and re-entering the reference in the instruction execution rebound trace are regarded as the first type of action node. The hovering hold action, hovering departure action, unit segment energy consumption change record and speed recovery action in the hovering energy consumption trail texture are regarded as the second type of action nodes; The communication link jitter record, communication link interruption record, communication link recovery record, attitude correction action, heading righting action, and speed correction action in the communication attitude clip are regarded as the third type of action node; Arrange the action nodes into a perturbation action chain according to the recorded time field; Based on the disturbance action chain, generate passage handling action indicators; When the disturbance action chain contains instruction execution bounce traces but does not contain hovering energy consumption trail textures and communication attitude fragments, the passage handling action is identified as speed smooth passage. When the disturbance action chain includes hover energy consumption trail texture, the passage action identifier is hover position replacement; When the disturbance action chain contains communication attitude clippings, the passage handling action is identified as a reordering of the transmission sequence; When the disturbance action chain contains both command execution bounce traces and communication attitude clips, and the passage segment co-occurrence record contains multiple UAV numbers, the passage handling action is identified as formation convergence passage; Finally, the task semantic passage segment identifier, passage segment co-occurrence record, and passage handling action identifier are encapsulated into a low-altitude disturbance memory kernel.

[0032] In this embodiment, through individual factor stripping and the co-occurrence unit of the passage segment, the command execution bounce trace, hovering energy consumption trail texture, and communication attitude fragments are first encapsulated into passage feature groups. Then, the passage feature groups are arranged individually according to the UAV number, and the same type of passage features appearing in at least two different mission semantic passage segment identifiers under the same UAV number are listed as individual UAV operation factors. Through the above processing, the passage features caused by the flight control execution characteristics, hovering execution characteristics, or communication execution characteristics of a single UAV can be separated before entering the passage segment environmental disturbance judgment, solving the deficiency in the prior art that easily misinterprets the execution performance of a single UAV as a low-altitude passage segment environmental problem.

[0033] After separating individual UAV operational factors, co-occurrence analysis is performed on the remaining traffic features according to the mission semantic traffic segment identifier. Under the same mission semantic traffic segment identifier, command execution bounce traces, hovering energy consumption trail textures, and communication attitude fragments corresponding to different UAV numbers are read. Through this processing, only when different UAV numbers exhibit the same type of traffic features under the same mission semantic traffic segment identifier, or when different UAV numbers exhibit different types of traffic features arranged consecutively according to the recording time field under the same mission semantic traffic segment identifier, is the current mission semantic traffic segment identifier listed as a traffic segment environmental disturbance factor, and a traffic segment co-occurrence record is generated. This processing method can distinguish the traffic problems jointly exhibited by multiple UAVs in passable locations such as building gaps, under bridges, near billboards, and at the edge of tree canopies from the occasional execution behavior of a single UAV, giving low-altitude traffic segment environmental disturbance factors a clear mission semantic traffic segment source and UAV number source.

[0034] This embodiment further generates passage handling action identifiers based on disturbance action chains, and encapsulates the task semantic passage segment identifier, passage segment co-occurrence record, and passage handling action identifier into a low-altitude disturbance memory kernel. Through this processing, command execution bounce traces can trigger speed smoothing passage, hovering energy consumption trail textures can trigger hovering position replacement, communication attitude entrainment fragments can trigger transmission order rearrangement, and command execution bounce traces and communication attitude entrainment fragments can trigger formation convergence passage when they co-occur under multiple UAV numbers. Thus, low-altitude passage anomalies that have occurred in historical missions are transformed into low-altitude disturbance memory kernels that can be called upon by subsequent path dynamic correction modules, enabling new missions to directly obtain the handling basis required for route matching and cluster scheduling when passing through the same task semantic passage segment, improving the problem in the prior art that historical flight anomalies are difficult to use for subsequent mission path planning.

[0035] Example 4 Please refer to Figure 1 Specifically: the path dynamic correction module includes a new task passage segment matching unit and a passage handling route trial matching unit; The new mission passage matching unit receives new UAV swarm missions and low-altitude disturbance memory cores; The new drone swarm missions include mission waypoints, mission actions, and mission sequence. First, a set of candidate routes is generated based on the task waypoints, task actions, and task operation sequence in the new UAV swarm mission. Each candidate route includes a candidate route number, a sequence of task waypoints, a sequence of task actions, and a task operation sequence. The candidate route set is divided according to the sequence of task actions, and the waypoints continuously covered by the same task action are encapsulated into a new task semantic passage segment. The new task semantic passage is matched with the low-altitude perturbation memory kernel, as follows: Simultaneously, read the new mission action category, waypoint entry, waypoint exit, intra-segment waypoint sequence, and mission operation sequence from the new mission semantic passage segment, and read the mission semantic passage segment identifier and passage segment co-occurrence record from the low-altitude disturbance memory kernel; The read content is compared with similar fields to obtain the memory kernel matching results. The memory kernel matching results include candidate route number, new task semantic passage segment number, matched low-altitude disturbance memory kernel, passage segment co-occurrence record and passage handling action identifier.

[0036] The passage handling route matching unit performs route matching processing on the candidate route set based on the candidate route number, new task semantic passage segment number and passage handling action identifier in the memory kernel matching results; When the passage action is marked as smooth passage, read the entry waypoint, the intra-segment waypoint sequence and the departure waypoint in the current new task semantic passage segment, and rewrite the speed switching action between the entry waypoint and the departure waypoint as a speed execution action that is continuously connected according to the waypoint sequence. When the passage handling action is identified as hover position replacement, read the hover operation waypoints and task operation sequence in the current new task semantic passage segment, and select the hover operation waypoints that do not match the hover position replacement identifier in the candidate route set as replacement waypoints within the same task operation sequence. When the passage handling action is identified as a return sequence reordering, the candidate route number to which the current new task semantic passage segment belongs, the UAV number participating in the same task operation sequence, and the task data return mission marker are read, and the new task semantic passage segment that hits the return sequence reordering marker is placed into the time-sharing passage sequence. When the passage action is marked as formation convergence passage, the passage action route trial unit reads the intra-segment waypoint sequence and the UAV numbers participating in the same segment passage in the current new task semantic passage segment, and rewrites the passage mode of multiple UAVs entering the same intra-segment waypoint sequence at the same time into a longitudinal continuous passage mode arranged by UAV number. Finally, the candidate route number, the new mission semantic passage segment number, the invoked low-altitude disturbance memory core, the passage handling action identifier, the rewritten waypoint sequence, the rewritten speed execution action, the replaced hovering operation waypoint, the time-sharing passage sequence, and the longitudinal continuous passage mode are encapsulated into a low-altitude memory trial route.

[0037] In this embodiment, a new task passage segment matching unit generates a candidate route set from the task waypoints, task actions, and task operation sequence in a new UAV swarm task. This candidate route set is then divided into new task semantic passage segments according to the task action sequence. Through this process, the new task no longer directly enters the path correction process as the entire candidate route. Instead, it is first broken down into passage segments with task action meanings, allowing subsequent path corrections to accurately apply to the new task semantic passage segments that hit the low-altitude disturbance memory core. This solves the problem in existing technologies where uniform adjustments to the entire route lead to unclear task action boundaries.

[0038] By matching new mission semantic passage segments with the low-altitude disturbance memory kernel, the co-occurrence records of passage segments and passage handling action identifiers formed in historical missions can be incorporated into the new mission path correction process. The new mission passage segment matching unit simultaneously reads the new mission action category, entry waypoint, departure waypoint, intra-segment waypoint sequence, and mission operation order from the new mission semantic passage segment, and reads the mission semantic passage segment identifiers and passage segment co-occurrence records from the low-altitude disturbance memory kernel to generate the memory kernel matching result. Through this process, historically occurring passage anomalies in passable surface locations such as building gaps, under bridges, near billboards, and tree canopy edges can be invoked when a new mission passes through the same mission semantic passage segment, solving the problem that historical flight anomalies are difficult to participate in new mission route processing in existing path planning.

[0039] This embodiment ultimately encapsulates the candidate route number, the new mission semantic passage segment number, the invoked low-altitude disturbance memory core, the passage handling action identifier, the rewritten waypoint sequence, the rewritten speed execution action, the replaced hovering operation waypoint, the time-sharing passage sequence, and the longitudinal continuous passage method into a low-altitude memory trial route. Through this encapsulation method, the output of the path dynamic correction module simultaneously retains the source of the candidate route, the source of the new mission semantic passage segment, the source of the historical disturbance memory, and the route trial processing result. This allows the subsequent cluster scheduling execution module to directly generate low-altitude disturbance memory scheduling route packages based on the low-altitude memory trial route, improving the problems of repeated route corrections, repeated hovering position adjustments, chaotic mission data return order, and mismatched formation entry order in existing UAV cluster low-altitude missions.

[0040] Example 5 Please refer to Figure 1 Specifically: the cluster scheduling and execution module includes an individual UAV operation profile arrangement unit and a low-altitude disturbance memory scheduling route package generation unit; The drone individual operation profile arrangement unit uses the drone number as an index to count the occurrence frequency of command execution rebound traces, hovering energy consumption trail textures, and communication attitude clips under the same drone number, and generates a drone individual operation profile. Four types of drone execution sequences are generated based on the type of passage handling action: The first category is a speed-smooth execution sequence, with the UAV numbers arranged in ascending order of the number of bounce marks appearing during instruction execution; The second category is the hover position replacement execution sequence, with the drone numbers arranged in ascending order of the number of times the hover energy consumption trail texture appears; The third category is the return sequence reordering execution sequence, where the UAV numbers are arranged in ascending order according to the frequency of occurrence of the communication attitude-carried fragments. The fourth category is the formation convergence and passage execution sequence. First, the UAV numbers are arranged in ascending order of the number of times the bounce marks appear during command execution, and then the UAV numbers are arranged in ascending order of the number of times the communication attitude-carried fragments appear. When two drone IDs appear the same number of times in the same sorting field, they are sorted according to the character order of the drone IDs recorded in the new drone cluster task, and the drone execution sequence is output. The low-altitude disturbance memory scheduling route package generation unit is based on the low-altitude memory trial route and the UAV execution sequence. It reads the passage handling action identifier in the low-altitude memory trial route and calls the UAV execution sequence according to the passage handling action identifier. When the passage handling action is marked as speed smooth passage, the speed smooth passage execution sequence is invoked, the drone number at the top of the sequence is written into the current new task semantic passage segment, and the rewritten speed execution action is bound. When the passage handling action is identified as hover position replacement, the hover position replacement execution sequence is invoked, the UAV number at the top of the sequence is written into the current new task semantic passage segment, and the replaced hover operation waypoint is bound; When the passage handling action is marked as "passing through order reordering", the return order reordering execution sequence is invoked, the drone number at the top of the sequence is written to the task data return location, and written to the time-sharing passage sequence according to the order; When the passage action is marked as "formation convergence passage", the formation convergence passage execution sequence is invoked, and the numbers of each UAV are written sequentially into the longitudinal continuous passage mode. The UAV number, candidate route number, new mission semantic passage segment number, invoked low-altitude disturbance memory core, passage handling action identifier, rewritten waypoint sequence, rewritten speed execution action, replaced hovering operation waypoint, time-sharing passage sequence, longitudinal continuous passage mode, and mission operation order are encapsulated into a low-altitude disturbance memory scheduling route package.

[0041] In this embodiment, the individual drone operation profile orchestration unit uses the drone number as an index to count the frequency of occurrence of command execution bounce traces, hovering energy consumption trail textures, and communication attitude fragments under the same drone number, generating an individual drone operation profile. Through the above processing, common features identified as individual drone operation factors in historical tasks can be transformed into drone selection criteria for new task scheduling. This allows cluster scheduling to no longer rely solely on drone idle status, current location, or remaining battery power, but rather to orchestrate based on each drone's historical performance in command execution, hovering operations, and communication transmission, solving the problem of difficulty in handling task execution differences caused by homogeneous allocation in existing cluster scheduling.

[0042] Based on the type of passage handling action, execution sequences for speed smoothing passage, hovering position replacement, return sequence rearrangement, and formation convergence passage are generated. UAVs are then numbered and arranged according to the frequency of command execution bounce traces, hovering energy consumption trail textures, and communication attitude fragments. Through this processing, different passage handling actions can call upon UAV execution sequences that match their execution requirements. Speed ​​smoothing passage is prioritized for UAVs with higher frequencies of command execution bounce traces, hovering position replacement is prioritized for UAVs with higher frequencies of hovering energy consumption trail textures, return sequence rearrangement is prioritized for UAVs with higher frequencies of communication attitude fragments, and formation convergence passage considers both command execution bounce trace and communication attitude fragment occurrences. This solves the problem in existing cluster scheduling where the same sorting rule is difficult to adapt to different low-altitude disturbance handling actions.

[0043] This embodiment ultimately encapsulates the UAV number, candidate route number, new mission semantic passage segment number, invoked low-altitude disturbance memory kernel, passage handling action identifier, rewritten waypoint sequence, rewritten speed execution action, replaced hovering operation waypoint, time-sharing passage sequence, longitudinal continuous passage mode, and mission operation sequence into a low-altitude disturbance memory scheduling route package. Through the above encapsulation method, the low-altitude disturbance memory kernel formed in historical missions, the individual UAV operation profile, and the mission operation sequence of the new mission are combined into a scheduling result executable by the UAV swarm. This enables the UAV swarm to execute missions according to the determined UAV number, route content, passage sequence, hovering position, and formation passage mode when passing through low-altitude passage segments with implicit passage problems. This improves the problems of repeated route corrections, repeated hovering position adjustments, chaotic mission data return order, and mismatched formation entry order in existing low-altitude swarm missions.

[0044] Example 6 A method for intelligent planning and scheduling of low-altitude paths for UAV swarms, please refer to... Figure 2 Specifically, it includes the following steps: Step 1: The mission route analysis module reads the mission route records, path planning instruction records, flight control execution records, hovering control records, unit segment energy consumption records, and communication link records from historical missions. It divides historical missions into mission semantic passage segments according to mission actions and generates mission semantic passage drafts. Step 2: The flight status recording and analysis module generates command execution rebound traces, hovering energy consumption trail textures, and communication attitude fragments based on the mission semantic communication draft. Step 3: The low-altitude disturbance memory map construction module performs individual UAV peeling, passage segment co-occurrence judgment, and action chain compression on command execution bounce traces, hovering energy consumption trail textures, and communication attitude-related fragments to generate a low-altitude disturbance memory kernel. Step 4: The path dynamic correction module divides the candidate routes of the new mission into new mission semantic passage segments, and generates low-altitude memory trial routes based on the matching results of the new mission semantic passage segments and the low-altitude disturbance memory core. Step 5: The cluster scheduling and execution module generates low-altitude disturbance memory scheduling route packages based on the low-altitude memory trial route and combined with the individual UAV operation profile and task operation sequence.

[0045] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended technical solutions and their equivalents.

Claims

1. A low-altitude path intelligent planning and scheduling system for unmanned aerial vehicle (UAV) swarms, characterized in that: It includes a mission route analysis module, a flight status record parsing module, a low-altitude disturbance memory map construction module, a path dynamic correction module, and a cluster scheduling execution module; The mission route analysis module reads mission route records, path planning instruction records, flight control execution records, hovering control records, unit segment energy consumption records, and communication link records from historical missions. It divides historical missions into mission semantic passage segments according to mission actions and generates mission semantic passage drafts. The flight status recording and analysis module generates command execution rebound traces, hovering energy consumption trail textures, and communication attitude fragments based on the mission semantic communication draft. The low-altitude disturbance memory map construction module performs individual UAV peeling, passage segment co-occurrence judgment and action chain compression on command execution bounce traces, hovering energy consumption trail textures and communication attitude clips to generate a low-altitude disturbance memory kernel. The path dynamic correction module divides the candidate routes of the new mission into new mission semantic passage segments, and generates low-altitude memory trial routes based on the matching results of the new mission semantic passage segments and the low-altitude disturbance memory core. The cluster scheduling and execution module generates low-altitude disturbance memory scheduling route packages based on low-altitude memory trial flight routes, combined with individual UAV operation profiles and task operation sequences.

2. The intelligent planning and scheduling system for low-altitude paths of unmanned aerial vehicle (UAV) swarms according to claim 1, characterized in that: The mission route analysis module includes a multi-source mission record arrangement unit and a mission semantic passage segment encapsulation unit; The multi-source mission record arrangement unit reads mission route records, path planning instruction records, flight control execution records, hovering control records, unit segment energy consumption records, and communication link records from historical missions, and extracts historical mission number, UAV number, record time field, waypoint number, and mission action field from the records; Based on the historical mission number and UAV number, the mission route records, path planning instruction records, flight control execution records, hovering control records, unit segment energy consumption records, and communication link records formed by the same UAV in the same historical mission are arranged according to the record time field to generate a single-aircraft mission execution time chain. After generating the single-machine task execution time chain, the path planning instruction content in the path planning instruction record, the flight control execution content in the flight control execution record, the hovering control content in the hovering control record, the unit segment energy consumption content in the unit segment energy consumption record, and the communication link content in the communication link record under the same waypoint number are grouped into the same waypoint execution segment. The task semantic passage segment encapsulation unit receives the single-machine task execution time chain and waypoint execution fragments, and marks the task action category for each waypoint execution fragment according to the task action fields in the path planning instruction record and the task route record; Locate the record position where the task action category changes in the single-machine task execution time chain, and use the task action category change position as the boundary of the passage segment; Waypoint execution segments that are located within the same passageway boundary and have the same task action category are merged into a single task semantic passageway. Multiple task semantic access segments formed by the same historical mission and the same drone are arranged according to the record time field to generate a task semantic access draft.

3. The intelligent planning and scheduling system for low-altitude paths of unmanned aerial vehicle (UAV) swarms according to claim 2, characterized in that: The flight status recording and parsing module includes a segment action sequence construction unit and a passage feature generation unit; The same action sequence construction unit reads multiple task semantic passage segments from the task semantic passage draft. Each task semantic passage segment contains historical task number, UAV number, task action category, entry action, departure action, path planning instruction segment, flight control execution segment, hovering control segment, unit segment energy consumption segment, and communication link segment. Using a single task semantic passage as the processing object, the path planning instruction fragment in the current task semantic passage is read, and the planning heading action, planning speed action, planning altitude action, planning hovering action and planning departure action in the path planning instruction fragment are arranged according to the recorded time field to form a planning action sequence; Read the flight control execution segment in the current task semantic passage, and arrange the actual heading action, actual speed action, actual altitude action, attitude correction action, heading recovery action and speed correction action in the flight control execution segment according to the recorded time field to form an execution action sequence; Read the hover control segment in the semantic passage of the current task, and arrange the hover entry action, hover hold action and hover exit action according to the recorded time field to form a hover action sequence; Read the unit segment energy consumption fragment in the current task semantic passage segment, arrange the unit segment energy consumption change records according to the record time field, and form an energy consumption change sequence; Read the communication link segments in the current task semantic passage segment, and arrange the communication link connection record, communication link interruption record, communication link recovery record and communication link jitter record according to the record time field to form a communication action sequence; The planning action sequence, execution action sequence, hovering action sequence, energy consumption change sequence, and communication action sequence are collectively encapsulated into a single action sequence.

4. The intelligent planning and scheduling system for low-altitude paths of unmanned aerial vehicle (UAV) swarms according to claim 3, characterized in that: The passage feature generation unit generates passage features within the same task semantic passage segment based on the same action sequence, namely instruction execution bounce trace, hovering energy consumption trail texture and communication posture clipping fragment. The method for obtaining the bounce trace of command execution is as follows: take the planned heading action and planned speed action in the planned action sequence as reference actions, and then find the action chain after the actual heading action or actual speed action is separated from the reference action in the executed action sequence. The action chain, in chronological order, includes: the action to detach from the reference action, the attitude correction action generated by the flight control execution segment, the heading correction action generated by the flight control execution segment, the speed correction action generated by the flight control execution segment, and the action to re-enter the reference action; the action chain, along with the historical mission number, UAV number, and mission action category of the mission semantic passage segment, is encapsulated as a command execution bounce trace; The method for obtaining hovering energy consumption trail texture is as follows: determine the hovering operation action chain based on the hovering entry action, hovering hold action and hovering departure action in the hovering action sequence, and then read the unit segment energy consumption change record located after the hovering departure action in the energy consumption change sequence. The energy consumption changes per unit segment of the hovering hold action, hovering departure action, and hovering departure action that continues with the passage action, as well as the speed recovery action in the sequence of actions, are encapsulated as hovering energy consumption trail textures. The method for obtaining communication attitude clips is as follows: search for communication link jitter records, communication link interruption records, or communication link recovery records in the communication action sequence, then read the execution action sequence during the communication action, and extract the attitude correction action, heading correction action, and speed correction action that occur in the same period. The communication link jitter record, communication link interruption record, or communication link recovery record, along with the attitude correction action, heading righting action, and speed correction action within the same period, are encapsulated into a communication attitude clip.

5. The intelligent planning and scheduling system for low-altitude paths of unmanned aerial vehicle (UAV) swarms according to claim 4, characterized in that: The low-altitude disturbance memory map construction module includes an individual factor stripping and passage segment co-occurrence unit and a disturbance action chain compression and memory core encapsulation unit; Individual factor stripping and passage co-occurrence unit encapsulates instruction execution bounce traces, hovering energy consumption trail textures and communication posture clipping fragments into passage feature groups; The access feature groups are arranged individually according to the drone number. Under the same drone number, the same access features that appear in at least two different mission semantic access segment identifiers are listed as individual drone operation factors. After separating the individual operational factors of the UAV, the remaining traffic features are analyzed for co-occurrence according to the traffic segment identifier of the mission semantics. For the same semantic passage segment identifier, read the instruction execution rebound traces, hovering energy consumption trail textures, and communication attitude fragments under different UAV numbers, and arrange them according to the recorded time field; When there are at least two different UAV numbers with the same type of passage feature under the same task semantic passage segment identifier, or when there are at least two different types of passage features with different UAV numbers arranged consecutively according to the record time field, the current task semantic passage segment identifier is listed as a passage segment environmental disturbance factor, and a passage segment co-occurrence record is generated.

6. The intelligent planning and scheduling system for low-altitude paths of unmanned aerial vehicle (UAV) swarms according to claim 5, characterized in that: The perturbation action chain compression and memory kernel encapsulation unit is based on the co-occurrence record of the passage segment and takes the semantic passage segment identifier of each task as the processing object to perform action chain compression processing on the passage features that participate in co-occurrence. In the action chain compression process, the action of leaving the reference, attitude correction, heading correction, speed correction and re-entering the reference in the instruction execution rebound trace are regarded as the first type of action node. The hovering hold action, hovering departure action, unit segment energy consumption change record and speed recovery action in the hovering energy consumption trail texture are regarded as the second type of action nodes; The communication link jitter record, communication link interruption record, communication link recovery record, attitude correction action, heading correction action, and speed correction action in the communication attitude clip are regarded as the third type of action node; Arrange the action nodes into a perturbation action chain according to the recorded time field; Based on the disturbance action chain, generate passage handling action indicators; When the disturbance action chain contains instruction execution bounce traces but does not contain hovering energy consumption trail textures or communication attitude fragments, the passage handling action is identified as speed smooth passage. When the disturbance action chain includes hover energy consumption trail texture, the passage action identifier is hover position replacement; When the disturbance action chain contains communication attitude clippings, the passage handling action is identified as a reordering of the transmission sequence. When the disturbance action chain contains both command execution bounce traces and communication attitude clips, and the passage segment co-occurrence record contains multiple UAV numbers, the passage handling action is identified as formation convergence passage; Finally, the task semantic passage segment identifier, passage segment co-occurrence record, and passage handling action identifier are encapsulated into a low-altitude disturbance memory kernel.

7. The intelligent planning and scheduling system for low-altitude paths of unmanned aerial vehicle (UAV) swarms according to claim 6, characterized in that: The path dynamic correction module includes a new task passage segment matching unit and a passage handling route trial matching unit; The new mission passage matching unit receives new UAV swarm missions and low-altitude disturbance memory cores; The new drone swarm missions include mission waypoints, mission actions, and mission sequence. First, a set of candidate routes is generated based on the task waypoints, task actions, and task operation sequence in the new UAV swarm mission. Each candidate route includes a candidate route number, a sequence of task waypoints, a sequence of task actions, and a task operation sequence. The candidate route set is divided according to the sequence of task actions, and the waypoints continuously covered by the same task action are encapsulated into a new task semantic passage segment. The new task semantic passage is matched with the low-altitude perturbation memory kernel, as follows: Simultaneously, read the new mission action category, waypoint entry, waypoint exit, intra-segment waypoint sequence, and mission operation sequence from the new mission semantic passage segment, and read the mission semantic passage segment identifier and passage segment co-occurrence record from the low-altitude disturbance memory kernel; The read content is compared with similar fields to obtain the memory kernel matching results. The memory kernel matching results include candidate route number, new task semantic passage segment number, matched low-altitude disturbance memory kernel, passage segment co-occurrence record and passage handling action identifier.

8. The intelligent planning and scheduling system for low-altitude paths of unmanned aerial vehicle (UAV) swarms according to claim 7, characterized in that: The passage handling route matching unit performs route matching processing on the candidate route set based on the candidate route number, new task semantic passage segment number and passage handling action identifier in the memory kernel matching results; When the passage action is marked as smooth passage, read the entry waypoint, the intra-segment waypoint sequence and the departure waypoint in the current new task semantic passage segment, and rewrite the speed switching action between the entry waypoint and the departure waypoint as a speed execution action that is continuously connected according to the waypoint sequence. When the passage handling action is identified as hover position replacement, read the hover operation waypoints and task operation sequence in the current new task semantic passage segment, and select the hover operation waypoints that do not match the hover position replacement identifier in the candidate route set as replacement waypoints within the same task operation sequence. When the passage handling action is identified as a return sequence reordering, the candidate route number to which the current new task semantic passage segment belongs, the UAV number participating in the same task operation sequence, and the task data return mission marker are read, and the new task semantic passage segment that hits the return sequence reordering marker is placed into the time-sharing passage sequence. When the passage action is marked as formation convergence passage, the passage action route trial unit reads the intra-segment waypoint sequence and the UAV numbers participating in the same segment passage in the current new task semantic passage segment, and rewrites the passage mode of multiple UAVs entering the same intra-segment waypoint sequence at the same time into a longitudinal continuous passage mode arranged by UAV number. Finally, the candidate route number, the new mission semantic passage segment number, the invoked low-altitude disturbance memory core, the passage handling action identifier, the rewritten waypoint sequence, the rewritten speed execution action, the replaced hovering operation waypoint, the time-sharing passage sequence, and the longitudinal continuous passage mode are encapsulated into a low-altitude memory trial route.

9. The intelligent planning and scheduling system for low-altitude paths of unmanned aerial vehicle (UAV) swarms according to claim 8, characterized in that: The cluster scheduling and execution module includes an individual UAV operation profile arrangement unit and a low-altitude disturbance memory scheduling route package generation unit; The drone individual operation profile arrangement unit uses the drone number as an index to count the occurrence frequency of command execution rebound traces, hovering energy consumption trail textures, and communication attitude clips under the same drone number, and generates a drone individual operation profile. Four types of drone execution sequences are generated based on the type of passage handling action: The first category is a speed-smooth execution sequence, with the UAV numbers arranged in ascending order of the number of bounce marks appearing during instruction execution; The second category is the hover position replacement execution sequence, with the drone numbers arranged in ascending order of the number of times the hover energy consumption trail texture appears; The third category is the return sequence reordering execution sequence, where the UAV numbers are arranged in ascending order according to the frequency of occurrence of the communication attitude-carried fragments. The fourth category is the formation convergence and passage execution sequence. First, the UAV numbers are arranged in ascending order of the number of times the bounce marks appear during command execution, and then the UAV numbers are arranged in ascending order of the number of times the communication attitude-carried fragments appear. When two drone IDs appear the same number of times in the same sorting field, they are sorted according to the character order of the drone IDs recorded in the new drone cluster task, and the drone execution sequence is output. The low-altitude disturbance memory scheduling route package generation unit is based on the low-altitude memory trial route and the UAV execution sequence. It reads the passage handling action identifier in the low-altitude memory trial route and calls the UAV execution sequence according to the passage handling action identifier. When the passage handling action is marked as speed smooth passage, the speed smooth passage execution sequence is invoked, the drone number at the top of the sequence is written into the current new task semantic passage segment, and the rewritten speed execution action is bound. When the passage handling action is identified as hover position replacement, the hover position replacement execution sequence is invoked, the UAV number at the top of the sequence is written into the current new task semantic passage segment, and the replaced hover operation waypoint is bound; When the passage handling action is marked as "passing through order reordering", the return order reordering execution sequence is invoked, the drone number at the top of the sequence is written to the task data return location, and written to the time-sharing passage sequence according to the order; When the passage action is marked as "formation convergence passage", the formation convergence passage execution sequence is invoked, and the numbers of each UAV are written sequentially into the longitudinal continuous passage mode. The UAV number, candidate route number, new mission semantic passage segment number, invoked low-altitude disturbance memory core, passage handling action identifier, rewritten waypoint sequence, rewritten speed execution action, replaced hovering operation waypoint, time-sharing passage sequence, longitudinal continuous passage mode, and mission operation order are encapsulated into a low-altitude disturbance memory scheduling route package.

10. A method for intelligent planning and scheduling of low-altitude paths for unmanned aerial vehicle (UAV) swarms, applied to the intelligent planning and scheduling system for low-altitude paths of UAV swarms as described in any one of claims 1 to 9, characterized in that: Includes the following steps: Step 1: The mission route analysis module reads the mission route records, path planning instruction records, flight control execution records, hovering control records, unit segment energy consumption records, and communication link records from historical missions. It divides historical missions into mission semantic passage segments according to mission actions and generates mission semantic passage drafts. Step 2: The flight status recording and analysis module generates command execution rebound traces, hovering energy consumption trail textures, and communication attitude fragments based on the mission semantic communication draft. Step 3: The low-altitude disturbance memory map construction module performs individual UAV peeling, passage segment co-occurrence judgment, and action chain compression on command execution bounce traces, hovering energy consumption trail textures, and communication attitude-related fragments to generate a low-altitude disturbance memory kernel. Step 4: The path dynamic correction module divides the candidate routes of the new mission into new mission semantic passage segments, and generates low-altitude memory trial routes based on the matching results of the new mission semantic passage segments and the low-altitude disturbance memory core. Step 5: The cluster scheduling and execution module generates low-altitude disturbance memory scheduling route packages based on the low-altitude memory trial route and combined with the individual UAV operation profile and task operation sequence.

Images