An unmanned device cluster cooperative intelligent control system

Abstract

The present application relates to unmanned equipment cluster cooperative control technical field, specifically relates to a kind of unmanned equipment cluster cooperative intelligent control system, including residual physical action analysis module obtains unmanned equipment cluster multi-source state data, identifies the residual physical action after control instruction termination, generates cluster control action residual atlas;The control right succession boundary construction module is based on the residual atlas, identifies control action non-repayment section, divides task succession type and constructs control right succession boundary diagram;The phase locking cooperative control module is based on the boundary diagram, and the cooperative control phase is divided, and the responsibility unmanned equipment and the release condition are configured, and phase locking control right package is generated, and the ordered migration of control right is realized.The present application realizes the dynamic physical constraint-oriented safe cooperative execution and continuous control right migration of unmanned equipment cluster, improves cluster control stability and task execution reliability.

Description

Technical Field

[0001] This invention relates to the field of unmanned equipment cluster collaborative control technology, and in particular to an unmanned equipment cluster collaborative intelligent control system. Background Technology

[0002] With the widespread application of intelligent unmanned equipment such as drones, unmanned vehicles, and unmanned ships in fields such as inspection, logistics transportation, emergency rescue, and collaborative operations, unmanned equipment swarms are gradually evolving from single-device independent operation to multi-device collaborative execution of complex tasks. During swarm task execution, multiple unmanned devices need to dynamically divide labor, coordinate paths, and hand over control according to task requirements to ensure the continuous completion of the overall task. However, during the execution of control commands, the physical effects of unmanned equipment, such as motion inertia, braking processes, mechanical vibrations, load release, wake disturbances, and equipment space occupancy, do not immediately disappear after the control command terminates. Instead, they continue to act on themselves or neighboring unmanned devices for a period of time, forming residual physical effects with temporal continuity and spatial diffusion. Therefore, accurately identifying and utilizing the aforementioned residual physical effects to achieve safe and efficient collaborative control of unmanned equipment swarms has become an important research direction in the field of intelligent control of unmanned swarms.

[0003] Existing collaborative control methods for unmanned equipment swarms typically rely on current equipment status, task priorities, path planning results, or real-time communication information for task allocation and control decisions. They lack systematic analysis of residual physical effects that persist after historical control actions have ended, making it impossible to accurately determine whether the release of control by a preceding unmanned equipment still affects subsequent equipment. Furthermore, existing methods often treat task succession as a direct switchover process under ideal conditions, failing to adequately consider the control transfer boundaries formed by factors such as motion inertia, braking delays, load release delays, mechanical vibrations, wake disturbances, and space occupancy. This makes unmanned equipment prone to task succession conflicts, premature control switching, and inter-equipment interference in complex dynamic environments, reducing the stability and security of collaborative execution in unmanned equipment swarms. Summary of the Invention

[0004] This invention provides a collaborative intelligent control system for unmanned equipment clusters. By acquiring historical control commands and multi-source state information of the unmanned equipment cluster, it identifies the residual physical effects after the termination of control commands and constructs a residual map of cluster control effects. Furthermore, based on the spatial influence range and duration of the residual physical effects, it analyzes the unpaid constraints generated by historical control actions on the execution of subsequent tasks and constructs a control right succession boundary map. On this basis, the cluster tasks are divided into multiple collaborative control phases, and the responsible unmanned equipment, the conditions for the release of preceding physical effects, and the allowed execution actions are configured to generate phase-locked control right packages. This enables the orderly migration of unmanned equipment cluster control rights according to physical constraint boundaries, thereby improving the safety, continuity, and reliability of collaborative control of unmanned equipment clusters in complex environments.

[0005] A collaborative intelligent control system for unmanned equipment clusters includes a residual physical effect analysis module, a control authority succession boundary construction module, and a phase-locked collaborative control module, wherein;

[0006] The residual physical action analysis module acquires the historical control commands, equipment motion status, actuator status, load status, environmental disturbance status, and relative spatial relationships between equipment of the unmanned equipment cluster. It identifies the residual physical actions that continue to exist after the control commands are terminated, including motion inertia, braking delay, load delay release, mechanical vibration, wake disturbance, and space occupancy. It determines the residual physical action spatial region, duration, and affected unmanned equipment for each residual physical action and generates a cluster control action residual map.

[0007] The control right succession boundary construction module, based on the cluster control action residual map, spatiotemporally correlates the cluster tasks to be executed with each residual physical action, identifies unpaid control action segments that still constrain unmanned equipment due to historical control actions, divides the cluster tasks to be executed into prohibited succession task segments, conditional succession task segments, and direct succession task segments, and constructs a control right succession boundary map based on the allowed entry conditions and succession order of each task segment.

[0008] The phase-locked collaborative control module divides the cluster task to be executed into multiple collaborative control phases based on the control right succession boundary diagram. It configures the responsible unmanned equipment, the release condition of the preceding physical action, and the allowed execution action for each collaborative control phase, and generates a phase-locked control right package. The responsible unmanned equipment obtains the execution permission of the current collaborative control phase after satisfying the corresponding release condition of the preceding physical action, and releases the execution permission after completing the current collaborative control phase.

[0009] Optionally, the residual physical effect analysis module includes:

[0010] Multi-source state synchronous acquisition: Acquire historical control commands, equipment motion status, actuator status, load application status, environmental disturbance status, and relative spatial relationships between equipment for each unmanned equipment in the unmanned equipment cluster. Then, perform time-series alignment according to the unmanned equipment identifier and unified timestamp. The historical control commands, equipment motion status, actuator status, load application status, environmental disturbance status, and relative spatial relationships between equipment that have completed time-series alignment are collected into multi-source state time-series data.

[0011] Residual physical effects identification: Based on the multi-source state time series data, taking the termination time of each historical control command as the starting point of analysis, the continuous change analysis of the equipment motion state, actuator state, load action state and environmental disturbance state after the command termination is performed. When the corresponding state has not recovered to a stable state after the command termination, the continuous change process is identified as a residual physical effect segment, and is marked as motion inertia, braking delay, load delayed release, mechanical vibration, wake disturbance or space occupation according to the source of the state.

[0012] Residual Map Construction: For each residual physical action segment, the corresponding residual physical action spatial region is determined, and it is determined whether the residual physical action spatial region overlaps with the position trajectory of each unmanned device in the unmanned device cluster except for the corresponding unmanned device. Unmanned devices whose position trajectory overlaps with the residual physical action spatial region are identified as affected unmanned devices. Using each residual physical action segment as a node, and using the historical control command source relationship, the residual physical action spatial region overlap relationship, the continuous stage inheritance relationship, and the unmanned device influence relationship as edges, each residual physical action spatial region, continuous stage, and affected unmanned device are associated to generate a cluster control action residual map.

[0013] Optionally, the identification of residual physical effects includes:

[0014] Termination State Reference Extraction: Based on the multi-source state time series data, according to the unmanned equipment identifier and historical control instruction identifier, the instruction termination time corresponding to each historical control instruction is extracted. Taking the instruction termination time as the starting point for analysis, the equipment motion state, actuator state, load action state and environmental disturbance state after the instruction termination are extracted. At the same time, the stable state value of each state is extracted from the corresponding historical stable running period. A comparison relationship between the instruction termination state and the stable state is established. The instruction termination time, instruction termination state value and stable state value are associated to generate termination state comparison data.

[0015] Continuous change process determination: Based on the termination state comparison data, starting from the termination time of each historical control command, the continuous change analysis of the equipment motion state, actuator state, load action state and environmental disturbance state is performed. The residual physical action intensity of each state after the command termination is calculated. When the residual physical action intensity is greater than the residual determination threshold, it is determined that the corresponding state has not yet recovered to a stable state. The continuous time process in which the state continues to exceed the residual determination threshold from the command termination time is determined as a continuous change process. When the residual physical action intensity is continuously lower than the residual determination threshold and reaches the residual decay duration threshold, it is determined that the continuous change process has ended.

[0016] Residual Fragment Classification and Labeling: Based on the continuous change process, residual physical effects are classified according to the data source and change characteristics of the corresponding state. Among them, the continuous change process formed by the continued change of velocity, acceleration or attitude after the command is terminated is labeled as motion inertia; the continuous change process formed by the incomplete switching of braking mechanism state or the failure of equipment speed to drop to a stable state is labeled as braking delay; the continuous change process formed by the incomplete switching from load holding state to load release state or the continued change of load application position is labeled as load delayed release; the continuous change process formed by the periodic reciprocating change of attitude of actuator or equipment is labeled as mechanical vibration; the continuous change process formed by the continued propagation of airflow disturbance or wake disturbance after the command is terminated is labeled as wake disturbance; and the continuous change process formed by the unmanned equipment body, mechanical working mechanism or incompletely released load continuing to occupy the target space is labeled as space occupancy. The unmanned equipment identifier, historical control command identifier, residual physical effect type, start time, end time and residual physical effect intensity corresponding to each continuous change process are associated to generate residual physical effect fragments.

[0017] Optionally, the residual map construction includes:

[0018] Residual physical action spatial region generation: Based on the residual physical action fragment, extract the position trajectory, equipment outline boundary, actuator range, load range and environmental disturbance propagation range of the corresponding unmanned equipment during the continuous phase, and generate corresponding action regions according to the spatial influence characteristics of different types of residual physical actions. Then, spatially associate the action regions within the continuous time period to form the residual physical action spatial region.

[0019] Affected unmanned equipment identification: When the location trajectory of any unmanned equipment in the unmanned equipment cluster, other than the corresponding unmanned equipment, enters the residual physical effect space area, and the location state continues to exist in the residual physical effect space area to reach the influence duration threshold, the corresponding unmanned equipment is identified as the affected unmanned equipment;

[0020] Residual Graph Association Generation: Based on the residual physical action fragments, residual physical action spatial regions, and affected unmanned equipment, the association relationships between residual physical actions are established. Each residual physical action fragment is used as a node, and historical control command source relationships, residual physical action spatial region overlap relationships, duration phase succession relationships, and unmanned equipment influence relationships are used as connecting edges to perform graph structure association on each residual physical action fragment. Among them, the historical control command source relationship is used to represent the correspondence between residual physical action fragments and historical control commands; the residual physical action spatial region overlap relationship is used to represent the overlap relationship between different residual physical action spatial regions; the duration phase succession relationship is used to represent the temporal continuity of residual physical action fragments; and the unmanned equipment influence relationship is used to represent the interaction relationship between residual physical action fragments and affected unmanned equipment. Finally, the residual physical action type, residual physical action spatial region, duration phase, affected unmanned equipment, and association edge information corresponding to each node are uniformly stored to generate a cluster control action residual graph.

[0021] Optionally, the generation of the residual physical action space region includes:

[0022] Residual Fragment Feature Extraction: Based on the residual physical action fragments, according to the unmanned equipment identifier and residual physical action type, the spatial influence factors of the corresponding unmanned equipment during the continuous phase are extracted, including position trajectory, equipment outer boundary, actuator range, load range, and environmental disturbance propagation range. The position trajectory consists of the spatial position of the unmanned equipment at each sampling time during the continuous phase. The equipment outer boundary is used to describe the space occupied by the unmanned equipment body. The actuator range is used to describe the spatial influence range generated by mechanical execution actions. The load range is used to describe the spatial influence range corresponding to the load release or holding process. The environmental disturbance propagation range is used to describe the spatial diffusion range generated by airflow disturbance, wake disturbance, and local obstacle influence. The spatial influence factors are then associated with the residual physical action fragments to generate residual spatial influence features.

[0023] Residual space region generation: Based on the residual space influence characteristics, the corresponding residual physical action space region is determined. Specifically, for motion inertia, braking delay, and mechanical vibration, the residual physical action space region is determined based on the residual physical action generated by the change of equipment motion state and the response process of the actuator, combined with the current position of the unmanned equipment and the radius of influence range. For load delay release, the residual physical action space region is determined based on the load application position and the spatial influence range during the load release process. For wake disturbance, the residual physical action space region is determined based on the propagation range of environmental disturbance. For space occupancy, the residual physical action space region is determined based on the equipment outer boundary and the current position of the unmanned equipment.

[0024] Spatial region association and fusion: Based on the residual physical action spatial region, the spatial continuity between residual physical action spatial regions in adjacent time periods is calculated. When the spatial continuity reaches the spatial continuity determination threshold, the residual physical action spatial regions in the corresponding time period are associated and fused to obtain the residual physical action spatial region in a continuous spatiotemporal range.

[0025] Optionally, the control succession boundary construction module includes:

[0026] Task Residual Association Analysis: Based on the cluster control action residual map, obtain the task area, task time window, target unmanned equipment and task execution order corresponding to the cluster task to be executed, and perform spatiotemporal association with the residual physical action space area and duration of the residual physical action segment, identify the unpaid control action segments that still constrain the unmanned equipment due to historical control actions, and generate control action constraint association data.

[0027] Successive task segment division: Based on the control action constraint association data, according to the spatial influence range, duration and entry conditions of the residual physical action and the replacement unmanned equipment, the successive constraint analysis is performed on the cluster tasks to be executed. The task segments directly restricted by the residual physical action are divided into prohibited successive task segments, the task segments that are executed after the constraint conditions are met are divided into conditional successive task segments, and the task segments that are not affected by the residual physical action are divided into direct successive task segments.

[0028] Construction of the continuation boundary graph: Based on the prohibited continuation task segments, conditional continuation task segments, and direct continuation task segments, associate the allowed entry conditions, continuation order, and constraint relationships corresponding to each task segment. Using task segments as nodes and the continuation and constraint relationships between task segments as edges, construct the control right continuation boundary graph.

[0029] Optionally, the task residual association resolution includes:

[0030] Task information extraction: Based on the residual map of cluster control action, the task area, task time window, target unmanned equipment, and task execution order corresponding to the cluster task to be executed are obtained, and the task information is converted into task constraint data associated with residual physical action fragments. The task area is used to describe the spatial range of the task to be executed, the task time window is used to describe the start and end time of the task execution, the target unmanned equipment is used to determine the unmanned equipment identifier for executing the task, and the task execution order is used to describe the succession relationship between each task fragment, thereby generating task description data to be executed.

[0031] Residual constraint correlation analysis: Based on the description data of the task to be executed, the task region is spatially correlated with the residual physical action space region in the residual map of cluster control action, and the task time window is temporally correlated with the duration of the residual physical action segment. The spatiotemporal correlation degree between the task to be executed and the residual physical action is calculated. When the spatiotemporal correlation degree reaches the spatiotemporal correlation judgment threshold, it is determined that the corresponding task segment to be executed is constrained by the residual physical action, and the control action correlation result is generated.

[0032] Unpaid Segment Identification: Based on the control action association results, combined with the residual physical action type, residual physical action spatial region, duration phase, and target unmanned equipment entry status, the unpaid control action segments that still affect the replacement unmanned equipment after the historical control actions have ended are identified. When the spatial region corresponding to the task to be executed is within the residual physical action spatial region, and the task execution time is within the residual physical action duration phase, the corresponding task segment is marked as an unpaid control action segment, and the residual physical action fragment, the replacement unmanned equipment, and the task execution order are associated to generate control action constraint association data.

[0033] Optionally, the segmentation of the successive task includes:

[0034] Succession constraint feature extraction: Based on the control action constraint association data, extract the residual physical action spatial region, residual physical action duration stage and replacement unmanned equipment entry conditions corresponding to each task segment to be executed, and establish the constraint relationship between the task segment and the residual physical action. The replacement unmanned equipment entry conditions include spatial entry range, entry time requirements and residual action decay requirements, and generate task succession constraint features.

[0035] Task continuation type determination: Based on the aforementioned task continuation constraint characteristics, the continuation constraint degree corresponding to each task segment to be executed is calculated according to the spatial correlation between the task area and the residual physical action space area, the temporal correlation between the task time window and the duration of the residual physical action, and the degree to which the entry conditions for the replacement unmanned equipment are met. When the continuation constraint degree reaches or exceeds the prohibited continuation constraint threshold, and the replacement unmanned equipment does not meet the minimum entry conditions, the corresponding task segment is classified as a prohibited continuation task segment. When the continuation constraint degree is between the conditional continuation constraint threshold and the prohibited continuation constraint threshold, and the replacement unmanned equipment meets the minimum entry conditions... When the entry conditions are met, the corresponding task segment is divided into conditional succession task segments. When the succession constraint degree is lower than the conditional succession constraint threshold, and there is no effective association between the task area and the residual physical effect space area, and there is no effective overlap between the task time window and the residual physical effect duration stage, the corresponding task segment is divided into direct succession task segments. The minimum entry conditions include the replacement unmanned equipment space occupancy area not entering the residual physical effect space area, the replacement unmanned equipment entry time being later than the end time of the residual physical effect duration stage, and the residual physical effect intensity being lower than the residual judgment threshold and continuously reaching the residual decay duration threshold.

[0036] Successive task segment classification generation: Based on the result of the task succession type determination, each task segment in the cluster task to be executed is classified and marked. The prohibited succession task segments, conditional succession task segments, and direct succession task segments are respectively associated with the corresponding residual physical action segments, residual physical action space regions, and conditions for the replacement unmanned equipment to enter, forming a set of task succession segments.

[0037] Optionally, the phase-locked cooperative control module includes:

[0038] Phase task partitioning: Based on the control right succession boundary diagram, according to the succession relationship, constraint relationship and control right migration order between task segment nodes, the cluster task to be executed is divided into multiple continuous collaborative control phases. Each collaborative control phase corresponds to one or more task segment nodes and is associated with the corresponding task area, execution order and control right migration relationship to form a collaborative control phase set.

[0039] Phase permission configuration: Based on the set of cooperative control phases, the responsible unmanned equipment, the conditions for the release of preceding physical actions, and the allowed actions are configured for each cooperative control phase. The responsible unmanned equipment is determined according to the target unmanned equipment corresponding to the task segment. The conditions for the release of preceding physical actions are determined according to the residual physical action constraints in the control right continuation boundary diagram, including the release of the residual physical action spatial region, the end of the residual physical action duration phase, and the fulfillment of the residual action decay requirements. The allowed actions are determined according to the task type of the corresponding task segment, and phase permission configuration data is generated.

[0040] Phase lock control package generation: Based on the phase permission configuration data, the responsible unmanned equipment, the preceding physical action release conditions, the allowed execution actions and control release conditions corresponding to each cooperative control phase are associated and encapsulated to generate a phase lock control package. The responsible unmanned equipment only obtains the current phase execution permission after satisfying the preceding physical action release conditions of the corresponding cooperative control phase, and releases the execution permission after completing the task segment corresponding to the current cooperative control phase.

[0041] Optionally, the phase permission configuration includes:

[0042] Responsible device configuration: Based on the set of cooperative control phases, obtain the task segment nodes and target unmanned devices corresponding to each cooperative control phase, and determine the corresponding responsible unmanned devices according to the task segment execution object. Bind the responsible unmanned device identifier to the corresponding cooperative control phase to form responsible device configuration data. The responsible unmanned device is used to obtain the execution permission of the corresponding task segment after the pre-order physical action release condition of the current cooperative control phase is met.

[0043] Release condition configuration: Based on the residual physical action constraint relationship in the control right succession boundary diagram, extract the preceding physical action constraint information corresponding to each cooperative control phase before execution, and configure the corresponding preceding physical action release condition. The preceding physical action release condition includes the residual physical action space region release condition, the residual physical action duration phase end condition, and the residual action decay condition, which are used to limit the triggering conditions for the responsible unmanned equipment to obtain the current cooperative control phase execution authority. The residual physical action space region release condition indicates that the space occupied by the responsible unmanned equipment does not overlap with the preceding residual physical action space region. The residual physical action duration phase end condition indicates that the current time has exceeded the end time of the corresponding residual physical action duration phase. The residual action decay condition indicates that the residual physical action intensity is lower than the residual judgment threshold and continues to reach the residual decay duration threshold.

[0044] Action association: Based on the configuration data of the responsible equipment and the conditions for the release of preceding physical actions, the permitted actions are determined according to the task type of the task segment corresponding to each cooperative control phase. The responsible unmanned equipment, the conditions for the release of preceding physical actions and the permitted actions are associated and encapsulated to generate phase permission configuration data. The permitted actions include task start, task hold, task migration and task end actions.

[0045] The beneficial effects of this invention are:

[0046] This invention, by constructing a residual physical effect analysis module, identifies the persistent physical effects of unmanned equipment clusters after the completion of historical control commands, such as motion inertia, braking delay, load release delay, mechanical vibration, wake disturbance, and space occupancy. It also generates a cluster control effect residual map by combining the equipment motion state, actuator state, load state, environmental disturbance state, and relative spatial relationship between equipment. This enables continuous perception and spatial description of the implicit residual effects after the traditional control process ends, avoiding the control conflict problem caused by neglecting the residual historical control actions when task scheduling is based solely on the current equipment state in existing unmanned equipment cluster collaborative control. This improves the completeness of state cognition during cluster task execution.

[0047] This invention constructs a control succession boundary map based on the residual map of cluster control actions. It performs correlation analysis between the spatial regions and duration stages of the cluster tasks to be executed and the residual physical actions, identifies the unresolved control action segments that still constrain the replacement of unmanned equipment due to historical control actions, and divides the tasks into prohibited succession task segments, conditional succession task segments, and direct succession task segments according to the degree of influence of residual physical actions. This achieves boundary constraints on the control transfer process of unmanned equipment, and can reduce the risk of equipment collisions, task interference, and abnormal control switching caused by the lingering of residual physical actions.

[0048] This invention constructs a phase-locked collaborative control module, which converts the control right succession boundary into multiple collaborative control phases. Each collaborative control phase is configured with a responsible unmanned device, the release condition of the preceding physical action, and the permitted actions, generating a phase-locked control right package. This ensures that the responsible unmanned device only acquires execution permission after satisfying the corresponding residual physical action release condition, and releases permission according to the control right succession boundary after completing the current collaborative control phase. This achieves phased control right migration for continuous task processes of unmanned device clusters, improving the stability, security, and task continuity of collaborative execution of unmanned device clusters in complex dynamic environments. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only for this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a schematic diagram of the system functional modules according to an embodiment of the present invention;

[0051] Figure 2 This is a schematic diagram of the residual physical effect analysis module in an embodiment of the present invention. Detailed Implementation

[0052] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. Those skilled in the art may employ other alternative methods to implement some well-known technologies; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0053] like Figures 1-2 As shown, a collaborative intelligent control system for unmanned equipment clusters includes a residual physical effect analysis module, a control authority succession boundary construction module, and a phase-locked collaborative control module, wherein;

[0054] The residual physical action analysis module acquires the historical control commands, equipment motion status, actuator status, load status, environmental disturbance status, and relative spatial relationships between equipment of the unmanned equipment cluster. It identifies the residual physical actions that continue to exist after the control commands are terminated, including motion inertia, braking delay, load delay release, mechanical vibration, wake disturbance, and space occupancy. It determines the residual physical action spatial area, duration, and affected unmanned equipment for each residual physical action and generates a cluster control action residual map.

[0055] The control succession boundary construction module is based on the residual map of cluster control actions. It spatiotemporally correlates the cluster tasks to be executed with each residual physical action, identifies the unpaid control action segments that still constrain the unmanned equipment due to historical control actions, divides the cluster tasks to be executed into prohibited succession task segments, conditional succession task segments, and direct succession task segments, and constructs the control succession boundary map according to the allowed entry conditions and succession order of each task segment.

[0056] The phase-locked collaborative control module, based on the control right succession boundary diagram, divides the cluster tasks to be executed into multiple collaborative control phases. It configures the responsible unmanned equipment, the release conditions of the preceding physical action, and the allowed actions for each collaborative control phase, and generates a phase-locked control right package. After the responsible unmanned equipment meets the corresponding release conditions of the preceding physical action, it obtains the execution permission of the current collaborative control phase, and releases the execution permission after completing the current collaborative control phase, so that the next responsible unmanned equipment can obtain the execution permission of the next collaborative control phase according to the control right succession boundary diagram.

[0057] The residual physical effects analysis module includes:

[0058] Multi-source state synchronous acquisition: This involves acquiring historical control commands, equipment motion states, actuator states, load states, environmental disturbance states, and relative spatial relationships between devices within the unmanned equipment cluster. These data are then time-aligned according to the unmanned equipment identifier and a unified timestamp. The time-aligned historical control commands, equipment motion states, actuator states, load states, environmental disturbance states, and relative spatial relationships are then aggregated into multi-source state time-series data, specifically including:

[0059] (1) According to the unmanned equipment identification, the historical control commands, equipment motion status, actuator status, load action status and environmental disturbance status of each unmanned equipment in the unmanned equipment cluster are obtained from the unmanned equipment control platform, airborne controller, actuator controller, load controller and environmental perception device. Among them, the historical control commands include command type, command issuance time, command execution time and command termination time; the equipment motion status includes position, speed, acceleration and attitude; the actuator status includes the operating status of braking mechanism, steering mechanism, drive mechanism and mechanical operation mechanism; the load action status includes load holding status, load release status and load action position; the environmental disturbance status includes airflow disturbance, wake disturbance and local obstacle change. The corresponding unmanned equipment identification, data type identification, original acquisition timestamp and data source identification are added to each data to form multi-source state original data.

[0060] (2) Based on the position and attitude in the multi-source state raw data, a unified three-dimensional coordinate system is established, and the position of each unmanned device is transformed to the unified three-dimensional coordinate system. For any two unmanned devices in the unmanned device cluster, the relative distance and relative orientation at the corresponding time are calculated to form the relative spatial relationship between the devices, which is expressed as:

[0061] ;

[0062] ;

[0063] in, For a moment No. The unmanned device and the first The relative distance between the unmanned devices , , They are time points No. The coordinates of an unmanned device in a unified three-dimensional coordinate system , , They are time points No. The coordinates of an unmanned device in a unified three-dimensional coordinate system For a moment No. The unmanned equipment is relative to the first The relative position vector of each unmanned device , They are time points No. The first unmanned device, the first The spatial location of each unmanned device is determined by associating its relative distance, relative position vector, and corresponding unmanned device identifier to obtain the relative spatial relationship between the devices.

[0064] (3) Based on the original data of multiple sources and the relative spatial relationship between devices, a unified sampling period is selected and a unified timestamp sequence is constructed. Historical control commands, device motion status, actuator status, load application status, environmental disturbance status, and relative spatial relationship between devices are mapped to the unified timestamp sequence. For continuous state data whose acquisition time is inconsistent with the unified timestamp, data from adjacent acquisition times are used for time interpolation, as follows:

[0065] ;

[0066] in, For the first Each unmanned device at a unified timestamp The corresponding interpolated state value, , The first An unmanned device at adjacent raw data collection times , The corresponding state value, ;

[0067] For discrete state data of command type, load holding state, load release state and actuator operation state, the state value that is closest to the unified timestamp and is still within the effective period is used. Then, according to the unmanned equipment identifier and unified timestamp, the historical control commands, equipment motion state, actuator state, load action state, environmental disturbance state and relative spatial relationship between equipment are associated and aggregated to generate multi-source state time series data.

[0068] Residual physical effects identification: Based on multi-source state time series data, taking the termination time of each historical control command as the starting point of analysis, the continuous change analysis of the equipment motion state, actuator state, load state and environmental disturbance state after the command termination is performed. When the corresponding state has not recovered to a stable state after the command termination, the continuous change process is identified as a residual physical effect segment, and is marked as motion inertia, braking delay, load delayed release, mechanical vibration, wake disturbance or space occupation according to the source of the state.

[0069] Residual Map Construction: For each residual physical action segment, the corresponding residual physical action spatial region is determined, and it is determined whether the residual physical action spatial region overlaps with the position trajectory of each unmanned device in the unmanned device cluster except for the corresponding unmanned device. Unmanned devices whose position trajectory overlaps with the residual physical action spatial region are identified as affected unmanned devices. Using each residual physical action segment as a node, and using the historical control command source relationship, the residual physical action spatial region overlap relationship, the continuity phase inheritance relationship, and the unmanned device influence relationship as edges, each residual physical action spatial region, continuity phase, and affected unmanned device are associated to generate a cluster control action residual map.

[0070] Residual physical effects identification includes:

[0071] Termination State Baseline Extraction: Based on multi-source state time-series data, according to the unmanned equipment identifier and historical control command identifier, the command termination time corresponding to each historical control command is extracted. Using the command termination time as the starting point for analysis, the equipment motion state, actuator state, load application state, and environmental disturbance state after command termination are extracted. Simultaneously, the stable state values ​​of each state are extracted from the corresponding historical stable operating segments. A correlation relationship between the command termination state and the stable state is established. The command termination time, command termination state value, and stable state value are correlated to generate termination state comparison data, specifically including:

[0072] (1) Based on multi-source state time series data, the execution process of each historical control command is correlated and analyzed according to the unmanned equipment identifier and the historical control command identifier, and the corresponding command type, command issuance time, command execution time and command termination time are extracted; among them, the command termination time is used as the time base for subsequent state change analysis, and the command termination time corresponding to each historical control command is bound with the corresponding unmanned equipment identifier to generate a set of command termination times, represented as:

[0073] ;

[0074] in, For the set of instruction termination times, For the first The instruction termination time corresponding to each historical control instruction. This represents the number of historical control commands.

[0075] (2) Taking the instruction termination time as the starting point for parsing, extract the device motion state, actuator state, load action state, and environmental disturbance state after the instruction termination, and obtain the state value corresponding to each state at the instruction termination time to generate the state at the instruction termination time. The unmanned device corresponds to the first The state value at the time of instruction termination is represented as follows:

[0076] ;

[0077] in, For the first The unmanned device corresponds to the first The state value of the class at the time of instruction termination. For the first The unmanned device at the time of command termination Corresponding to the The actual collected values ​​of the class state, Assign a status type number, including equipment motion status, actuator status, load application status, and environmental disturbance status;

[0078] For the first The unmanned device corresponds to the first The specific state values ​​of a class at the time of instruction termination include:

[0079] 1) The status values ​​of the equipment's motion state include position status values. Speed ​​state value Acceleration state value and attitude state values ,in, , , The first The roll angle, pitch angle, and yaw angle of an unmanned device at time t;

[0080] 2) The status values ​​of the actuators include the status values ​​of the braking mechanisms. Steering mechanism status values Drive mechanism status values and mechanical operating mechanism status values , represented as:

[0081] ;

[0082] ;

[0083] ;

[0084] ;

[0085] in, This represents the current braking force. For the maximum braking amount, Indicates no braking. Indicates full braking. The current steering angle of the steering mechanism. This is the current drive output. For maximum drive output, The current workload of mechanical operating mechanisms, The maximum amount of execution required to complete the corresponding action;

[0086] 3) The state values ​​of the load application state include the load holding state values. Load release state value and load application location state value , represented as:

[0087] ;

[0088] ;

[0089] ;

[0090] in, This represents the actual holding amount of the current load holding mechanism. This represents the maximum holding amount when the load holding mechanism is fully closed. This indicates that the load is in a fully held state. This indicates that the load has been fully released, when At this time, it indicates that the load is in the process of transitioning from holding to releasing. This represents the actual amount of force released by the current load release mechanism. The maximum amount of force required for the load release mechanism to complete one full release action. This indicates that the load release is complete. This indicates that the load has not yet begun to be released. At this time, it indicates that the load is in the release process. , , These are the coordinates of the load application point in three directions within a unified three-dimensional coordinate system;

[0091] 4) The state values ​​of environmental disturbances include the state values ​​of airflow disturbance intensity. wake disturbance intensity state value and the influence of local obstacles on state values , represented as:

[0092] ;

[0093] ;

[0094] ;

[0095] ;

[0096] in, Let V be the airflow velocity vector at the location of the unmanned equipment. The amplitude of the velocity disturbance in the wake region. This represents the current impact of local obstacles. The maximum amount of obstacle impact, For the first An obstacle at any moment The corresponding space occupied area or space occupied volume For the first The unmanned device and the first The spatial distance between obstacles To avoid the default minimum positive number with a denominator of zero;

[0097] (3) Select data from the historical stable operating periods of each unmanned device from the multi-source state time series data, perform statistical analysis on the device motion state, actuator state, load action state and environmental disturbance state, extract the stable state value corresponding to each state, and establish a comparison relationship between the state at the command termination time and the stable state value. Associate the command termination time, the state value at the command termination time and the stable state value according to the unmanned device identifier, the historical control command identifier and the state type to generate termination state comparison data, which is expressed as:

[0098] ;

[0099] in, For the first The unmanned device corresponds to the first The stable state value of the class state, For the first The first unmanned device during a historically stable operating period The sampling time corresponds to the first sampling time. The state value of the class state, This is the number of samples used to calculate the steady-state values;

[0100] Historical stable operating periods are defined as continuous time intervals during which unmanned equipment remains unaffected by control command switching, task status changes, or external disturbances, and where the equipment's motion state, actuator state, load state, and environmental disturbance state remain stable within a certain time range. The stability of the operating state is determined by the rate of change of state. The unmanned device corresponds to the first Class state in time window The rate of change of state within is expressed as:

[0101] ;

[0102] in, For the first The unmanned device corresponds to the first Class state at time The rate of change of state, For the first The unmanned device corresponds to the first Class at time The state value, The state sampling time interval is defined as the condition that is met within a continuous time window. And the duration reaches the stable operation duration threshold. When, the corresponding time window is determined as the first. The unmanned device corresponds to the first Stable execution segments of class states, For the first The stable change threshold corresponding to the class state;

[0103] ;

[0104] in, , The first Mean and standard deviation of historical rate of change of class state For the first Fluctuation tolerance coefficient corresponding to the class state;

[0105] ;

[0106] in, For the first The duration of a historically stable operating segment. The standard deviation of the duration of historically stable operating segments. For duration adjustment factor, The number of historically stable running segments;

[0107] Continuous Change Process Determination: Based on the termination state comparison data, starting from the termination time of each historical control command, a continuous change analysis is performed on the equipment motion state, actuator state, load application state, and environmental disturbance state. The residual physical effect intensity of each state after command termination is calculated. When the residual physical effect intensity is greater than the residual determination threshold, the corresponding state is determined to have not yet recovered to a stable state. The continuous time process in which this state continues to exceed the residual determination threshold from the command termination time is defined as a continuous change process. When the residual physical effect intensity continuously falls below the residual determination threshold and reaches the residual decay duration threshold, the continuous change process is determined to have ended, as shown below:

[0108] ;

[0109] in, For a moment No. The unmanned device corresponds to the first The intensity of residual physical effects in the class state, For the first The unmanned device corresponds to the first The state value of the class at the time of instruction termination. For the first The unmanned device corresponds to the first The stable state value of the class state;

[0110] ;

[0111] ;

[0112] in, For the first The residual determination threshold corresponding to the class state , The first Mean and standard deviation of residual physical effects on the state. For the first The residual fluctuation adjustment coefficient corresponding to the class state, The threshold for the duration of residual decay. , These represent the mean and standard deviation of the duration of fading of historical residual physical effects. This is the decay time adjustment factor;

[0113] Residual Fragment Classification and Labeling: Based on the continuous change process, residual physical effects are classified according to the data source and change characteristics of the corresponding state. Among them, the continuous change process formed by the continued change of velocity, acceleration or attitude after the command is terminated is labeled as motion inertia; the continuous change process formed by the incomplete switching of braking mechanism state or the failure of equipment speed to drop to a stable state is labeled as braking delay; the continuous change process formed by the incomplete switching from load holding state to load release state or the continued change of load application position is labeled as load delayed release; the continuous change process formed by the periodic reciprocating change of actuator or equipment attitude is labeled as mechanical vibration; the continuous change process formed by the continued propagation of airflow disturbance or wake disturbance after the command is terminated is labeled as wake disturbance; and the continuous change process formed by the unmanned equipment body, mechanical operating mechanism or incompletely released load continuing to occupy the target space is labeled as space occupancy. The unmanned equipment identifier, historical control command identifier, residual physical effect type, start time, end time and residual physical effect intensity corresponding to each continuous change process are associated to generate residual physical effect fragments.

[0114] Residual map construction includes:

[0115] Residual physical action spatial region generation: Based on residual physical action fragments, extract the position trajectory, equipment outline boundary, actuator range, load range and environmental disturbance propagation range of the corresponding unmanned equipment during the continuous phase, and generate corresponding action regions according to the spatial impact characteristics of different types of residual physical actions. Then, spatially correlate the action regions within a continuous time period to form the residual physical action spatial region.

[0116] Affected Unmanned Equipment Identification: When the location trajectory of any unmanned equipment in the unmanned equipment cluster, excluding the corresponding unmanned equipment, enters the residual physical effect space area, and this location state persists within the residual physical effect space area for a duration threshold of influence, the corresponding unmanned equipment is identified as an affected unmanned equipment. The determination condition is as follows: and ,in, For the first An unmanned device at all times Location, For a moment No. The unmanned device corresponds to the first The spatial region of residual physical effects, For the first The moment when an unmanned device leaves the area of ​​residual physical influence. For the first The moment when an unmanned device first enters the area of ​​residual physical effects. To affect the duration threshold;

[0117] ;

[0118] in, , These represent the mean and standard deviation of the duration of the influence of historical residual physical effects. To influence the duration adjustment factor;

[0119] Residual Graph Association Generation: Based on residual physical action fragments, residual physical action spatial regions, and affected unmanned equipment, the association relationships between residual physical actions are established. Each residual physical action fragment is used as a node, and historical control command source relationships, residual physical action spatial region overlap relationships, duration phase succession relationships, and unmanned equipment influence relationships are used as connecting edges to perform graph structure association on each residual physical action fragment. Specifically, historical control command source relationships represent the correspondence between residual physical action fragments and historical control commands; residual physical action spatial region overlap relationships represent the overlap between different residual physical action spatial regions; duration phase succession relationships represent the temporal continuity of residual physical action fragments; and unmanned equipment influence relationships represent the interaction relationships between residual physical action fragments and affected unmanned equipment. Finally, the residual physical action type, residual physical action spatial region, duration phase, affected unmanned equipment, and association edge information corresponding to each node are uniformly stored to generate a cluster control action residual graph, represented as:

[0120] ;

[0121] in, For the residual map of cluster control, It is a set of residual physical action fragment nodes. It is the set of associated edges between residual physical action fragments.

[0122] The generation of residual physical effects in the spatial region includes:

[0123] Residual Fragment Feature Extraction: Based on residual physical action fragments, according to the unmanned equipment identifier and residual physical action type, the spatial influencing factors of the corresponding unmanned equipment during the continuous phase are extracted, including position trajectory, equipment outer boundary, actuator range, load range, and environmental disturbance propagation range. Among them, the position trajectory consists of the spatial position of the unmanned equipment at each sampling time during the continuous phase, the equipment outer boundary is used to describe the space occupied by the unmanned equipment body, the actuator range is used to describe the spatial influence range generated by mechanical execution actions, the load range is used to describe the spatial influence range corresponding to the load release or holding process, and the environmental disturbance propagation range is used to describe the spatial diffusion range generated by airflow disturbance, wake disturbance, and local obstacle influence. The spatial influencing factors are then associated with residual physical action fragments to generate residual spatial influence features.

[0124] Residual space region generation: Based on the residual space influence characteristics, the corresponding residual physical action space region is determined. Specifically, for motion inertia, braking delay, and mechanical vibration, the residual physical action space region is determined based on the residual physical action generated by the changes in the equipment's motion state and the response process of the actuator, combined with the current position of the unmanned equipment and the radius of influence range. For load delay release, the residual physical action space region is determined based on the load's position and the spatial influence range during the load release process. For wake disturbance, the residual physical action space region is determined based on the propagation range of environmental disturbance. For space occupancy, the residual physical action space region is determined based on the equipment's outer boundary and the current position of the unmanned equipment.

[0125] Spatial Region Association and Fusion: Based on the spatial regions of residual physical effects, the spatial continuity between the spatial regions of residual physical effects in adjacent time periods is calculated. When the spatial continuity reaches the spatial continuity determination threshold, the spatial regions of residual physical effects in the corresponding time periods are associated and fused to obtain the spatial regions of residual physical effects within a continuous spatiotemporal range, represented as:

[0126] ;

[0127] in, For the first The unmanned device corresponds to the first Spatial continuity of residual physical effects at adjacent time points. This represents the residual physical interaction space region at the next sampling time. The sampling time interval;

[0128] ;

[0129] in, The threshold for determining spatial continuity. , These represent the average value and the degree of fluctuation of the spatial continuity of the spatial region due to historical residual physical effects, respectively. This is the spatially continuous adjustment coefficient.

[0130] The control succession boundary construction module includes:

[0131] Task Residual Correlation Analysis: Based on the residual map of cluster control actions, the task region, task time window, target unmanned equipment and task execution order corresponding to the cluster task to be executed are obtained, and spatiotemporally correlated with the residual physical action space region and duration of the residual physical action fragments. The unpaid control action segments that still constrain the unmanned equipment due to historical control actions are identified, and control action constraint correlation data is generated.

[0132] Successive task segment division: Based on the control action constraint association data, according to the spatial influence range, duration and entry conditions of the residual physical action and the replacement unmanned equipment, the successive constraint analysis is performed on the cluster tasks to be executed. The task segments directly restricted by the residual physical action are divided into prohibited successive task segments, the task segments that are executed after the constraint conditions are met are divided into conditional successive task segments, and the task segments that are not affected by the residual physical action are divided into direct successive task segments.

[0133] Succession Boundary Graph Construction: Based on prohibited succession task segments, conditional succession task segments, and direct succession task segments, the allowed entry conditions, succession order, and constraints corresponding to each task segment are associated. Using task segments as nodes and the succession and constraint relationships between task segments as edges, a control right succession boundary graph is constructed to describe the control right transfer boundary during the task succession process of an unmanned equipment cluster. Specifically, this includes:

[0134] (1) Based on prohibited continuation task segments, conditional continuation task segments, and direct continuation task segments, extract the task area, task time window, target unmanned equipment, allowed entry conditions, and task segment type corresponding to each task segment, and set each task segment as a task segment node. Among them, the prohibited continuation task segment node is configured with a prohibition to obtain control rights flag, the conditional continuation task segment node is configured with a control rights flag after meeting the allowed entry conditions, and the direct continuation task segment node is configured with a direct control rights flag, forming a task segment node set. ;

[0135] (2) Based on the task execution order of the cluster tasks to be executed, determine whether there is a successor relationship between any two task segments. When the subsequent task segment needs to be executed after the previous task segment is completed or releases control, establish a successor relationship edge between the corresponding task segment nodes. When the subsequent task segment also needs to meet the allowed entry conditions, establish a constraint relationship edge between the corresponding task segment nodes. The spatial entry condition, entry time condition, and residual effect decay condition are used as edge attributes to form an edge set. The connection relationships between task fragment nodes are represented as follows:

[0136] ;

[0137] in, For the first The task fragment node and the first The connection status between task segment nodes;

[0138] Constraints represent relationships whereby a subsequent task segment, upon gaining control or commencing execution, is subject to the residual physical effects, sequential order, and permitted entry conditions of the preceding task segment. These constraints include:

[0139] 1) Spatial constraint relationship: that is, before the unmanned equipment corresponding to the next task segment enters the task area, its spatial occupation area shall not overlap with the residual physical action space area corresponding to the previous task segment.

[0140] 2) Time constraint relationship: that is, the next task segment can only release control after the previous task segment, and gain control after the corresponding residual physical effect lasts for a period of time.

[0141] 3) Residual effect constraint relationship: that is, before the next task segment starts to be executed, the residual physical effect intensity corresponding to the previous task segment needs to be lower than the residual judgment threshold and continue to reach the residual decay duration threshold.

[0142] (3) Based on the set of task fragment nodes With edge set Connect the task fragment nodes according to the task execution order, and associate the task fragment type, allowed entry conditions, and control release requirements with the corresponding nodes or connecting edges to construct a control succession boundary graph. .

[0143] Task residual association analysis includes:

[0144] Task Information Extraction: Based on the residual map of cluster control actions, the task region, task time window, target unmanned equipment, and task execution order corresponding to the cluster task to be executed are obtained. The task information is then converted into task constraint data associated with residual physical action fragments. The task region is used to describe the spatial range of the task to be executed, the task time window is used to describe the start and end time of the task execution, the target unmanned equipment is used to determine the unmanned equipment identifier for executing the task, and the task execution order is used to describe the succession relationship between each task fragment, thus generating description data of the task to be executed.

[0145] Residual constraint correlation analysis: Based on the description data of the task to be executed, the task region is spatially correlated with the spatial region of residual physical actions in the residual map of cluster control actions, and the task time window is temporally correlated with the duration of the residual physical action segment. The spatiotemporal correlation degree between the task to be executed and the residual physical action is calculated. When the spatiotemporal correlation degree reaches the spatiotemporal correlation judgment threshold, it is determined that the corresponding task segment to be executed is constrained by the residual physical action, and the control action correlation result is generated, represented as:

[0146] ;

[0147] ;

[0148] ;

[0149] ;

[0150] in, For the first The degree of spatiotemporal correlation between each task fragment to be executed and the residual physical effects. The degree of spatial correlation between the mission area and the area of ​​residual physical effects. The degree of temporal correlation between the mission time window and the duration of residual physical effects. , These are spatial correlation weights and temporal correlation weights, respectively. For the first The task area corresponding to each task segment to be executed. For the first The task time window corresponding to each pending task segment. For the first The unmanned device corresponds to the first The sustained phase of residual physical action, The threshold for determining spatiotemporal correlation. , These represent the average and standard deviation of historical spatiotemporal correlation, respectively. This is the spatiotemporal correlation adjustment coefficient;

[0151] Unpaid Segment Identification: Based on the control action association results, combined with the residual physical action type, residual physical action spatial region, duration phase, and target unmanned equipment entry status, the unpaid control action segments that still affect the replacement unmanned equipment after the historical control actions have ended are identified. When the spatial region corresponding to the task to be executed is within the residual physical action spatial region, and the task execution time is within the residual physical action duration phase, the corresponding task segment is marked as an unpaid control action segment, and the residual physical action fragment, the replacement unmanned equipment, and the task execution order are associated to generate control action constraint association data.

[0152] The subsequent task segments are divided into:

[0153] Succession constraint feature extraction: Based on the control action constraint association data, extract the residual physical action spatial region, residual physical action duration stage and replacement unmanned equipment entry conditions corresponding to each task segment to be executed, and establish the constraint relationship between task segments and residual physical actions. Among them, the replacement unmanned equipment entry conditions include spatial entry range, entry time requirements and residual action decay requirements, generating task succession constraint features.

[0154] The spatial access range is determined based on the task area to be executed, the residual physical effect space area, and the dimensions of the unmanned equipment itself. It describes the effective spatial range within the task area that the replacement unmanned equipment can enter. For example, if after a certain unmanned equipment completes its load release, there is still a load movement influence range around its release area, then the replacement unmanned equipment must avoid this residual physical effect space area before entering, as shown below:

[0155] ;

[0156] in, For the first A replacement for unmanned equipment is in time. The space occupied;

[0157] The entry time requirement indicates that the time for the replacement unmanned equipment to enter the task area must avoid the period of residual physical effects. In other words, the replacement unmanned equipment cannot enter the corresponding area immediately before the residual effects of the previous control actions have ended. This is expressed as:

[0158] ;

[0159] in, For the first The moment when unmanned equipment takes over and enters the mission area. For the first The unmanned device corresponds to the first The moment when the residual physical action ends;

[0160] The residual effect decay requirement indicates that the intensity of residual physical effects has been reduced to a level that will not affect the continuation of control, meaning that the motion inertia, braking delay, load release delay, mechanical vibration, wake disturbance, or space occupation caused by historical control actions have completely decayed, and is expressed as:

[0161] and ;

[0162] Task succession type determination: Based on task succession constraint characteristics, the succession constraint degree corresponding to each task segment to be executed is calculated according to the spatial correlation between the task area and the residual physical action space area, the temporal correlation between the task time window and the duration of the residual physical action, and the degree to which the entry conditions of the replacement unmanned equipment are met. The succession constraint degree is used to characterize the restriction strength of the historical control action residue on the task succession process. The higher the spatial correlation degree, the higher the temporal correlation degree, and the lower the degree to which the entry conditions of the replacement unmanned equipment are met, the higher the corresponding succession constraint degree. When the succession constraint degree reaches or exceeds the prohibited succession constraint threshold, and the replacement unmanned equipment does not meet the minimum entry conditions, the corresponding task segment is classified as a prohibited succession task segment. When the constraint level is between the conditional continuation constraint threshold and the prohibition continuation constraint threshold, and the replacement unmanned equipment meets the minimum entry conditions, the corresponding task segment is classified as a conditional continuation task segment. When the continuation constraint level is lower than the conditional continuation constraint threshold, and there is no effective connection between the task area and the residual physical effect space area, and no effective overlap between the task time window and the residual physical effect duration phase, the corresponding task segment is classified as a direct continuation task segment. The minimum entry conditions include that the space occupied by the replacement unmanned equipment has not entered the residual physical effect space area, the entry time of the replacement unmanned equipment is later than the end time of the residual physical effect duration phase, and the residual physical effect intensity is lower than the residual judgment threshold and continues to reach the residual decay duration threshold, expressed as:

[0163] ;

[0164] ;

[0165] ;

[0166] ;

[0167] in, For the first The degree of continuity constraint corresponding to each task segment to be executed For the first The minimum entry condition satisfaction level corresponding to each pending task segment. , , These are spatial association weight, temporal association weight, and entry condition weight, respectively. For the first The minimum number of unmanned devices that have met the entry conditions for each pending task segment. For the first The minimum total number of entry conditions corresponding to each pending task segment. To prohibit the continuation of constraint thresholds, , These represent the mean and standard deviation of the degree of constraint on the continuation of historical prohibited task segments. To prohibit the continuation of the threshold adjustment coefficient, The threshold for conditional continuation constraints. , These represent the mean and standard deviation of the constraint degree for the continuation of historical condition-based task segments. This is the conditional continuation threshold adjustment coefficient;

[0168] Successive task segment classification and generation: Based on the result of the task succession type determination, each task segment in the cluster task to be executed is classified and marked. The prohibited succession task segments, conditional succession task segments, and direct succession task segments are respectively associated with the corresponding residual physical action segments, residual physical action space regions, and conditions for the entry of the replacement unmanned equipment, forming a set of task succession segments.

[0169] The phase-locked cooperative control module includes:

[0170] Phase Task Partitioning: Based on the control right succession boundary graph, and according to the succession relationships, constraints, and control right transition order among task fragment nodes, the cluster tasks to be executed are divided into multiple continuous collaborative control phases. Each collaborative control phase corresponds to one or more task fragment nodes and is associated with corresponding task regions, execution order, and control right transition relationships, forming a set of collaborative control phases, specifically including:

[0171] (1) Based on the control right succession boundary diagram, extract the succession relationship, constraint relationship and control right migration order between each task segment node, and aggregate the task segment nodes according to the continuous execution relationship. The task segment nodes with continuous control right migration relationship and satisfy the same execution constraint are divided into the same collaborative control phase to form an initial collaborative control phase set. Each collaborative control phase corresponds to one or more task segment nodes and is associated with the corresponding task area and target unmanned equipment. The control right migration order is the order in which the execution rights of different task segments are transferred from one responsible unmanned equipment to another responsible unmanned equipment during the execution of continuous tasks by the unmanned equipment cluster.

[0172] (2) Based on the initial set of collaborative control phases, the switching boundary between adjacent collaborative control phases is determined according to the control transfer relationship and constraint relationship between task segment nodes. When the task segment node corresponding to the next collaborative control phase needs to wait for the previous collaborative control phase to complete the release of control rights and meets the corresponding residual physical action release condition, a phase transfer relationship is established, and the end node of the previous collaborative control phase and the start node of the next collaborative control phase are associated to generate a phase boundary relationship.

[0173] (3) Based on the phase boundary relationship, the task segment nodes, task areas, execution order and control transfer relationship corresponding to each cooperative control phase are associated and encapsulated to generate a cooperative control phase set. The cooperative control phase set includes the execution order of each cooperative control phase, the association relationship of the responsible unmanned equipment and the control transfer boundary between phases, which is used to generate the phase locking control package in the future.

[0174] Phase permission configuration: Based on the set of collaborative control phases, the responsible unmanned equipment, the conditions for the release of preceding physical actions, and the allowed actions are configured for each collaborative control phase. The responsible unmanned equipment is determined according to the target unmanned equipment corresponding to the task segment. The conditions for the release of preceding physical actions are determined according to the constraints of residual physical actions in the control right continuation boundary diagram, including the release of residual physical action spatial area, the end of the residual physical action duration phase, and the fulfillment of residual action fading requirements. The allowed actions are determined according to the task type of the corresponding task segment, and phase permission configuration data is generated.

[0175] Phase-locked control package generation: Based on phase permission configuration data, the responsible unmanned equipment, the conditions for releasing the preceding physical action, the allowed execution action and the conditions for releasing control rights corresponding to each cooperative control phase are associated and encapsulated to generate a phase-locked control package. The responsible unmanned equipment only obtains the execution rights of the current phase after the conditions for releasing the preceding physical action of the corresponding cooperative control phase are met, and releases the execution rights after completing the task segment corresponding to the current cooperative control phase.

[0176] Phase permission configuration includes:

[0177] Responsible equipment configuration: Based on the set of cooperative control phases, the task segment nodes and target unmanned equipment corresponding to each cooperative control phase are obtained, and the corresponding responsible unmanned equipment is determined according to the execution object of the task segment. The responsible unmanned equipment identifier is bound to the corresponding cooperative control phase to form responsible equipment configuration data. The responsible unmanned equipment is used to obtain the execution permission of the corresponding task segment after the pre-sequence physical action release condition of the current cooperative control phase is met.

[0178] Release condition configuration: Based on the residual physical action constraint relationship in the control right succession boundary diagram, extract the preceding physical action constraint information corresponding to each cooperative control phase before execution, and configure the corresponding preceding physical action release conditions. Among them, the preceding physical action release conditions include the residual physical action space region release condition, the residual physical action duration phase end condition, and the residual action decay condition, which are used to limit the triggering conditions for the responsible unmanned equipment to obtain the current cooperative control phase execution authority. The residual physical action space region release condition indicates that the space occupied by the responsible unmanned equipment does not overlap with the preceding residual physical action space region. The residual physical action duration phase end condition indicates that the current time has exceeded the end time of the corresponding residual physical action duration phase. The residual action decay condition indicates that the residual physical action intensity is lower than the residual judgment threshold and continues to reach the residual decay duration threshold.

[0179] Action association: Based on the configuration data of the responsible equipment and the conditions for the release of preceding physical actions, the permissible actions are determined according to the task type of the task segment corresponding to each cooperative control phase. The responsible unmanned equipment, the conditions for the release of preceding physical actions and the permissible actions are associated and encapsulated to generate phase permission configuration data. The permissible actions include task start, task hold, task migration and task end actions, which are used to limit the control behaviors that the responsible unmanned equipment can perform within the current cooperative control phase.

[0180] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0181] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A collaborative intelligent control system for unmanned equipment clusters, characterized in that, It includes a residual physical effect analysis module, a control right succession boundary construction module, and a phase-locked collaborative control module, among which; The residual physical action analysis module acquires the historical control commands, equipment motion status, actuator status, load status, environmental disturbance status, and relative spatial relationships between equipment of the unmanned equipment cluster. It identifies the residual physical actions that continue to exist after the control commands are terminated, including motion inertia, braking delay, load delay release, mechanical vibration, wake disturbance, and space occupancy. It determines the residual physical action spatial region, duration, and affected unmanned equipment for each residual physical action and generates a cluster control action residual map. The control right succession boundary construction module, based on the cluster control action residual map, spatiotemporally correlates the cluster tasks to be executed with each residual physical action, identifies unpaid control action segments that still constrain unmanned equipment due to historical control actions, divides the cluster tasks to be executed into prohibited succession task segments, conditional succession task segments, and direct succession task segments, and constructs a control right succession boundary map based on the allowed entry conditions and succession order of each task segment. The phase-locked collaborative control module divides the cluster task to be executed into multiple collaborative control phases based on the control right succession boundary diagram. It configures the responsible unmanned equipment, the release condition of the preceding physical action, and the allowed execution action for each collaborative control phase, and generates a phase-locked control right package. The responsible unmanned equipment obtains the execution permission of the current collaborative control phase after satisfying the corresponding release condition of the preceding physical action, and releases the execution permission after completing the current collaborative control phase.

2. The unmanned equipment cluster collaborative intelligent control system according to claim 1, characterized in that, The residual physical effect analysis module includes: Multi-source state synchronous acquisition: Acquire historical control commands, equipment motion status, actuator status, load application status, environmental disturbance status, and relative spatial relationships between equipment for each unmanned equipment in the unmanned equipment cluster. Then, perform time-series alignment according to the unmanned equipment identifier and unified timestamp. The historical control commands, equipment motion status, actuator status, load application status, environmental disturbance status, and relative spatial relationships between equipment that have completed time-series alignment are collected into multi-source state time-series data. Residual physical effects identification: Based on the multi-source state time series data, taking the termination time of each historical control command as the starting point of analysis, the continuous change analysis of the equipment motion state, actuator state, load action state and environmental disturbance state after the command termination is performed. When the corresponding state has not recovered to a stable state after the command termination, the continuous change process is identified as a residual physical effect segment, and is marked as motion inertia, braking delay, load delayed release, mechanical vibration, wake disturbance or space occupation according to the source of the state. Residual Map Construction: For each residual physical action segment, the corresponding residual physical action spatial region is determined, and it is determined whether the residual physical action spatial region overlaps with the position trajectory of each unmanned device in the unmanned device cluster except for the corresponding unmanned device. Unmanned devices whose position trajectory overlaps with the residual physical action spatial region are identified as affected unmanned devices. Using each residual physical action segment as a node, and using the historical control command source relationship, the residual physical action spatial region overlap relationship, the continuous stage inheritance relationship, and the unmanned device influence relationship as edges, each residual physical action spatial region, continuous stage, and affected unmanned device are associated to generate a cluster control action residual map.

3. The unmanned equipment cluster collaborative intelligent control system according to claim 2, characterized in that, The identification of residual physical effects includes: Termination State Reference Extraction: Based on the multi-source state time series data, according to the unmanned equipment identifier and historical control instruction identifier, the instruction termination time corresponding to each historical control instruction is extracted. Taking the instruction termination time as the starting point for analysis, the equipment motion state, actuator state, load action state and environmental disturbance state after the instruction termination are extracted. At the same time, the stable state value of each state is extracted from the corresponding historical stable running period. A comparison relationship between the instruction termination state and the stable state is established. The instruction termination time, instruction termination state value and stable state value are associated to generate termination state comparison data. Continuous change process determination: Based on the termination state comparison data, starting from the termination time of each historical control command, the continuous change analysis of the equipment motion state, actuator state, load action state and environmental disturbance state is performed. The residual physical action intensity of each state after the command termination is calculated. When the residual physical action intensity is greater than the residual determination threshold, it is determined that the corresponding state has not yet recovered to a stable state. The continuous time process in which the state continues to exceed the residual determination threshold from the command termination time is determined as a continuous change process. When the residual physical action intensity is continuously lower than the residual determination threshold and reaches the residual decay duration threshold, it is determined that the continuous change process has ended. Residual Fragment Classification and Labeling: Based on the continuous change process, residual physical effects are classified according to the data source and change characteristics of the corresponding state. Among them, the continuous change process formed by the continued change of velocity, acceleration or attitude after the command is terminated is labeled as motion inertia; the continuous change process formed by the incomplete switching of braking mechanism state or the failure of equipment speed to drop to a stable state is labeled as braking delay; the continuous change process formed by the incomplete switching from load holding state to load release state or the continued change of load application position is labeled as load delayed release; the continuous change process formed by the periodic reciprocating change of attitude of actuator or equipment is labeled as mechanical vibration; the continuous change process formed by the continued propagation of airflow disturbance or wake disturbance after the command is terminated is labeled as wake disturbance; and the continuous change process formed by the unmanned equipment body, mechanical working mechanism or incompletely released load continuing to occupy the target space is labeled as space occupancy. The unmanned equipment identifier, historical control command identifier, residual physical effect type, start time, end time and residual physical effect intensity corresponding to each continuous change process are associated to generate residual physical effect fragments.

4. The unmanned equipment cluster collaborative intelligent control system according to claim 3, characterized in that, The construction of the residual map includes: Residual physical action spatial region generation: Based on the residual physical action fragment, extract the position trajectory, equipment outline boundary, actuator range, load range and environmental disturbance propagation range of the corresponding unmanned equipment during the continuous phase, and generate corresponding action regions according to the spatial influence characteristics of different types of residual physical actions. Then, spatially associate the action regions within the continuous time period to form the residual physical action spatial region. Affected unmanned equipment identification: When the location trajectory of any unmanned equipment in the unmanned equipment cluster, other than the corresponding unmanned equipment, enters the residual physical effect space area, and the location state continues to exist in the residual physical effect space area to reach the influence duration threshold, the corresponding unmanned equipment is identified as the affected unmanned equipment; Residual Graph Association Generation: Based on the residual physical action fragments, residual physical action spatial regions, and affected unmanned equipment, the association relationships between residual physical actions are established. Each residual physical action fragment is used as a node, and historical control command source relationships, residual physical action spatial region overlap relationships, duration phase succession relationships, and unmanned equipment influence relationships are used as connecting edges to perform graph structure association on each residual physical action fragment. Among them, the historical control command source relationship is used to represent the correspondence between residual physical action fragments and historical control commands; the residual physical action spatial region overlap relationship is used to represent the overlap relationship between different residual physical action spatial regions; the duration phase succession relationship is used to represent the temporal continuity of residual physical action fragments; and the unmanned equipment influence relationship is used to represent the interaction relationship between residual physical action fragments and affected unmanned equipment. Finally, the residual physical action type, residual physical action spatial region, duration phase, affected unmanned equipment, and association edge information corresponding to each node are uniformly stored to generate a cluster control action residual graph.

5. The unmanned equipment cluster collaborative intelligent control system according to claim 4, characterized in that, The generation of the residual physical action space region includes: Residual Fragment Feature Extraction: Based on the residual physical action fragments, according to the unmanned equipment identifier and residual physical action type, the spatial influence factors of the corresponding unmanned equipment during the continuous phase are extracted, including position trajectory, equipment outer boundary, actuator range, load range, and environmental disturbance propagation range. The position trajectory consists of the spatial position of the unmanned equipment at each sampling time during the continuous phase. The equipment outer boundary is used to describe the space occupied by the unmanned equipment body. The actuator range is used to describe the spatial influence range generated by mechanical execution actions. The load range is used to describe the spatial influence range corresponding to the load release or holding process. The environmental disturbance propagation range is used to describe the spatial diffusion range generated by airflow disturbance, wake disturbance, and local obstacle influence. The spatial influence factors are then associated with the residual physical action fragments to generate residual spatial influence features. Residual space region generation: Based on the residual space influence characteristics, the corresponding residual physical action space region is determined. Specifically, for motion inertia, braking delay, and mechanical vibration, the residual physical action space region is determined based on the residual physical action generated by the change of equipment motion state and the response process of the actuator, combined with the current position of the unmanned equipment and the radius of influence range. For load delay release, the residual physical action space region is determined based on the load application position and the spatial influence range during the load release process. For wake disturbance, the residual physical action space region is determined based on the propagation range of environmental disturbance. For space occupancy, the residual physical action space region is determined based on the equipment outer boundary and the current position of the unmanned equipment. Spatial region association and fusion: Based on the residual physical action spatial region, the spatial continuity between residual physical action spatial regions in adjacent time periods is calculated. When the spatial continuity reaches the spatial continuity determination threshold, the residual physical action spatial regions in the corresponding time period are associated and fused to obtain the residual physical action spatial region in a continuous spatiotemporal range.

6. The unmanned equipment cluster collaborative intelligent control system according to claim 1, characterized in that, The control succession boundary construction module includes: Task Residual Association Analysis: Based on the cluster control action residual map, obtain the task area, task time window, target unmanned equipment and task execution order corresponding to the cluster task to be executed, and perform spatiotemporal association with the residual physical action space area and duration of the residual physical action segment, identify the unpaid control action segments that still constrain the unmanned equipment due to historical control actions, and generate control action constraint association data. Successive task segment division: Based on the control action constraint association data, according to the spatial influence range, duration and entry conditions of the residual physical action and the replacement unmanned equipment, the successive constraint analysis is performed on the cluster tasks to be executed. The task segments directly restricted by the residual physical action are divided into prohibited successive task segments, the task segments that are executed after the constraint conditions are met are divided into conditional successive task segments, and the task segments that are not affected by the residual physical action are divided into direct successive task segments. Construction of the continuation boundary graph: Based on the prohibited continuation task segments, conditional continuation task segments, and direct continuation task segments, associate the allowed entry conditions, continuation order, and constraint relationships corresponding to each task segment. Using task segments as nodes and the continuation and constraint relationships between task segments as edges, construct the control right continuation boundary graph.

7. The unmanned equipment cluster collaborative intelligent control system according to claim 6, characterized in that, The task residual association resolution includes: Task information extraction: Based on the residual map of cluster control action, the task area, task time window, target unmanned equipment, and task execution order corresponding to the cluster task to be executed are obtained, and the task information is converted into task constraint data associated with residual physical action fragments. The task area is used to describe the spatial range of the task to be executed, the task time window is used to describe the start and end time of the task execution, the target unmanned equipment is used to determine the unmanned equipment identifier for executing the task, and the task execution order is used to describe the succession relationship between each task fragment, thereby generating task description data to be executed. Residual constraint correlation analysis: Based on the description data of the task to be executed, the task region is spatially correlated with the residual physical action space region in the residual map of cluster control action, and the task time window is temporally correlated with the duration of the residual physical action segment. The spatiotemporal correlation degree between the task to be executed and the residual physical action is calculated. When the spatiotemporal correlation degree reaches the spatiotemporal correlation judgment threshold, it is determined that the corresponding task segment to be executed is constrained by the residual physical action, and the control action correlation result is generated. Unpaid Segment Identification: Based on the control action association results, combined with the residual physical action type, residual physical action spatial region, duration phase, and target unmanned equipment entry status, the unpaid control action segments that still affect the replacement unmanned equipment after the historical control actions have ended are identified. When the spatial region corresponding to the task to be executed is within the residual physical action spatial region, and the task execution time is within the residual physical action duration phase, the corresponding task segment is marked as an unpaid control action segment, and the residual physical action fragment, the replacement unmanned equipment, and the task execution order are associated to generate control action constraint association data.

8. The unmanned equipment cluster collaborative intelligent control system according to claim 7, characterized in that, The segmentation of the successive task includes: Succession constraint feature extraction: Based on the control action constraint association data, extract the residual physical action spatial region, residual physical action duration stage and replacement unmanned equipment entry conditions corresponding to each task segment to be executed, and establish the constraint relationship between the task segment and the residual physical action. The replacement unmanned equipment entry conditions include spatial entry range, entry time requirements and residual action decay requirements, and generate task succession constraint features. Task continuation type determination: Based on the aforementioned task continuation constraint characteristics, the continuation constraint degree corresponding to each task segment to be executed is calculated according to the spatial correlation between the task area and the residual physical action space area, the temporal correlation between the task time window and the duration of the residual physical action, and the degree to which the entry conditions for the replacement unmanned equipment are met. When the continuation constraint degree reaches or exceeds the prohibited continuation constraint threshold, and the replacement unmanned equipment does not meet the minimum entry conditions, the corresponding task segment is classified as a prohibited continuation task segment. When the continuation constraint degree is between the conditional continuation constraint threshold and the prohibited continuation constraint threshold, and the replacement unmanned equipment meets the minimum entry conditions... When the entry conditions are met, the corresponding task segment is divided into conditional succession task segments. When the succession constraint degree is lower than the conditional succession constraint threshold, and there is no effective association between the task area and the residual physical effect space area, and there is no effective overlap between the task time window and the residual physical effect duration stage, the corresponding task segment is divided into direct succession task segments. The minimum entry conditions include the replacement unmanned equipment space occupancy area not entering the residual physical effect space area, the replacement unmanned equipment entry time being later than the end time of the residual physical effect duration stage, and the residual physical effect intensity being lower than the residual judgment threshold and continuously reaching the residual decay duration threshold. Successive task segment classification generation: Based on the result of the task succession type determination, each task segment in the cluster task to be executed is classified and marked. The prohibited succession task segments, conditional succession task segments, and direct succession task segments are respectively associated with the corresponding residual physical action segments, residual physical action space regions, and conditions for the replacement unmanned equipment to enter, forming a set of task succession segments.

9. The unmanned equipment cluster collaborative intelligent control system according to claim 8, characterized in that, The phase-locked collaborative control module includes: Phase task partitioning: Based on the control right succession boundary diagram, according to the succession relationship, constraint relationship and control right migration order between task segment nodes, the cluster task to be executed is divided into multiple continuous collaborative control phases. Each collaborative control phase corresponds to one or more task segment nodes and is associated with the corresponding task area, execution order and control right migration relationship to form a collaborative control phase set. Phase permission configuration: Based on the set of cooperative control phases, the responsible unmanned equipment, the conditions for the release of preceding physical actions, and the allowed actions are configured for each cooperative control phase. The responsible unmanned equipment is determined according to the target unmanned equipment corresponding to the task segment. The conditions for the release of preceding physical actions are determined according to the residual physical action constraints in the control right continuation boundary diagram, including the release of the residual physical action spatial region, the end of the residual physical action duration phase, and the fulfillment of the residual action decay requirements. The allowed actions are determined according to the task type of the corresponding task segment, and phase permission configuration data is generated. Phase lock control package generation: Based on the phase permission configuration data, the responsible unmanned equipment, the preceding physical action release conditions, the allowed execution actions and control release conditions corresponding to each cooperative control phase are associated and encapsulated to generate a phase lock control package. The responsible unmanned equipment only obtains the current phase execution permission after satisfying the preceding physical action release conditions of the corresponding cooperative control phase, and releases the execution permission after completing the task segment corresponding to the current cooperative control phase.

10. The unmanned equipment cluster collaborative intelligent control system according to claim 9, characterized in that, The phase permission configuration includes: Responsible device configuration: Based on the set of cooperative control phases, the task segment nodes and target unmanned devices corresponding to each cooperative control phase are obtained, and the corresponding responsible unmanned devices are determined according to the task segment execution object. The responsible unmanned device identifier is bound to the corresponding cooperative control phase to form responsible device configuration data. The responsible unmanned device is used to obtain the execution permission of the corresponding task segment after the pre-order physical action release condition of the current cooperative control phase is met. Release condition configuration: Based on the residual physical action constraint relationship in the control right succession boundary diagram, extract the preceding physical action constraint information corresponding to each cooperative control phase before execution, and configure the corresponding preceding physical action release condition. The preceding physical action release condition includes the residual physical action space region release condition, the residual physical action duration phase end condition, and the residual action decay condition, which are used to limit the triggering conditions for the responsible unmanned equipment to obtain the current cooperative control phase execution authority. The residual physical action space region release condition indicates that the space occupied by the responsible unmanned equipment does not overlap with the preceding residual physical action space region. The residual physical action duration phase end condition indicates that the current time has exceeded the end time of the corresponding residual physical action duration phase. The residual action decay condition indicates that the residual physical action intensity is lower than the residual judgment threshold and continues to reach the residual decay duration threshold. Action association: Based on the configuration data of the responsible equipment and the conditions for the release of preceding physical actions, the permitted actions are determined according to the task type of the task segment corresponding to each cooperative control phase. The responsible unmanned equipment, the conditions for the release of preceding physical actions and the permitted actions are associated and encapsulated to generate phase permission configuration data. The permitted actions include task start, task hold, task migration and task end actions.

Images