AGV cluster charging scheduling and task coordination method based on energy consumption prediction

Abstract

The application relates to the technical field of AGV cluster control, in particular to an AGV cluster charging scheduling and task coordination method based on energy consumption prediction, which comprises the following steps which are cyclically executed: acquiring AGV state data, task attribute data and charging station state data; predicting future energy consumption by using a hybrid energy consumption prediction model, outputting a residual power change curve and marking a low power moment; calculating a comprehensive charging urgency index to mark AGVs to be charged and select optimal charging piles for the AGVs to be charged; triggering a relay task mechanism for the AGVs to be charged which have not completed tasks, so that idle AGVs undertake the tasks; dynamically allocating charging gun power and updating the AGV state to be idle after charging. The application solves the problem of AGV power failure and shutdown in the middle of the way through closed-loop scheduling, and improves the operation continuity and overall operation efficiency of the cluster.

Description

Technical Field

[0001] This application relates to the field of AGV cluster control technology, specifically to an AGV cluster charging scheduling and task coordination method based on energy consumption prediction. Background Technology

[0002] AGVs (Automated Guided Vehicles), as core execution units in intelligent logistics and flexible manufacturing systems, have been widely used in industrial scenarios requiring high-frequency material handling, such as e-commerce warehousing, semiconductor manufacturing, and automobile assembly. In these scenarios, AGV clusters composed of multiple AGVs typically need to operate 24 hours a day without interruption, and their power management directly affects the continuity and stability of the entire production system. With the continuous acceleration of production cycles and the increasing complexity of handling tasks, the power scheduling problem of AGV clusters has evolved from a simple energy replenishment requirement to a comprehensive scheduling problem involving task coordination, resource allocation, and energy consumption optimization.

[0003] In existing technologies, a typical AGV cluster charging scheduling method is to set a fixed power percentage threshold for each AGV. When an AGV detects that its current remaining power is lower than the threshold, it immediately stops the currently executing transport task and triggers a charging request. Then, it reports its location and request information to the scheduling center via wireless communication. The scheduling center allocates the nearest available charging pile to the AGV based on the real-time occupancy of each charging station. The AGV travels to the charging pile according to the planned path to complete the charging. After being fully charged, it rejoins the task queue to wait for the next task allocation.

[0004] However, in real-world complex operating conditions, the actual energy consumption rate of AGVs during long-distance, heavy-load, or frequent start-stop complex tasks differs significantly from their idle state. The aforementioned fixed-threshold triggering scheme only acquires current battery information at the moment of charging triggering, failing to predict energy consumption demands for the remaining distance of the current task or detect changes in discharge characteristics caused by battery health degradation. Therefore, when an AGV performs a task with an energy consumption rate exceeding expectations, it is highly likely to run out of power before reaching the charging station or precisely at the midpoint of the task, forcing material handling to be interrupted and equipment to stop. This problem is particularly prominent in large-scale AGV clusters and high-cycle production lines, because a sudden stop of a single AGV not only affects its own workstation but also triggers a chain reaction due to broken material connections, leading to a sharp decline in the turnover efficiency of the entire production line or storage area. Summary of the Invention

[0005] To address the technical problem that existing AGV cluster charging scheduling methods based on fixed threshold triggers cannot predict the actual energy consumption rate of AGVs during heavy-load or long-distance tasks, leading to AGVs often running out of power and suddenly stopping during task execution, causing material handling interruptions, this application provides an AGV cluster charging scheduling and task coordination method based on energy consumption prediction. By predicting the future energy consumption of each AGV in advance to identify AGVs that are about to run out of power, the optimal charging pile is planned for each AGV by comprehensively considering its power danger, task urgency, and charging station congestion. At the same time, idle AGVs are triggered to take over its unfinished tasks. During the charging phase, power is dynamically allocated according to the urgency, forming a closed-loop scheduling.

[0006] Firstly, this application provides a method for AGV cluster charging scheduling and task coordination based on energy consumption prediction, including the following steps executed in a loop: S1. Obtain the AGV status data, task attribute data, and charging station status data for each AGV in the AGV cluster; The AGV status data includes the current working status, remaining power percentage, motion parameters, load parameters, environmental parameters, and historical energy consumption records. The task attribute data includes the remaining task distance and task deadline. The charging station status data includes the queue length of each charging pile. S2. Based on the AGV status data of each AGV, use the hybrid energy consumption prediction model to predict the energy consumption of each AGV in the future period, output the predicted energy consumption value, calculate the remaining power change curve of each AGV based on the predicted energy consumption value, and mark the time when its power is lower than the safety threshold. The hybrid energy consumption prediction model includes a physical mechanism-based energy consumption analytical model and a pre-trained residual compensation model. The energy consumption analytical model outputs a theoretical energy consumption baseline, while the residual compensation model, built on a neural network, outputs a residual compensation value. The hybrid energy consumption prediction model adds the theoretical energy consumption baseline output by the energy consumption analytical model to the residual compensation value output by the residual compensation model to obtain the predicted energy consumption value. S3. Based on the remaining power change curve, task attribute data and charging station status data of each AGV, calculate the comprehensive charging urgency index of each AGV, mark the AGV with the comprehensive charging urgency index exceeding the preset urgency index threshold as the AGV to be charged, and select the optimal charging station for each AGV to be charged that minimizes the weighted sum of the empty driving energy consumption to the charging station and the expected queuing waiting time. S4. For AGVs that are waiting to be charged but whose current tasks are not yet completed, select candidate AGVs from the surrounding preset radius. Under the constraint that the remaining power after taking on the task is higher than the safety value, select the candidate AGV with the smallest energy consumption increment to take on the unfinished task and issue a takeover instruction. Among them, the candidate AGV is the AGV whose current working state is idle; S5. Control each AGV to be charged to arrive at its corresponding charging pile, and then dynamically allocate the output power of each charging gun according to the comprehensive charging urgency index of all AGVs currently charging and those in the queue. After charging is completed, each AGV to be charged updates its current working status to idle status.

[0007] It should be further noted that the input to the energy consumption analysis model is the current travel speed from the AGV status data. , running path slope Load weight Ambient temperature Instantaneous value of motor current and remaining battery percentage The output is the theoretical energy consumption baseline per unit time step. ; The data processing flow of the energy consumption analysis model is as follows: S201. Calculate the mechanical power required for the AGV to overcome resistance per unit time based on the dynamic equation. The formula is:

[0008] in, The total mass of the current AGV is calculated using the following formula:

[0009] The current unloaded mass of the AGV. This represents the current load weight of the AGV. It is the acceleration due to gravity. The coefficient of rolling friction is air density, This is the drag coefficient. For windward area; S202. Calculate the theoretical energy consumption baseline of the current AGV. The calculation formula is:

[0010] in, For a unit time step, Let I be the motor efficiency function with respect to the current driving speed v and the instantaneous value of the motor current I. For information regarding remaining battery percentage, State of Charge (SOC), and ambient temperature The battery discharge efficiency function; The residual compensation model is a long short-term memory neural network, whose input is the past... The residual sequence at each time step ,in The time window length, ; For the first The actual energy consumption per time step The theoretical energy consumption baseline is output by the energy consumption analytical model at the same time step; the output of the residual compensation model is the future energy consumption baseline. Residual compensation value at each time step ,in To predict the number of time steps, For the first The residual compensation value for each prediction time step; The hybrid energy consumption prediction model adds the theoretical energy consumption baseline output by the energy consumption analytical model to the residual compensation value output by the residual compensation model to obtain the predicted energy consumption value. The calculation formula is as follows:

[0011] in For the first Predicted energy consumption for each prediction time step. For the first The theoretical energy consumption baseline for each predicted time step. For the first The residual compensation value for each prediction time step.

[0012] It should be further explained that the pre-training process of the residual compensation model includes: Extract the actual energy consumption at each time step from the historical operation data of the AGV. Compare actual energy consumption with the theoretical energy consumption baseline at the same time step. The residual labels are obtained by subtraction. ; Initialize the residual compensation model to continuously The residual sequence at each time step is used as the training sample input to initialize the residual compensation model, and then... The residual labels at each time step are used as the prediction target for training. During the training process, the backpropagation algorithm and the adaptive moment estimation optimizer are used to train the network parameters of the residual compensation model, so as to minimize the mean square error between the predicted residual and the residual label, until the loss function converges on the validation set, and the pre-trained residual compensation model is obtained.

[0013] It should be further explained that in step S2, the energy consumption of each AGV in the future time period is predicted using a pre-trained hybrid energy consumption prediction model to obtain the predicted energy consumption value. Then, the remaining power change curve is calculated based on the predicted energy consumption value, and the time when the power is lower than the safety threshold is marked. Specifically, this includes: The current travel speed in the AGV status data , running path slope Load weight Ambient temperature Instantaneous value of motor current and remaining battery percentage Input the hybrid energy consumption prediction model and output the predicted energy consumption value; The remaining electricity consumption change curve is calculated by accumulating the predicted energy consumption values. The calculation formula is as follows:

[0014] in For the first The percentage of remaining battery power at each predicted time step. This represents the current percentage of battery remaining. For the first Predicted energy consumption for each prediction time step. This refers to the battery's rated capacity. Will satisfy The earliest time is marked as the moment when the battery level falls below the safe threshold, where This is a preset safe power threshold.

[0015] It should be further explained that in step S2, the hybrid energy consumption prediction model is executed in a rolling manner with a fixed time window. In each scheduling cycle, the latest acquired AGV status data is input into the model to update the final predicted energy consumption value, the remaining power change curve, and the marked time.

[0016] It should be further noted that in step S3, the comprehensive charging urgency index... The calculation formula is:

[0017] in, The preset weighting coefficient for the risk level of electricity consumption. The preset task urgency weighting coefficient, The preset weighting coefficient for the charging station congestion penalty factor; Electricity risk level The calculation formula is:

[0018] in, For the Sigmoid function, To predict the amount of electricity needed to reach the next charging station, This is the preset scaling factor; Task urgency The calculation formula is:

[0019] in, The remaining time is calculated using the following formula:

[0020] The deadline for the task. For the current moment, The formula for estimating travel time is as follows:

[0021] The remaining distance for the mission. Average driving speed This is the preset sensitivity coefficient; Charging station congestion penalty factor The calculation formula is:

[0022] This refers to the queue length of the available charging stations. This represents the total number of charging guns at the charging station. This is the preset magnification factor.

[0023] It should be further explained that, , , The fitness function is determined offline through a genetic algorithm based on historical operational data. The optimization objective is to minimize the weighted sum of the weighted task interruption rate and charging wait time.

[0024] in, For task interruption rate, This refers to the average charging wait time. The preset penalty coefficient for task interruption rate. A preset penalty coefficient for average charging wait time.

[0025] It should be further noted that in step S3, the selection of the optimal charging station is based on minimizing the objective function. The calculation formula is as follows:

[0026] in, For the first The AGV waiting to be charged arrived at the first The distance between charging stations This refers to the mechanical power of the AGV when it is running unloaded. Average driving speed This is the energy consumption conversion factor for empty driving. For charging piles The estimated waiting time Time cost weighting.

[0027] It should be further explained that, The estimation is achieved by virtually scheduling the currently queued AGVs according to their respective comprehensive charging urgency indices.

[0028] It should be further noted that in step S4, the Hungarian algorithm is used to solve for the task allocation matrix and cost matrix elements. The calculation formula is:

[0029] in, For the first The remaining task distance of the AGV waiting to be charged. For the first The journey of the candidate AGVs to the handover point. For the first The average driving speed of the candidate AGVs For loading and unloading time.

[0030] It should be further noted that in step S4, the constraint is that the remaining power of the candidate AGV after undertaking the task is higher than the safety value, and the formula is:

[0031] in, For the first The current remaining battery percentage of each candidate AGV. For the first The amount of electricity required for each candidate AGV to undertake the task. For the battery's rated capacity, This is a preset safe power threshold.

[0032] It should be further explained that in step S4, among the feasible solutions that satisfy the constraints, the energy consumption increment is selected. The minimum allocation scheme, The calculation formula is:

[0033] in, For candidate number Increased energy consumption of candidate AGVs due to undertaking additional tasks.

[0034] It should be further noted that in step S5, the output power of each charging gun is dynamically allocated using a proportional-integral controller, and the control law formula is as follows:

[0035] in, This is the serial number of the charging gun that is currently charging. For the first A charging gun in Output power at any moment As the reference power, This is the proportionality coefficient. The integral coefficient is... For the first The overall charging urgency index of each charging gun for the AGV currently charging. The average comprehensive charging urgency index for all AGVs currently charging is τ, where τ is the integral variable.

[0036] It should be further noted that in step S5, the allocation process simultaneously satisfies both the total power constraint and the minimum charging current constraint for each AGV. The formula for the total power constraint is:

[0037] in, This refers to the upper limit of the transformer capacity for charging stations; The formula for the minimum charging current constraint of AGV is:

[0038] in, Battery voltage, This is the minimum sustaining current required by the battery management system.

[0039] It should be further noted that step S5 also includes monitoring the rate of battery recovery for each AGV during the charging process. and battery temperature When the remaining battery percentage of the AGV reaches the preset safe battery threshold... When its overall charging urgency index ranking drops below that of newly entered AGVs waiting to be charged at the charging station, the output power of each charging gun is recalculated and adjusted according to the power distribution control law.

[0040] Secondly, this application provides an AGV cluster charging scheduling and task coordination system based on energy consumption prediction, used to implement the above-mentioned AGV cluster charging scheduling and task coordination method, including: The data acquisition module is used to acquire AGV status data, task attribute data, and charging station status data for each AGV in the AGV cluster. The remaining power consumption curve generation module is used to predict the energy consumption of each AGV in the future period based on the AGV status data of each AGV and using a hybrid energy consumption prediction model. It outputs the predicted energy consumption value, calculates the remaining power consumption curve of each AGV based on the predicted energy consumption value, and marks the time when its power consumption is lower than the safety threshold. The optimal charging pile selection module is used to calculate the comprehensive charging urgency index of each AGV based on the remaining power change curve, task attribute data and charging station status data of each AGV, mark the AGV with the comprehensive charging urgency index exceeding the preset urgency index threshold as the AGV to be charged, and select the optimal charging pile for each AGV to be charged that minimizes the weighted sum of the empty driving energy consumption to the charging pile and the expected queuing waiting time. The task handover module is used to select candidate AGVs from a preset radius for AGVs that are waiting to be charged but whose current tasks have not been completed. Under the constraint that the remaining power after taking on the task is higher than the safety value, the module selects the candidate AGV with the smallest energy consumption increment to take on the unfinished task and issues a handover instruction. The charging management module controls each AGV to arrive at its corresponding charging pile. Then, it dynamically allocates the output power of each charging gun based on the comprehensive charging urgency index of all AGVs currently charging and those in the queue. After charging is completed, each AGV updates its current working status to idle status.

[0041] Thirdly, this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the above-described AGV cluster charging scheduling and task coordination method.

[0042] Fourthly, this application provides a storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described AGV cluster charging scheduling and task coordination method.

[0043] As can be seen from the above technical solutions, this application has the following advantages: 1. This application acquires AGV status data and uses a hybrid energy consumption prediction model to predict the energy consumption of each AGV in future time periods, outputs the remaining power change curve and marks low power times. Based on this, a comprehensive charging urgency index is calculated to mark AGVs to be charged and select the optimal charging pile for them. At the same time, a relay task mechanism is triggered for AGVs that have not completed their current tasks to achieve seamless task continuation. During the charging phase, the output power of each charging gun is dynamically allocated according to the comprehensive charging urgency index, and the AGV working status is updated to idle after charging is completed. The above steps are executed cyclically to form a complete closed-loop control link, which solves the problem that existing technologies lack the ability to predict future energy consumption trends and task continuation mechanisms, causing AGVs to run out of power and be forced to stop in the middle of complex tasks, resulting in material flow interruptions. This enables AGV clusters to accurately judge charging timing, seamless task continuation, and reasonable allocation of charging power in continuous operation scenarios, significantly improving the operational continuity and overall operating efficiency of AGV clusters in high-cycle production environments.

[0044] 2. This application calculates a comprehensive charging urgency index by weighting and summing factors of power risk, task urgency, and charging station congestion penalty factor. Based on this index, AGVs exceeding a preset threshold are marked as AGVs awaiting charging. Then, for each AGV awaiting charging, the optimal charging pile that minimizes the weighted sum of empty running energy consumption and expected queuing waiting time is selected. This allows the charging decision to simultaneously reflect the AGV's own power risk, the urgency of the current task, and the real-time congestion status of the charging station. This avoids queuing congestion and ineffective empty runs caused by multiple AGVs rushing to the same charging pile at the same time, and reduces overall energy consumption and single charging waiting time.

[0045] 3. This application employs a relay task mechanism to trigger AGVs waiting to be charged that have not yet completed their current tasks. It selects candidate AGVs currently in an idle state from within a preset radius of the AGV. Under the constraint that the remaining battery power after taking over is higher than a safe value, the candidate AGV with the smallest energy consumption increment is selected to take over the unfinished task. Simultaneously, the output power of each charging gun is dynamically allocated based on the comprehensive charging urgency index of all currently charging and queued AGVs. After charging is completed, the current working state of the AGV is updated to idle. This collaborative mechanism of charging decision-making, task succession, and power allocation ensures that AGVs waiting to be charged can leave the site for charging without waiting to complete their current tasks, and that unfinished tasks are seamlessly taken over by other idle AGVs. This solves the problem of material flow stagnation and production rhythm disruption caused by forced task interruption in existing technologies, ensuring the continuity of production tasks and on-time delivery rate. Attached Figure Description

[0046] To more clearly illustrate the technical solution of this application, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a flowchart of an AGV cluster charging scheduling and task coordination method based on energy consumption prediction in one embodiment of this application.

[0048] Figure 2 This is a schematic block diagram of an AGV cluster charging scheduling and task coordination system based on energy consumption prediction in one embodiment of this application.

[0049] Figure 3 This is a schematic diagram of the hardware structure of an electronic device in one embodiment of this application. Detailed Implementation

[0050] To make the purpose, features, and advantages of this application more apparent and understandable, specific embodiments and accompanying drawings will be used to clearly and completely describe the technical solution protected by this application. Obviously, the embodiments described below are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0051] The following describes in detail the AGV cluster charging scheduling and task coordination method involved in this application. Specific details such as particular system structures and technologies are presented for illustrative purposes rather than limiting, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application can also be implemented in other embodiments without these specific details.

[0052] In the AGV cluster charging scheduling and task coordination method involved in this application, the term "comprising" indicates the presence of the described feature, whole, step, operation, element, and / or component, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or sets thereof. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0053] To facilitate a clear description of the technical solutions of this application, the terms "first" and "second" are used to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that the terms "first" and "second" do not necessarily imply that they are different.

[0054] The terms "one embodiment" or "some embodiments" used in this application mean that one or more embodiments of this application include the specific features, structures, or characteristics described in that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this application do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized.

[0055] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0056] The AGV cluster charging scheduling and task coordination method provided in this application embodiment is executed by a computer device. Correspondingly, the AGV cluster charging scheduling and task coordination system based on energy consumption prediction runs in the computer device.

[0057] Figure 1 This is a flowchart of an embodiment of the AGV cluster charging scheduling and task coordination method based on energy consumption prediction, according to one embodiment of this application. Figure 1 The executing entity can be an AGV cluster charging scheduling and task coordination system. Depending on different requirements, the order of the steps in this flowchart can be changed, and some can be omitted.

[0058] like Figure 1 As shown, the AGV cluster charging scheduling and task coordination method based on energy consumption prediction includes the following steps executed in a loop: Step S1: Obtain the AGV status data, task attribute data, and charging station status data for each AGV in the AGV cluster; The AGV status data includes the current working status, remaining battery percentage, motion parameters, load parameters, environmental parameters, and historical energy consumption records. The task attribute data includes the remaining task distance and task deadline. The charging station status data includes the queue length of each charging pile.

[0059] By acquiring AGV status data, task attribute data, and charging station status data for each AGV in the AGV cluster, including AGV status data such as current working status, remaining battery percentage, motion parameters, load parameters, environmental parameters, and historical energy consumption records; task attribute data such as remaining task distance and task deadline; and charging station status data such as queue length of each charging pile, the scheduling system can comprehensively obtain all the input information required for subsequent charging scheduling decisions. This provides a sufficient data foundation for the accurate calculation of energy consumption prediction models, the reasonable assessment of comprehensive urgency indices, and the effective triggering of relay task mechanisms.

[0060] Step S2: Based on the AGV status data of each AGV, use the hybrid energy consumption prediction model to predict the energy consumption of each AGV in the future period, output the predicted energy consumption value, calculate the remaining power change curve of each AGV based on the predicted energy consumption value, and mark the time when its power is lower than the safety threshold. The hybrid energy consumption prediction model includes a physical mechanism-based energy consumption analytical model and a pre-trained residual compensation model. The energy consumption analytical model outputs a theoretical energy consumption baseline, while the residual compensation model, built on a neural network, outputs a residual compensation value. The hybrid energy consumption prediction model adds the theoretical energy consumption baseline output by the energy consumption analytical model to the residual compensation value output by the residual compensation model to obtain the predicted energy consumption value.

[0061] By using AGV status data for each AGV and employing a hybrid energy consumption prediction model that includes a physical mechanism-based energy consumption analysis model and a neural network-based pre-trained residual compensation model, the energy consumption of each AGV in future time periods is predicted and the predicted energy consumption value is output. Then, based on the predicted energy consumption value, the remaining power change curve of each AGV is calculated and the time when its power is below the safety threshold is marked. This allows the scheduling system to perceive the energy consumption change trend and potential risks of AGVs in advance before they reach the low power state, thereby fundamentally avoiding the situation where AGVs run out of power and are forced to stop in the middle of a complex task due to a lack of prediction capabilities.

[0062] In some specific embodiments, the input to the energy consumption analysis model is the current travel speed from the AGV status data. , running path slope Load weight Ambient temperature Instantaneous value of motor current and remaining battery percentage The output is the theoretical energy consumption baseline per unit time step. ; The data processing flow of the energy consumption analysis model is as follows: S201. Calculate the mechanical power required for the AGV to overcome resistance per unit time based on the dynamic equation. The formula is:

[0063] in, The total mass of the current AGV is calculated using the following formula:

[0064] The current unloaded mass of the AGV. This represents the current load weight of the AGV. It is the acceleration due to gravity. The coefficient of rolling friction is air density, This is the drag coefficient. For windward area; S202. Calculate the theoretical energy consumption baseline of the current AGV. The calculation formula is:

[0065] in, For a unit time step, Let I be the motor efficiency function with respect to the current driving speed v and the instantaneous value of the motor current I. For information regarding remaining battery percentage, State of Charge (SOC), and ambient temperature The battery discharge efficiency function; The residual compensation model is a long short-term memory neural network, whose input is the past... The residual sequence at each time step ,in The time window length, ; For the first The actual energy consumption per time step The theoretical energy consumption baseline is output by the energy consumption analytical model at the same time step; the output of the residual compensation model is the future energy consumption baseline. Residual compensation value at each time step ,in To predict the number of time steps, For the first The residual compensation value for each prediction time step; The hybrid energy consumption prediction model adds the theoretical energy consumption baseline output by the energy consumption analytical model to the residual compensation value output by the residual compensation model to obtain the predicted energy consumption value. The calculation formula is as follows:

[0066] in For the first Predicted energy consumption for each prediction time step. For the first The theoretical energy consumption baseline for each predicted time step. For the first The residual compensation value for each prediction time step.

[0067] By setting a physical mechanism-based energy consumption analytical model in the hybrid energy consumption prediction model, this model calculates the mechanical power required to overcome resistance per unit time based on the dynamic equation. Its calculation integrates rolling friction components, slope resistance components, and air resistance components, enabling the theoretical energy consumption baseline to be accurately calculated based on specific working conditions such as the real-time driving speed of the AGV, the slope of the running path, and the load weight. At the same time, by adding the theoretical energy consumption baseline with the residual compensation value output by the long short-term memory neural network, the final predicted energy consumption value is obtained. This ensures that the hybrid prediction result both follows physical laws and can correct for nonlinear disturbance factors in actual operation.

[0068] In some specific embodiments, the pre-training process of the residual compensation model includes: Extract the actual energy consumption at each time step from the historical operation data of the AGV. Compare actual energy consumption with the theoretical energy consumption baseline at the same time step. The residual labels are obtained by subtraction. ; Initialize the residual compensation model to continuously The residual sequence at each time step is used as the training sample input to initialize the residual compensation model, and then... The residual labels at each time step are used as the prediction target for training. During the training process, the backpropagation algorithm and the adaptive moment estimation optimizer are used to train the network parameters of the residual compensation model, so as to minimize the mean square error between the predicted residual and the residual label, until the loss function converges on the validation set, and the pre-trained residual compensation model is obtained.

[0069] By extracting the actual energy consumption of each time step from the historical operation data of the AGV during the pre-training process of the residual compensation model, the residual label is obtained by subtracting the actual energy consumption from the theoretical energy consumption baseline at the same time step. The residual sequence of N consecutive time steps is used as the training sample input into the residual compensation model, and the residual label of the next k time steps is used as the prediction target for training. The backpropagation algorithm and adaptive moment estimation optimizer are used to train the network parameters to minimize the mean square error between the predicted residual and the residual label. This enables the residual compensation model to effectively learn from historical data the influence of nonlinear factors such as battery aging, uneven road surface, and wind resistance fluctuations on energy consumption that cannot be covered by the theoretical model.

[0070] In some specific embodiments, a pre-trained hybrid energy consumption prediction model is used to predict the energy consumption of each AGV in future time periods to obtain predicted energy consumption values. Then, the remaining power consumption change curve is calculated based on the predicted energy consumption values, and the time when the power consumption is lower than the safety threshold is marked. Specifically, this includes: The current travel speed in the AGV status data , running path slope Load weight Ambient temperature Instantaneous value of motor current and remaining battery percentage Input the hybrid energy consumption prediction model and output the predicted energy consumption value; The remaining electricity consumption change curve is calculated by accumulating the predicted energy consumption values. The calculation formula is as follows:

[0071] in For the first The percentage of remaining battery power at each predicted time step. This represents the current percentage of battery remaining. For the first Predicted energy consumption for each prediction time step. This refers to the battery's rated capacity. Will satisfy The earliest time is marked as the moment when the battery level falls below the safe threshold, where This is a preset safe power threshold.

[0072] By using a pre-trained hybrid energy consumption prediction model to predict the energy consumption of each AGV in the future, the model inputs the current driving speed, running path slope, load weight, ambient temperature, instantaneous value of motor current, and percentage of remaining power from the AGV status data to output the predicted energy consumption value. The model then calculates the remaining power change curve based on the predicted energy consumption value and marks the earliest time when the remaining power is below the preset safety threshold as the time when the power is below the safety threshold. This allows the scheduling system to accurately locate the time node when each AGV is about to reach the critical power level and thus formulate a charging plan in advance.

[0073] In some specific embodiments, the hybrid energy consumption prediction model is executed in a rolling manner with a fixed time window. In each scheduling cycle, the latest acquired AGV status data is input into the model to update the final predicted energy consumption value, the remaining power change curve, and the marked time.

[0074] By enabling the hybrid energy consumption prediction model to be executed in a rolling manner with a fixed time window, the latest acquired AGV status data is input into the model and the final predicted energy consumption value, remaining power change curve, and marker time are updated in each scheduling cycle. This allows the energy consumption prediction results to be continuously revised as the AGV operating conditions change in real time, ensuring the timeliness and accuracy of the prediction information in a continuously changing dynamic operating environment.

[0075] Step S3: Based on the remaining power change curve, task attribute data and charging station status data of each AGV, calculate the comprehensive charging urgency index of each AGV, mark the AGVs whose comprehensive charging urgency index exceeds the preset urgency index threshold as AGVs to be charged, and select the optimal charging station for each AGV to be charged that minimizes the weighted sum of the empty driving energy consumption to the charging station and the expected queuing waiting time.

[0076] By calculating the comprehensive charging urgency index of each AGV based on its remaining power change curve, task attribute data, and charging station status data, and marking AGVs with a comprehensive charging urgency index exceeding a preset threshold as AGVs to be charged, the charging trigger decision can simultaneously consider the AGV's own power risk, the urgency of the current task, and the real-time queuing congestion of the charging station. At the same time, by selecting the optimal charging station for each AGV to be charged that minimizes the weighted sum of the empty driving energy consumption to the charging station and the expected queuing waiting time, the AGV will not consume too much extra power by choosing a charging station that is too far away, nor will it be stuck in the charging station area for a long time by choosing a charging station with an excessively long queue.

[0077] In some specific embodiments, the comprehensive charging urgency index The calculation formula is:

[0078] in, The preset weighting coefficient for the risk level of electricity consumption. The preset task urgency weighting coefficient, The preset weighting coefficient for the charging station congestion penalty factor; Electricity risk level The calculation formula is:

[0079] in, For the Sigmoid function, To predict the amount of electricity needed to reach the next charging station, This is the preset scaling factor; Task urgency The calculation formula is:

[0080] in, The remaining time is calculated using the following formula:

[0081] The deadline for the task. For the current moment, The formula for estimating travel time is as follows:

[0082] The remaining distance for the mission. Average driving speed This is the preset sensitivity coefficient; Charging station congestion penalty factor The calculation formula is:

[0083] This refers to the queue length of the available charging stations. This represents the total number of charging guns at the charging station. This is the preset magnification factor.

[0084] By designing the comprehensive charging urgency index as a weighted sum of power risk, task urgency, and charging station congestion penalty factor, the power risk is obtained by mapping the current remaining power percentage to the predicted power required to reach the next charging station using a Sigmoid function. The task urgency is obtained by mapping the task deadline to the estimated travel time using an exponential function. The charging station congestion penalty factor is obtained by adjusting the ratio of the queue length of the candidate charging stations to the total number of charging guns using an amplification factor. This allows the final urgency assessment result to simultaneously reflect the combined impact of three dimensions: the AGV's own power risk, the urgency of the task being performed, and the real-time congestion status of the target charging station.

[0085] In some specific embodiments, , , The fitness function is determined offline through a genetic algorithm based on historical operational data. The optimization objective is to minimize the weighted sum of the weighted task interruption rate and charging wait time.

[0086] in, For task interruption rate, This refers to the average charging wait time. The preset penalty coefficient for task interruption rate. A preset penalty coefficient for average charging wait time.

[0087] By optimizing the weight coefficients of each item in the comprehensive charging urgency index offline using a genetic algorithm based on historical operating data, and taking the weighted sum of minimizing task interruption rate and average charging waiting time as the optimization objective, the weight ratio of each item can be automatically optimized based on actual operating data rather than relying on manual experience. Under different operating conditions, a weight combination that adapts to the current operating characteristics can be obtained.

[0088] In some specific embodiments, the selection of the optimal charging station is based on minimizing the objective function. The calculation formula is as follows:

[0089] in, For the first The AGV waiting to be charged arrived at the first The distance between charging stations This refers to the mechanical power of the AGV when it is running unloaded. Average driving speed This is the energy consumption conversion factor for empty driving. For charging piles The estimated waiting time Time cost weighting.

[0090] By minimizing the weighted sum of the empty-running energy consumption of the AGV to be charged and the expected queuing time during the optimal charging pile selection process, the AGV to be charged can simultaneously consider both the driving energy consumption and queuing time when selecting a charging pile, thus achieving a quantitative balance between optimizing the empty-running energy consumption of a single AGV and the queuing efficiency of the global charging station.

[0091] In some specific embodiments, The estimation is achieved by virtually scheduling the currently queued AGVs according to their respective comprehensive charging urgency indices.

[0092] By virtually scheduling the AGVs currently in the queue according to their comprehensive charging urgency index, the expected queuing time for charging piles can be estimated. This allows AGVs waiting to be charged to anticipate the queuing status of a charging pile in the future, rather than making decisions based solely on the current queue length. This avoids the bias of underestimating or overestimating the actual waiting time due to considering only the instantaneous queue length.

[0093] Step S4: For AGVs that are waiting to be charged but whose current tasks are not yet completed, select candidate AGVs from the surrounding preset radius. Under the constraint that the remaining power after taking on the task is higher than the safety value, select the candidate AGV with the smallest energy consumption increment to take on the unfinished task and issue a takeover instruction. Among them, the candidate AGV is the AGV that is currently in an idle state.

[0094] By selecting idle AGVs from a preset radius around an AGV that is currently not completing its task, and under the constraint that the remaining power of the candidate AGV is higher than a safe value after taking on the task, the candidate AGV with the smallest energy consumption increment is selected to take over the unfinished task of the AGV to be charged and a takeover instruction is issued. This allows the AGV to leave the site immediately to the charging station without waiting to complete its current task, while the task it was originally undertaking is seamlessly taken over by other idle AGVs, thus ensuring that the material handling task is not interrupted due to the AGV leaving the site for charging.

[0095] In some specific embodiments, the Hungarian algorithm is used to solve for the task allocation matrix and cost matrix elements. The calculation formula is:

[0096] in, For the first The remaining task distance of the AGV waiting to be charged. For the first The journey of the candidate AGVs to the handover point. For the first The average driving speed of the candidate AGVs For loading and unloading time.

[0097] By employing the Hungarian algorithm to solve the task allocation matrix, and defining the elements of the cost matrix as the sum of the remaining task distance of the original AGV and the distance of the candidate AGV to the handover point, divided by the average speed of the candidate AGV, plus the loading and unloading time, multiple factors such as the remaining task distance of the original AGV, the distance of the candidate AGV to the handover point, the average speed of the candidate AGV, and the loading and unloading time can be uniformly quantified and incorporated into the optimization consideration of task allocation. The goal is to seek the globally optimal matching scheme between all candidate AGVs and all tasks to be continued, rather than making decisions independently for each one.

[0098] In some specific embodiments, the constraint is that the remaining power of the candidate AGV after undertaking the task is higher than a safe value, and the formula is:

[0099] in, For the first The current remaining battery percentage of each candidate AGV. For the first The amount of electricity required for each candidate AGV to undertake the task. For the battery's rated capacity, This is a preset safe power threshold.

[0100] By setting a constraint that the remaining power of candidate AGVs after accepting a task must be higher than a preset safe power threshold, the scheduling system will not sacrifice the power safety of the candidate AGVs when allocating successive tasks in pursuit of task completion efficiency. This ensures that while the current task is successfully continued, the AGVs accepting the task will still have a basic power margin to independently perform other tasks or go to the charging station for a period of time in the future.

[0101] In some specific embodiments, among the feasible solutions that satisfy the constraints, the energy consumption increment is selected. The minimum allocation scheme, The calculation formula is:

[0102] in, For candidate number Increased energy consumption of candidate AGVs due to undertaking additional tasks.

[0103] By selecting the allocation scheme with the smallest energy consumption increment among feasible solutions that meet the power safety constraints, when multiple candidate AGVs can meet the condition that the remaining power after undertaking the task is higher than the safety value, the scheduling system will prioritize the AGV with the smallest increase in energy consumption due to undertaking the additional task as the final successor. This minimizes the overall additional energy consumption caused by task transfer while ensuring the successful succession of the task.

[0104] Step S5: Control each AGV to be charged to arrive at its corresponding charging pile. Then, dynamically allocate the output power of each charging gun according to the comprehensive charging urgency index of all AGVs currently charging and those in the queue. After charging is completed, each AGV to be charged updates its current working status to idle status.

[0105] By controlling each AGV to be charged to arrive at its corresponding charging pile, the output power of each charging gun is dynamically allocated according to the comprehensive charging urgency index of all AGVs currently charging and those in the queue. This allows the charging station capacity resources to be configured differently according to the urgency, with AGVs with higher urgency receiving higher charging rates. At the same time, after each AGV to be charged finishes charging, its current working status is updated to idle, so that AGVs that have completed charging can be reinstated into the task allocation and relay candidate evaluation scope in subsequent scheduling cycles, thereby achieving overall optimization of charging resource utilization efficiency.

[0106] In some specific embodiments, the dynamic allocation of the output power of each charging gun is regulated by a proportional-integral controller, and the control law formula is as follows:

[0107] in, This is the serial number of the charging gun that is currently charging. For the first A charging gun in Output power at any moment As the reference power, This is the proportionality coefficient. The integral coefficient is... For the first The overall charging urgency index of each charging gun for the AGV currently charging. The average comprehensive charging urgency index for all AGVs currently charging is τ, where τ is the integral variable.

[0108] By using a proportional-integral controller to adjust the output power of each charging gun during the charging phase, the control law incorporates the deviation between the comprehensive charging urgency index of the AGV corresponding to each charging gun and the average urgency of all currently charging AGVs as the adjustment basis. This allows the output power of each charging gun to be dynamically and continuously adjusted according to the real-time changes in the relative urgency ranking of each AGV. AGVs with urgency higher than the average level receive a charging rate higher than the benchmark power, while AGVs with urgency lower than the average level have their charging rates reduced accordingly.

[0109] In some specific embodiments, the allocation process simultaneously satisfies both the total power constraint and the minimum charging current constraint for each AGV. The formula for the total power constraint is:

[0110] in, This refers to the upper limit of the transformer capacity for charging stations; The formula for the minimum charging current constraint of AGV is:

[0111] in, Battery voltage, This is the minimum sustaining current required by the battery management system.

[0112] By simultaneously ensuring that the total power does not exceed the upper limit of the charging station transformer capacity and that the charging power of each AGV is not lower than the power value corresponding to the minimum sustaining current required by the battery management system during the power allocation process, dynamic power allocation will not cause the power supply system to overload due to the cumulative power demand of each charging gun exceeding the upper limit of the charging station transformer capacity, nor will it cause charging interruption or battery damage due to the power allocated to a certain AGV being too low and falling below the minimum sustaining current required by its battery management system. This ensures power supply safety and equipment safety while pursuing differentiated charging efficiency.

[0113] In some specific embodiments, step S5 further includes monitoring the rate of battery recovery for each AGV during the charging process. and battery temperature When the remaining battery percentage of the AGV reaches the preset safe battery threshold... When its overall charging urgency index ranking drops below that of newly entered AGVs waiting to be charged at the charging station, the output power of each charging gun is recalculated and adjusted according to the power distribution control law.

[0114] By monitoring the power recovery rate and battery temperature of each AGV during charging, when the remaining power percentage of an AGV reaches a preset safe power threshold or its overall charging urgency index ranking drops below that of newly entering AGVs waiting to be charged, the output power of each charging gun is recalculated and adjusted according to the proportional-integral controller control law. This allows AGVs that have completed their main charging needs to release some power to newly entering high-urgency AGVs in a timely manner, thus achieving dynamic reuse and efficient flow of power resources in the time dimension under the condition of limited charging station capacity.

[0115] The following are embodiments of the AGV cluster charging scheduling and task coordination system based on energy consumption prediction provided in this application. This AGV cluster charging scheduling and task coordination system based on energy consumption prediction belongs to the same inventive concept as the AGV cluster charging scheduling and task coordination methods in the above embodiments. For details not described in detail in the embodiments of the AGV cluster charging scheduling and task coordination system, please refer to the embodiments of the above-mentioned AGV cluster charging scheduling and task coordination methods based on energy consumption prediction.

[0116] like Figure 2 As shown, the AGV cluster charging scheduling and task coordination system based on energy consumption prediction includes: The data acquisition module is used to acquire AGV status data, task attribute data, and charging station status data for each AGV in the AGV cluster. The remaining power consumption curve generation module is used to predict the energy consumption of each AGV in the future period based on the AGV status data of each AGV and using a hybrid energy consumption prediction model. It outputs the predicted energy consumption value, calculates the remaining power consumption curve of each AGV based on the predicted energy consumption value, and marks the time when its power consumption is lower than the safety threshold. The optimal charging pile selection module is used to calculate the comprehensive charging urgency index of each AGV based on the remaining power change curve, task attribute data and charging station status data of each AGV, mark the AGV with the comprehensive charging urgency index exceeding the preset urgency index threshold as the AGV to be charged, and select the optimal charging pile for each AGV to be charged that minimizes the weighted sum of the empty driving energy consumption to the charging pile and the expected queuing waiting time. The task handover module is used to select candidate AGVs from a preset radius for AGVs that are waiting to be charged but whose current tasks have not been completed. Under the constraint that the remaining power after taking on the task is higher than the safety value, the module selects the candidate AGV with the smallest energy consumption increment to take on the unfinished task and issues a handover instruction. The charging management module controls each AGV to arrive at its corresponding charging pile. Then, it dynamically allocates the output power of each charging gun based on the comprehensive charging urgency index of all AGVs currently charging and those in the queue. After charging is completed, each AGV updates its current working status to idle status.

[0117] The AGV cluster charging scheduling and task coordination system in this embodiment is used to implement an AGV cluster charging scheduling and task coordination method based on energy consumption prediction.

[0118] This application also provides an electronic device for implementing the various embodiments of this application. Figure 3 To illustrate the hardware structure of an electronic device according to various embodiments of this application, as shown in the following diagram... Figure 3 As shown, the electronic device includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor.

[0119] Those skilled in the art will understand that the electronic device structure involved in the embodiments of this application does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0120] In embodiments of this application, electronic devices include, but are not limited to, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices may also represent various forms of mobile devices and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of this application described and / or claimed herein.

[0121] In this application embodiment, the processor can be implemented using at least one of an Application-Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a processor, a controller, a microcontroller, a microprocessor, or an electronic unit designed to perform the functions described herein. In some cases, such implementations can be implemented within a controller. For software implementations, implementations such as processes or functions can be implemented with separate software modules that allow the performance of at least one function or operation. The software code can be implemented by a software application (or program) written in any suitable programming language, and the software code can be stored in memory and executed by the controller.

[0122] In addition, the electronic device includes some functional modules not shown, which will not be described in detail here.

[0123] Those skilled in the art will understand that the various aspects of the electronic device provided in this application can be implemented as a system, method, or program product. Therefore, the various aspects of this application can be specifically implemented in the following forms: a completely hardware implementation, a completely software implementation (including firmware, microcode, etc.), or a combination of hardware and software aspects, collectively referred to herein as a "circuit," "module," or "system."

[0124] This application also provides a storage medium storing a program product capable of implementing an AGV cluster charging scheduling and task coordination method based on energy consumption prediction. In some possible implementations, various aspects of this application can also be implemented as a program product comprising program code that, when run on a terminal device, causes the terminal device to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of this application.

[0125] The storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example,, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0126] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for AGV cluster charging scheduling and task coordination based on energy consumption prediction, characterized in that, include: S1. Obtain the AGV status data, task attribute data, and charging station status data for each AGV in the AGV cluster; The AGV status data includes the current working status, remaining power percentage, motion parameters, load parameters, environmental parameters, and historical energy consumption records. The task attribute data includes the remaining task distance and task deadline. The charging station status data includes the queue length of each charging pile. S2. Based on the AGV status data of each AGV, use the hybrid energy consumption prediction model to predict the energy consumption of each AGV in the future period, output the predicted energy consumption value, calculate the remaining power change curve of each AGV based on the predicted energy consumption value, and mark the time when its power is lower than the safety threshold. The hybrid energy consumption prediction model includes a physical mechanism-based energy consumption analytical model and a pre-trained residual compensation model. The energy consumption analytical model outputs a theoretical energy consumption baseline, while the residual compensation model, built on a neural network, outputs a residual compensation value. The hybrid energy consumption prediction model adds the theoretical energy consumption baseline output by the energy consumption analytical model to the residual compensation value output by the residual compensation model to obtain the predicted energy consumption value. S3. Based on the remaining power change curve, task attribute data and charging station status data of each AGV, calculate the comprehensive charging urgency index of each AGV, mark the AGV with the comprehensive charging urgency index exceeding the preset urgency index threshold as the AGV to be charged, and select the optimal charging station for each AGV to be charged that minimizes the weighted sum of the empty driving energy consumption to the charging station and the expected queuing waiting time. S4. For AGVs that are waiting to be charged but whose current tasks are not yet completed, select candidate AGVs from the surrounding preset radius. Under the constraint that the remaining power after taking on the task is higher than the safety value, select the candidate AGV with the smallest energy consumption increment to take on the unfinished task and issue a takeover instruction. Among them, the candidate AGV is the AGV that is currently in an idle state; S5. Control each AGV to be charged to arrive at its corresponding charging pile, and then dynamically allocate the output power of each charging gun according to the comprehensive charging urgency index of all AGVs currently charging and those in the queue. After charging is completed, each AGV to be charged updates its current working status to idle status.

2. The AGV cluster charging scheduling and task coordination method as described in claim 1, characterized in that, The input to the energy consumption analysis model is the current travel speed from the AGV status data. , running path slope Load weight Ambient temperature Instantaneous value of motor current and remaining battery percentage The output is the theoretical energy consumption baseline per unit time step. ; The data processing flow of the energy consumption analysis model is as follows: S201. Calculate the mechanical power required for the AGV to overcome resistance per unit time based on the dynamic equation. The formula is: in, The total mass of the current AGV is calculated using the following formula: The current unloaded mass of the AGV. This represents the current load weight of the AGV. It is the acceleration due to gravity. The coefficient of rolling friction is air density, This is the drag coefficient. For windward area; S202. Calculate the theoretical energy consumption baseline of the current AGV. The calculation formula is: in, For a unit time step, Let I be the motor efficiency function with respect to the current driving speed v and the instantaneous value of the motor current I. For information regarding remaining battery percentage, State of Charge (SOC), and ambient temperature The battery discharge efficiency function; The residual compensation model is a long short-term memory neural network, whose input is the past... The residual sequence at each time step ,in The time window length, ; For the first The actual energy consumption per time step The output of the energy consumption analytical model at the same time step is the theoretical energy consumption baseline; the output of the residual compensation model is the future energy consumption baseline. Residual compensation value at each time step ,in To predict the number of time steps, For the first The residual compensation value for each prediction time step; The hybrid energy consumption prediction model adds the theoretical energy consumption baseline output by the energy consumption analytical model to the residual compensation value output by the residual compensation model to obtain the predicted energy consumption value. The calculation formula is as follows: in For the first Predicted energy consumption for each prediction time step. For the first The theoretical energy consumption baseline for each predicted time step. For the first The residual compensation value for each prediction time step.

3. The AGV cluster charging scheduling and task coordination method as described in claim 2, characterized in that, In step S2, the energy consumption of each AGV in the future time period is predicted using a pre-trained hybrid energy consumption prediction model to obtain the predicted energy consumption value. Then, the remaining power change curve is calculated based on the predicted energy consumption value, and the time when the power is lower than the safety threshold is marked. Specifically, this includes: The current travel speed in the AGV status data , running path slope Load weight Ambient temperature Instantaneous value of motor current and remaining battery percentage Input the hybrid energy consumption prediction model and output the predicted energy consumption value; The remaining electricity consumption change curve is calculated by accumulating the predicted energy consumption values. The calculation formula is as follows: in For the first The percentage of remaining battery power at each predicted time step. This represents the current percentage of battery remaining. For the first Predicted energy consumption for each prediction time step. This refers to the battery's rated capacity. Will satisfy The earliest time is marked as the moment when the battery level falls below the safe threshold, where This is a preset safe power threshold.

4. The AGV cluster charging scheduling and task coordination method as described in claim 1, characterized in that, In step S3, the overall charging urgency index The calculation formula is: in, The preset weighting coefficient for the risk level of electricity consumption. The preset task urgency weighting coefficient, The preset weighting coefficient for the charging station congestion penalty factor; Electricity risk level The calculation formula is: in, For the Sigmoid function, To predict the amount of electricity needed to reach the next charging station, This is the preset scaling factor; Task urgency The calculation formula is: in, The remaining time is calculated using the following formula: The deadline for the task. For the current moment, The formula for estimating travel time is as follows: The remaining distance for the mission. Average driving speed This is the preset sensitivity coefficient; Charging station congestion penalty factor The calculation formula is: This refers to the queue length of the available charging stations. This represents the total number of charging guns at the charging station. This is the preset magnification factor.

5. The AGV cluster charging scheduling and task coordination method as described in claim 1, characterized in that, The selection of the optimal charging station in step S3 is based on minimizing the objective function. The calculation formula is as follows: in, For the first The AGV waiting to be charged arrived at the first The distance between charging stations This refers to the mechanical power of the AGV when it is running unloaded. Average driving speed This is the energy consumption conversion factor for empty driving. For charging piles The estimated waiting time Time cost weighting.

6. The AGV cluster charging scheduling and task coordination method as described in claim 1, characterized in that, In step S4, the Hungarian algorithm is used to solve for the task allocation matrix and cost matrix elements. The calculation formula is: in, For the first The remaining task distance of the AGV waiting to be charged. For the first The journey of the candidate AGVs to the handover point. For the first The average driving speed of the candidate AGVs For loading and unloading time.

7. The AGV cluster charging scheduling and task coordination method as described in claim 6, characterized in that, The constraint in step S4 is that the remaining power of the candidate AGV after undertaking the task is higher than the safety value, and the formula is: in, For the first The current remaining battery percentage of each candidate AGV. For the first The amount of electricity required for each candidate AGV to undertake the task. For the battery's rated capacity, This is a preset safe power threshold.

8. The AGV cluster charging scheduling and task coordination method as described in claim 7, characterized in that, In step S4, among the feasible solutions that satisfy the constraints, the energy consumption increment is selected. The minimum allocation scheme, The calculation formula is: in, For candidate number Increased energy consumption of candidate AGVs due to undertaking additional tasks.

9. The AGV cluster charging scheduling and task coordination method as described in claim 1, characterized in that, In step S5, the output power of each charging gun is dynamically allocated using a proportional-integral controller, and the control law formula is as follows: in, This is the serial number of the charging gun that is currently charging. For the first A charging gun in Output power at any moment As the reference power, This is the proportionality coefficient. The integral coefficient is... For the first The overall charging urgency index of each charging gun for the AGV currently charging. The average comprehensive charging urgency index for all AGVs currently charging is τ, where τ is the integral variable.

10. The AGV cluster charging scheduling and task coordination method as described in claim 9, characterized in that, In step S5, the allocation process simultaneously satisfies both the total power constraint and the minimum charging current constraint for each AGV. The formula for the total power constraint is: in, This refers to the upper limit of the transformer capacity for charging stations; The formula for the minimum charging current constraint of AGV is: in, Battery voltage, This is the minimum sustaining current required by the battery management system.

Images