A method, device and storage medium for multi-machine cooperative control of a cargo cleaning device

Abstract

The application provides a multi-machine cooperative control method, device and equipment for a cleaning equipment and a storage medium. The method comprises the following steps: when a task allocation center detects a re-distribution condition, a request is generated and an affected area (including a historical and newly-added task area) is determined, a space-time lock request containing a lock validity period is sent to the equipment in the area, and the period is set based on a historical communication delay and a time allowance. After the equipment receives the request, the path planning is frozen and a state is fed back. After the task allocation center verifies that all the equipment is ready and the linear speed is lower than a threshold, a new task set is broadcasted and the freezing is released. Through the space-time transaction lock mechanism, the application effectively solves the conflict caused by the asynchronous task re-distribution and path planning in the multi-machine cooperative control of the cleaning equipment, improves the operation safety and efficiency, and enhances the stability and robustness of the system.

Description

Technical Field

[0001] This application belongs to the field of warehouse cleaning equipment control, and in particular relates to a multi-machine collaborative control method, device, equipment and storage medium for warehouse cleaning equipment. Background Technology

[0002] In existing multi-machine collaborative control technologies for cabin clearing equipment, the task allocation process typically includes stages such as hierarchical task allocation, local equivalent transformation, distributed collaborative decision-making, and dynamic reallocation triggering. However, these methods suffer from the problem of asynchronous task relocation and local path planning during dynamic reallocation, leading to spatiotemporal overlap between old and new task paths and consequently causing motion conflicts between equipment. For example, when the task allocation center triggers dynamic reallocation, due to communication delays between equipment clusters, some equipment determines that the frozen path planning has received the new task and started execution based on the lock validity period, while other equipment is still executing the old task, resulting in unpredictable motion conflicts in the handover area. Summary of the Invention

[0003] The purpose of this application is to overcome the defects in the prior art and provide a method, apparatus, equipment and storage medium for multi-machine collaborative control of a tank cleaning device.

[0004] This application provides a multi-machine collaborative control method for tank cleaning equipment, including:

[0005] When the task allocation center detects a reassignment condition, it generates a reassignment request; based on the reassignment request, it determines the affected area, which includes historical task areas and newly added areas; it sends a time-space lock request to each device in the affected area, which includes a lock validity period, which is determined based on the maximum historical communication delay and time margin.

[0006] Upon receiving the spatiotemporal lock request, each device in the affected area determines the freeze path plan based on the lock validity period and reports its ready status.

[0007] The task allocation center determines the current linear velocity of the device from each ready state. When all devices complete the ready state feedback and each current linear velocity is less than a set low speed threshold, the task allocation center broadcasts a new task set to each device in the affected area and releases the frozen path planning.

[0008] Optionally, determining the freeze path plan and feeding back the ready status based on the lock validity period includes:

[0009] The affected area is divided into a discrete grid set;

[0010] Each device determines its spatial inclusion relationship with the discrete grid set based on its own position coordinates;

[0011] When the spatial inclusion relationship is established, confirm the frozen path planning.

[0012] Optionally, the lock validity period is determined based on the maximum historical communication delay and time margin, including:

[0013] Obtain the set of valid time data for historical lock periods;

[0014] Get the current cluster communication latency data set;

[0015] Extract the maximum delay value from the communication delay data set;

[0016] The lock validity period is obtained by weighted fusion calculation of the valid time data set and the maximum delay value.

[0017] Optionally, the task allocation center broadcasts a new task set to each device in the affected area and unfreezes the path planning, including:

[0018] Detect the spatial overlap between the faulty task area and the affected area;

[0019] When the spatial overlap relationship meets the preset geometric conditions, the frozen path planning is released.

[0020] Optionally, when all the devices complete the readiness status feedback, including:

[0021] Verifying whether all devices have reported a ready status includes: establishing a unique set of identifiers for devices that have reported a ready status; obtaining a complete set of identifiers for the initially locked devices; and performing element matching between the unique set of identifiers and the complete set of identifiers to verify whether all devices have completed the ready status feedback.

[0022] Optionally, each of the current linear velocities being less than a set low-speed threshold includes:

[0023] Acquire real-time motion vector data of the device and calculate the magnitude of the motion vector data;

[0024] Determine whether the modulus value is within a preset motion range.

[0025] Optional, also includes:

[0026] Monitor the validity period status of the lock;

[0027] When the validity period status indicates that the lock has expired, the verification process is terminated, and the task status of the device is restored to the original state before receiving the time lock request.

[0028] This application also provides a multi-machine collaborative control device for tank cleaning equipment, including:

[0029] The time-space lock module generates a reassignment request when the task allocation center detects a reassignment condition; determines the affected area based on the reassignment request, the affected area including historical task areas and newly added areas; and sends a time-space lock request to each device in the affected area, the time-space lock request including a lock validity period, the lock validity period being determined based on the historical maximum communication delay and time margin.

[0030] The feedback module determines the freeze path plan and feeds back the ready status for each device in the affected area after receiving the spatiotemporal lock request, based on the lock validity period.

[0031] The judgment module determines the current linear velocity of the device from each ready state. When all devices complete the ready state feedback and each current linear velocity is less than a set low speed threshold, the task allocation center broadcasts a new task set to each device in the affected area and releases the frozen path planning.

[0032] This application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described method.

[0033] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform the above-described method.

[0034] The beneficial effects of this application are:

[0035] This application provides a multi-machine collaborative control method for cabin cleaning equipment, comprising: a task allocation center generating a reassignment request when a reassignment condition is detected; determining an affected area based on the reassignment request, the affected area including a historical task area and a newly added area; sending a spacetime lock request to each device in the affected area, the spacetime lock request including a lock validity period, the lock validity period being determined based on the historical maximum communication delay and time margin; each device in the affected area, upon receiving the spacetime lock request, determining a frozen path plan based on the lock validity period and feeding back a ready status; the task allocation center determining the current linear velocity of the device from each ready status; when all devices complete the ready status feedback, and each current linear velocity is less than a set low-speed threshold, the task allocation center broadcasts a new task set to each device in the affected area and releases the frozen path plan. This application, by introducing a spacetime transaction lock mechanism, resolves the conflict caused by asynchronous task reassignment and path planning in multi-machine collaborative cabin cleaning equipment, improves operational safety and efficiency, and enhances system stability and robustness. Attached Figure Description

[0036] Figure 1 This is a schematic diagram of the multi-machine collaborative control process of the tank cleaning equipment in this application;

[0037] Figure 2 This is a schematic diagram of the spatiotemporal transaction lock flowchart in this application;

[0038] Figure 3 This is a schematic diagram of the spatiotemporal transaction lock mechanism interaction in this application. Detailed Implementation

[0039] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it is to be understood that various forms of implementation of the present disclosure are intended and should not be limited to the embodiments set forth herein. Rather, the embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0040] Please refer to Figures 1-3 The present application provides a multi-machine collaborative control method for a tank cleaning device, comprising:

[0041] S101. When the task allocation center detects a redistribution condition, it generates a redistribution request; based on the redistribution request, it determines the affected area, which includes historical task areas and newly added areas; it sends a time-space lock request to each device in the affected area, which includes a lock validity period, which is determined based on the historical maximum communication delay and time margin.

[0042] The task allocation center detects redistribution conditions based on a set of state-aware triggering mechanisms defined by the system. These triggering conditions are determined at three levels: individual, local, and global, ensuring that the task management strategy has local awareness and execution efficiency.

[0043] First, at the individual level, when the duration of continuous blocking of any device meets the following conditions... ≥ When this occurs, it indicates that the device has fallen into a state of path infeasibility or execution obstruction under the current task. The system determines that its execution has failed, thereby triggering the reallocation process.

[0044] The above The duration of the continuous blocking of device i, the This is the preset maximum blocking time threshold.

[0045] Secondly, at the local level, if device i senses the set of neighboring devices... The variance of the spatial distribution satisfies:

[0046]

[0047] in, For the collection of neighboring devices, The sample variance function representing the location coordinates of neighboring devices. Let j be the current position coordinate vector of the neighboring device j. =0.2 This indicates that the surrounding devices are in a spatially clustered state with a lack of dynamic changes. Based on the lock validity period, the frozen path planning is determined. Due to mutual obstruction or path overlap, the system locally stagnates. The system identifies this as a potential congestion area, thereby triggering reallocation to reorganize tasks and release paths.

[0048] Finally, at the global level, when the system detects that the number of devices in the cluster that are unable to communicate or report failures exceeds 10% of the total, the system determines that there is a tendency for large-scale failures, immediately suspends some tasks, and restructures the scheduling strategy.

[0049] These triggering conditions can individually activate the task redistribution process. The system monitors the three criteria in parallel every cycle, comprehensively ensuring the robustness and stability of the system operation from three aspects: time continuity, spatial status, and global health.

[0050] In addition to the aforementioned triggering events, spatial judgment logic linked to the locked area has been introduced, making the task management strategy more efficient.

[0051] When the system detects a blocked area The area currently under lockdown When the spatial overlap between them reaches or exceeds 60%, the following condition is met:

[0052]

[0053] The system determines that this type of blocking problem stems from an ongoing task reconstruction or path freeze state based on the lock's validity period. Therefore, it skips the regular reallocation process and directly activates the fault rescue module to quickly dispatch backup equipment to the affected area to perform recovery tasks, prioritizing system stability and task continuity.

[0054] in, This is a congested area. For the locked region, Area represents the region area calculation function.

[0055] Meanwhile, to improve overall scheduling flexibility, when a new task is dynamically inserted and the Euclidean distance between the insertion point and any current lock region satisfies... Under certain conditions, the task can be asynchronously responded to by nearby, unlocked local devices without waiting for global unlocking, thus achieving spatial isolation of task insertion and parallel response. This mechanism enables the reallocation logic to have state awareness by precisely controlling the spatial overlap ratio and distance threshold, avoiding unnecessarily triggering global resource reconfiguration during task switching.

[0056] Based on the redistribution request, the task allocation center determines the affected areas:

[0057]

[0058] Among them, the For the historical mission area, the stated For newly added areas, ensure that the affected areas cover all relevant spaces.

[0059] In the spatial dimension, the task allocation center will lock the area to be locked. Rasterization is performed to obtain a discrete raster set G={ , , ..., Each grid cell represents a passable location unit, simplifying device location determination. Wherein... These are discrete raster elements.

[0060] Subsequently, the task allocation center sends a time-space lock request to every device within the affected area, the request including the lock validity period. The lock validity period is determined based on the historical maximum communication delay and time margin, using a dynamic adjustment strategy. The system obtains a set of valid time data for historical lock periods and a set of current cluster communication delay data. , , ..., }, where τi is the communication delay time of device i. Extract the maximum delay value max( from the communication delay dataset). , , ..., The lock validity period is obtained by weighting and fusing the valid time data set and the maximum delay value.

[0061]

[0062] in, The effective time of the current lockout period. The effective time of the previous locking period, the This represents the maximum current communication delay. The time margin is reflected by a weighting coefficient of 0.2, ensuring that the lock validity period covers the synchronization requirements under the worst-case communication conditions.

[0063] This formula uses a sliding weighted average to dynamically adapt to network conditions and incorporates the maximum communication latency as a reference, ensuring that all devices can receive lock requests and task updates even under the slowest communication conditions.

[0064] The task allocation center performs this process before broadcasting new tasks to ensure lock validity. satisfy:

[0065]

[0066] Where Δt is the time margin compensation value.

[0067] By employing spatial discretization and temporal control, unified scheduling of critical areas is achieved. Regarding task scheduling priority, the system sets nested preemption rules. When a fault rescue task is triggered and its target area overlaps with the currently locked area, the system allows for forced unlocking and interruption of the original task, transferring equipment within the occupied area to obstacle avoidance standby mode. This mechanism enables rapid spatial state identification, adaptive control of time windows, and dynamic preemption of priority scheduling, improving system response efficiency.

[0068] S102. After receiving the time-space lock request, each of the devices in the affected area determines the freeze path plan according to the lock validity period and reports the ready status.

[0069] Each device in the affected area, upon receiving a spacetime lock request, determines its validity period based on the lock's expiration date. Entering frozen state.

[0070] First, each device determines its spatial inclusion relationship with the discrete grid set G based on its own location coordinates. The task allocation center has divided the affected area into discrete grid sets G={ , , ..., Each device only needs to calculate whether it is in any grid based on its current pose (including position and orientation). Inside. This occurs when the spatial containment relationship holds, i.e., the current location of the device. Upon falling into the grid set, the device confirms that it is within the locked area and performs path planning freeze. Wherein... This is the current position of device i.

[0071] The specific mechanism for freezing path planning is implemented through an equivalent transformation module. Upon receiving a spacetime lock request and confirming entry into a frozen state, this module automatically switches to a special operating mode to ensure that the device does not interfere with task reconstruction due to erroneous actions during task switching. At this time, the original dynamic potential field function no longer applies, and a simplified potential field function is adopted instead:

[0072]

[0073] in, This represents the potential energy in the frozen state. This is the obstacle repulsion coefficient, which controls the avoidance intensity for static obstacles. Location of the obstacle. This is the position retention coefficient. This indicates the location at the moment the device freezes. This is the current position of device i.

[0074] The potential field function consists of two parts: the first term is the obstacle repulsion term, which maintains basic environmental safety perception; the second term is the position-maintaining term, which minimizes the current position. Location at the time of freezing The square of the distance guides the device to stabilize in place. By minimizing this function, the device maintains a static state near its current position, making minimal avoidance only when an obstacle is detected approaching.

[0075] At the same time, the control module forcibly limits the output speed and sets an upper limit. A speed of 0.1 m / s ensures that motion response is limited to emergency obstacle avoidance. This freeze mechanism, through a combination of potential field control and velocity constraints, places the equipment in a waiting state, prohibiting it from executing any goal-based behavioral instructions without receiving mission updates. Upon entering the freeze state, the equipment immediately reports its readiness status to the mission allocation center, including its unique identifier and current status data. This feedback process is completed within the lock's validity period, ensuring synchronized responses from all equipment. This mechanism effectively supports system consistency and spatial security during mission switching.

[0076] S103. The task allocation center determines the current linear velocity of the device from each of the ready states. When all the devices complete the ready state feedback and each of the current linear velocities is less than a set low speed threshold, the task allocation center broadcasts a new task set to each of the devices in the affected area and releases the frozen path planning.

[0077] After receiving the readiness status feedback from each device, the task allocation center executes an atomicity verification mechanism to ensure the synchronization and security of task switching.

[0078] First, the task allocation center verifies whether all devices have reported a ready status by establishing a unique set of identifiers for the devices that have reported the status and matching them element by element with the complete set of identifiers for the initially locked devices. The system then counts the number of devices in the current locked area that have reported a frozen status and compares this count with the total number of devices in the initial locked area. Only when both counts exactly match does the system confirm that all relevant devices have received the lock request and successfully entered the frozen state.

[0079] Wherein, the unique identifier is the unique number of the device, and the identifier set is the set of numbers of all devices in the initial affected area.

[0080] Secondly, the task allocation center extracts the current linear velocity data of the device from each ready state and determines whether each current linear velocity is less than a set low-speed threshold. Specifically, the system acquires the device's real-time motion vector data and calculates the magnitude of the motion vector data. , wherein Let be the linear velocity vector of device i. Determine if the magnitude is within the preset motion range, i.e., satisfy . ≤0.2m / s. This threshold is set below the low-speed obstacle avoidance range to ensure the device is nearly stationary in space, allowing only necessary minute movements to avoid obstacles or maintain stability. Wherein... This represents the linear velocity magnitude.

[0081] When the above two verifications are within the lock validity period When all processes pass (i.e., all devices have completed their readiness status feedback and each current linear velocity is less than the low-speed threshold), the task allocation center broadcasts a new task set to every device in the affected area. The new task set includes updated subtask target locations and path instructions. Subsequently, the system unfreezes path planning, the equivalent transformation module restores the original dynamic potential field function, and the devices generate local behavior instructions based on the new tasks. The original dynamic potential field function is:

[0082]

[0083] Among them, the For device i at its current location The energy of the potential field that is felt For the target gravitational coefficient, the For the mission objective point, The coefficient of repulsion force of the obstacle. Location of the obstacle. The repulsion coefficient between equipment. For the collection of neighboring devices, Set a value for the safety distance. Let j be the location of the neighboring device.

[0084] Repulsion coefficient Dynamically adjust based on task relevance:

[0085]

[0086] in, This represents the maximum value of the repulsion coefficient. β represents the minimum repulsion coefficient, and β is the decay rate parameter. Let be the angle between the target direction vectors of devices i and j. This is the preset included angle threshold.

[0087] The device's motion direction and velocity control vector are obtained by calculating the negative gradient of the total potential field function.

[0088]

[0089] in, This is the control vector for the direction and speed of the equipment's movement.

[0090] Each device adjusts its speed and path based on its real-time location in the neighborhood using an adaptive consensus protocol to satisfy:

[0091] and

[0092] The collaborative constraints were used to complete the cabin clearing mission.

[0093] in, The speed difference between device i and device j. For speed threshold, Let be the position coordinate vector of device i at time t. This is the target point's position vector.

[0094] This strategy combines angle judgment with target spatial distance to dynamically adjust the repulsion weight, balancing safety and cooperation.

[0095] Before broadcasting a new task set, the system detects the spatial overlap between the faulty task region and the affected region. When the spatial overlap meets preset geometric conditions, such as a blocked region... With locked area The degree of overlap satisfies:

[0096]

[0097] Instead of broadcasting a new task set, the frozen path planning is directly unfrozen to prioritize handling emergency events. The preset geometric condition mentioned here is a spatial overlap threshold.

[0098] Furthermore, the validity period of the lock is monitored. When the validity period status indicates that the lock has expired, that is, when the current time exceeds [a certain threshold], [the lock will be released]. The system terminates the verification process and triggers a timeout rollback mechanism. Following the principle of minimum disturbance, the system revoks the task update instruction, restoring the frozen device's task state to its original state and path plan before receiving the time-space lock request. This prevents partial updates from causing task inconsistencies, path overlaps, or scheduling conflicts within the system. This mechanism ensures overall operational stability through time limit control.

[0099] The entire implementation process concludes with the execution of the new task set. The equipment then enters the local behavior generation phase, where subtasks are converted into local behavior instruction sets based on the equivalent transformation module, enabling multi-machine collaborative control. In the hierarchical allocation phase, the task allocation center, based on the ship's 3D map and equipment cluster status, decomposes the global task into subtask sets using a hierarchical strategy and allocates them to individual devices according to regional priority. Regional division is achieved through connected component analysis.

[0100]

[0101] Among them, the A subregion that has spatial connectivity.

[0102] Weight allocation uses an adaptive weighting function:

[0103]

[0104] in, For device i to area The adaptation weight, This represents the remaining battery power of device i. Let i be the current position of device i. For the region The geometric center, where ε is a minimal constant to prevent division by zero. The task binding layer is based on the maximum weight allocation principle: .

[0105] This achieves an efficient and balanced allocation of initial tasks. This foundation provides task boundaries and objectives for the redistribution phase. All formulas and details are derived from the disclosed information, ensuring the completeness and atomicity of implementation.

[0106] This application also provides a multi-machine collaborative control device for tank cleaning equipment, including:

[0107] The time-space lock module generates a reassignment request when the task allocation center detects a reassignment condition; determines the affected area based on the reassignment request, the affected area including historical task areas and newly added areas; and sends a time-space lock request to each device in the affected area, the time-space lock request including a lock validity period, the lock validity period being determined based on the historical maximum communication delay and time margin.

[0108] The feedback module determines the freeze path plan and feeds back the ready status for each device in the affected area after receiving the spatiotemporal lock request, based on the lock validity period.

[0109] The judgment module determines the current linear velocity of the device from each ready state. When all devices complete the ready state feedback and each current linear velocity is less than a set low speed threshold, the task allocation center broadcasts a new task set to each device in the affected area and releases the frozen path planning.

[0110] This application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described method.

[0111] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform the above-described method.

[0112] The above description of the embodiments is provided to enable those skilled in the art to understand and apply this application. Those skilled in the art will readily make various modifications to the above embodiments and apply the general principles described herein to other embodiments without inventive effort. Therefore, this application is not limited to the above embodiments, and any improvements and modifications made to this application based on the disclosure thereof should be within the scope of protection of this application.

Claims

1. A multi-machine collaborative control method for a tank cleaning system, characterized in that, include: When the task allocation center detects a reassignment condition, it generates a reassignment request; The affected areas are determined based on the reallocation request, and the affected areas include historical task areas and newly added areas; A time-space lock request is sent to each device within the affected area. The time-space lock request includes a lock validity period, which is determined based on the historical maximum communication delay and time margin. The process includes: obtaining the time margin data set; obtaining the current cluster communication delay data set; extracting the maximum delay value from the communication delay data set; and performing a weighted fusion calculation on the valid time data set and the maximum delay value to obtain the lock validity period. The time margin is the valid time of a historical locking period. Upon receiving the spatiotemporal lock request, each device in the affected area determines the freeze path plan based on the lock validity period and reports its ready status. The task allocation center determines the current linear velocity of the device from each ready state. When all devices complete the ready state feedback and each current linear velocity is less than a set low speed threshold, the task allocation center broadcasts a new task set to each device in the affected area and releases the frozen path planning.

2. The method according to claim 1, characterized in that, Determine the freeze path planning based on the lock validity period and provide feedback on the ready status, including: The affected area is divided into a discrete grid set; Each device determines its spatial inclusion relationship with the discrete grid set based on its own position coordinates; When the spatial inclusion relationship is established, confirm the frozen path planning.

3. The method according to claim 1, characterized in that, The task allocation center broadcasts a new task set to each device in the affected area and unfreezes the path planning, including: Detect the spatial overlap between the faulty task area and the affected area; When the spatial overlap relationship meets the preset geometric conditions, the frozen path planning is released.

4. The method according to claim 1, characterized in that, When all the aforementioned devices complete the readiness status feedback, including: Verifying whether all devices have reported a ready status includes: establishing a unique set of identifiers for devices that have reported a ready status; obtaining a complete set of identifiers for the initially locked devices; and performing element matching between the unique set of identifiers and the complete set of identifiers to verify whether all devices have completed the ready status feedback.

5. The method according to claim 1, characterized in that, And each of the aforementioned current linear velocities being less than a set low-speed threshold includes: Acquire real-time motion vector data of the device and calculate the magnitude of the motion vector data; Determine whether the modulus value is within a preset motion range.

6. The method according to claim 1, characterized in that, Also includes: Monitor the validity period status of the lock; When the validity period status indicates that the lock has expired, the verification process is terminated, and the task status of the device is restored to the original state before receiving the time lock request.

7. A multi-machine collaborative control device for a tank cleaning system, characterized in that, include: The time-space lock module generates a reassignment request when the task allocation center detects a reassignment condition; The affected areas are determined based on the reallocation request, and the affected areas include historical task areas and newly added areas; A time-space lock request is sent to each device within the affected area. The time-space lock request includes a lock validity period, which is determined based on the historical maximum communication delay and time margin. The process includes: obtaining the time margin data set; obtaining the current cluster communication delay data set; extracting the maximum delay value from the communication delay data set; and performing a weighted fusion calculation on the valid time data set and the maximum delay value to obtain the lock validity period. The time margin is the valid time of a historical locking period. The feedback module determines the freeze path plan and feeds back the ready status for each device in the affected area after receiving the spatiotemporal lock request, based on the lock validity period. The judgment module determines the current linear velocity of the device from each ready state. When all devices complete the ready state feedback and each current linear velocity is less than a set low speed threshold, the task allocation center broadcasts a new task set to each device in the affected area and releases the frozen path planning.

8. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed in a computer, causes the computer to perform the method described in any one of claims 1-7.

Images