Task prediction based embodied intelligent mobile robot wireless charging scheduling and energy storage collaborative optimization method and system

Abstract

The application discloses a task prediction-based embodied intelligent mobile robot wireless charging scheduling and energy storage collaborative optimization method and system, comprising the following steps: acquiring robot cluster state data, task queue data, charging station and energy storage unit state data; predicting the task energy consumption demand of each robot in the future time window by using a deep learning energy consumption prediction model; dynamically adjusting the state of charge target value according to the task urgency, and determining the charging demand and charging urgency, charging demand quantity in combination with the predicted energy consumption and the current state of charge; jointly optimizing the charging time and charging quantity of each robot under various system constraints, with the goal of minimizing the total downtime, charging cost and battery health loss of the cluster; when the charging period conflicts, performing competitive scheduling according to the charging urgency; at the same time, performing charging and discharging collaborative scheduling on the charging station energy storage buffer unit according to the time-of-use electricity price; finally, issuing and executing the scheduling instructions.

Description

Technical Field

[0001] This invention belongs to the field of robot charging scheduling and energy storage technology, and particularly relates to a method and system for wireless charging scheduling and energy storage collaborative optimization of embodied intelligent mobile robots based on task prediction. Background Technology

[0002] Embodying intelligent mobile robots, powered by energy storage batteries, have been widely used in warehousing and logistics, intelligent manufacturing, and hospital delivery, requiring regular charging. Current mainstream charging solutions suffer from the following shortcomings: charging triggering is based on a fixed low-power threshold, lacking task predictability and easily leading to task delays or wasted charging resources; charging decisions fail to achieve comprehensive optimization of "when to charge, how much to charge, and where to charge," and ignore the competitive scheduling of charging station resources in scenarios with multiple embodying intelligent mobile robots; the coordination between the charging station's energy storage buffer unit and the charging needs of the embodying intelligent mobile robots is insufficient, potentially causing instantaneous grid load peaks and failing to utilize peak-valley electricity pricing to reduce operating costs; multi-robot power allocation is only "on-demand response," failing to incorporate future tasks into global optimization; theoretical energy consumption calculations do not explicitly incorporate safety power constraints for handling sudden operating conditions, resulting in insufficient system robustness.

[0003] Some existing technologies have attempted to improve charging decisions through power prediction. For example, Chinese patent CN119116759A discloses a dynamic power warning system and method for mobile charging robots, which predicts power consumption and allocates it according to orders using deep learning. However, this only achieves warning judgment and does not solve the problem of comprehensive charging optimization. It allocates power through priority coefficients, but this is still an on-demand response scheduling method, lacking a global consideration of future tasks. None of the above solutions achieve coordinated scheduling of the intelligent mobile robot's future task energy consumption, charging station resource status, and energy storage buffer units, lacking a systematic solution. Summary of the Invention

[0004] To address the problems existing in the prior art, this invention provides a method and system for wireless charging scheduling and energy storage collaborative optimization of embodied intelligent mobile robots based on task prediction. This transforms passive charging into active predictive charging, realizes global optimization of charging resources for multiple robots and collaborative energy storage, reduces operating costs, extends battery life, and ensures safe system operation.

[0005] The technical solution of this invention is implemented as follows: A method for coordinated optimization of wireless charging scheduling and energy storage for embodied intelligent mobile robots based on task prediction includes the following steps: S1. Obtain the status data of each robot in the embodied intelligent mobile robot cluster, the task queue data, the wireless charging station status data, and the energy storage buffer unit status data. S2. Use the energy consumption prediction model to predict the task energy consumption requirements of each robot in the future time window. S3. Dynamically adjust the target value of the state of charge according to the urgency of the task, and determine the robots with charging needs and their charging urgency and charging demand indicators by combining the task energy consumption requirements and the current state of charge. S4. With the goal of minimizing the total downtime, charging cost and battery health deterioration of the embodied intelligent mobile robot cluster, the charging timing and charging amount of each robot are jointly optimized under the condition of satisfying the system operation constraints. S5. When the charging time of multiple robots conflicts, the charging station resources will be competitively and collaboratively scheduled based on the charging urgency index. S6. Based on time-of-use electricity pricing and robot charging load forecasts, coordinate the charging and discharging of the energy storage buffer unit at the charging station. S7. Dynamically adjust the target value of the state of charge of each robot according to the urgency of the task, the joint optimization results and the competitive scheduling results, and feed back the adjusted target value of the state of charge to S3 to form a dynamic adjustment cycle. S8. Send the scheduling instructions to each robot and the energy storage controller of the charging station.

[0006] Furthermore, in S2, the energy consumption prediction model adopts a long short-term memory network or a Transformer architecture. The input features include at least the task type code, task duration, task path length, task load weight, average power consumption of historical similar tasks, robot current charge state and running state; the output is the predicted power consumption value for each time step within the future time window, and the predicted power consumption value is integrated to obtain the predicted value of total energy consumption demand. Furthermore, in S3, determining that a robot has a charging requirement includes: setting a safe charge state limit; when the robot's current remaining available power is lower than the safe reserved power determined by the safe charge state limit, or when the value of the current remaining available power minus the predicted task energy consumption requirement is lower than the safe reserved power, it is determined that the robot has a charging requirement. Furthermore, in S3, the charging urgency index The calculation formula is: ; in, The current state of charge of the i-th embodied intelligent mobile robot. To the limit of safe state of charge, It represents the urgency of the current battery level approaching the safe threshold. The closer When the urgency level increases, the charging urgency level rises; , , Preset weighting coefficients; For the task time limit, As a task priority weight; Charging demand indicators The calculation formula is: ; in, The target value for the state of charge is set to 90% when performing emergency tasks and 80% when performing routine tasks, and is subsequently dynamically determined by S7. The rated battery capacity of the i-th embodied intelligent mobile robot; The net energy value that the i-th embodied intelligent mobile robot needs to charge its battery is always a non-negative value; After charging, the state of charge should be restored to at least 10% of the safe state of charge limit. Furthermore, in step S4, a multi-objective optimization model is established with the objectives of minimizing the total downtime, charging cost, and battery health deterioration of the embodied intelligent mobile robot cluster, to obtain the optimal charging time and optimal charging amount for each embodied intelligent mobile robot; the objective function of the multi-objective optimization model is: ; in, Let i be the downtime of the i-th embodied intelligent mobile robot due to charging. For charging costs, The charging process degrades battery health. These are preset weighting coefficients. Furthermore, in S4, the system operation constraints include: charging station capacity constraints, charging power constraints, state of charge safety constraints, and task completion constraints; The capacity constraint for charging stations is: the number of charging stations serving at the same time cannot exceed the number of available charging stations. The charging power constraint is: the charging power of a single robot shall not exceed the rated maximum power; The state of charge safety constraints are as follows: during robot operation, the real-time state of charge shall not be lower than the safety lower limit, and after charging, the state of charge shall not exceed the upper limit and shall be restored to at least 10% of the level exceeding the safety lower limit; The task completion constraint is: the robot must complete charging and arrive at the task location before the task deadline. Furthermore, in S5, the competitive scheduling of charging station resources includes: allocating charging time slots in descending order of charging urgency index; For robots that do not receive priority charging time, alternative charging plans are generated based on their charging urgency index and task time limit. Alternative charging options include a standby charging option or a partial charging option; The waiting charging plan involves postponing the charging start time to the earliest available idle time at the charging station, recalculating the state of charge after charging is completed and whether it can meet the task time limit. If it can, it will be used as an alternative charging plan. Some charging solutions involve reducing the charging amount to the minimum value that meets the minimum charging guarantee constraint if the waiting time is unacceptable. If none of the alternative charging options can meet the task time limit, a system alarm will be triggered. If the alternative solution adjusts the optimal charging amount or optimal charging timing, the adjusted parameters will be fed back to S4 for re-evaluation and optimization. Furthermore, in step S6, the coordinated scheduling of the energy storage buffer unit specifically includes: Obtain time-of-use electricity pricing information from the power grid; During periods of low electricity prices, the grid charges the energy storage buffer unit; during periods of high electricity prices or peak periods of robot charging load, the energy storage buffer unit prioritizes power supply. When the state of charge of the energy storage buffer unit is lower than the preset lower limit, the power supply is switched to the grid directly. When the mains power is interrupted, the energy storage buffer unit switches to emergency power supply mode. Furthermore, in S8, the scheduling instructions generated in S4-S7 are sent to each robot and the energy storage controller of the charging station through a wireless communication network. Each robot goes to the designated charging station for charging within a specified time according to the scheduling instructions. The status data during the execution process is collected in real time and fed back to the scheduling center for iterative optimization. The execution feedback data includes: real-time collection of the actual charging start time, actual charging duration, actual power consumption curve and task completion status of each robot; comparison of the collected data with the predicted data and calculation of the prediction error; when the accumulated error exceeds the preset threshold, the incremental retraining of the energy consumption prediction model is triggered. The system for wireless charging scheduling and energy storage collaborative optimization of embodied intelligent mobile robots based on task prediction is characterized by applying the wireless charging scheduling and energy storage collaborative optimization method for embodied intelligent mobile robots based on task prediction as described above. Compared with the prior art, the present invention achieves the following beneficial effects: 1. Shift from passive charging to proactive predictive charging. By using deep learning models to predict the future energy consumption needs of embodied intelligent mobile robots, charging strategies can be planned in advance before the battery level reaches a low threshold, avoiding forced interruptions to charging during task execution, and improving task completion rate and system operating efficiency.

[0007] 2. Multi-objective joint optimization for global optimum. By incorporating downtime, charging cost, and battery health into a unified optimization framework, it overcomes the limitations of existing technologies that only consider a single objective, achieving system-level comprehensive optimum.

[0008] 3. Resource competition and coordination at charging stations to avoid congestion. In scenarios where multiple intelligent mobile robots request charging simultaneously, priority ranking and resource allocation are based on the urgency of charging, effectively avoiding congestion and waiting caused by multiple robots simultaneously competing for limited charging resources.

[0009] 4. Coordinated optimization of energy storage side to reduce costs and increase efficiency. Integrate the energy storage buffer units at the charging station into unified scheduling, utilize peak-valley electricity price differences to reduce operating costs, and smooth grid load fluctuations, thereby reducing the charging station's grid access capacity requirements.

[0010] 5. Dynamic SOC management extends battery life. The target SOC value is dynamically adjusted based on the urgency of the task, avoiding unnecessary deep charging and prolonged high SOC float charging, effectively extending battery cycle life.

[0011] 6. Engineered safety margin management. Safety power constraints are integrated into the charging demand judgment and optimization model to ensure that the embodied intelligent mobile robot maintains a minimum safe power reserve under any working condition, avoiding downtime due to energy consumption fluctuations. Attached Figure Description

[0012] Figure 1 This is a flowchart of the wireless charging scheduling and energy storage collaborative optimization method for embodied intelligent mobile robots based on task prediction provided in Embodiment 1 of the present invention; Figure 2 This is a state data diagram of three embodied intelligent mobile robots in the task prediction-based wireless charging scheduling and energy storage collaborative optimization method for embodied intelligent mobile robots provided in Embodiment 1 of the present invention. Figure 3 This is a flowchart of the charging station resource competition and cooperation scheduling process of the wireless charging scheduling and energy storage collaborative optimization method for embodied intelligent mobile robots based on task prediction provided in Embodiment 1 of the present invention. Figure 4 This is a flowchart of the collaborative scheduling of energy storage buffer units in the wireless charging scheduling and energy storage collaborative optimization method for embodied intelligent mobile robots based on task prediction provided in Embodiment 1 of the present invention. Figure 5 This is a system module architecture diagram of the wireless charging scheduling and energy storage collaborative optimization method for embodied intelligent mobile robots based on task prediction provided in Embodiment 1 of the present invention; Figure 6 This is a diagram of the embodied intelligent mobile robot provided in Embodiment 1 of the present invention, which is a wireless charging scheduling and energy storage collaborative optimization method for embodied intelligent mobile robots based on task prediction. Figure 7This is a schematic diagram of the energy consumption prediction model of the wireless charging scheduling and energy storage collaborative optimization method for embodied intelligent mobile robots based on task prediction provided in Embodiment 2 of the present invention. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions in the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0014] Example 1 like Figures 1 to 7 A method for coordinated optimization of wireless charging scheduling and energy storage for embodied intelligent mobile robots based on task prediction includes the following steps: S1. Obtain the status data of each robot in the embodied intelligent mobile robot cluster, the task queue data, the wireless charging station status data, and the energy storage buffer unit status data. The status data is real-time status data, which includes at least: current location coordinates, current SOC value, current operating speed, and current power consumption; Task queue data includes historical task data and current task queue data; Historical task data should include at least: historical task type, task duration, and power consumption curve during task execution; The current task queue data includes at least: a list of tasks to be executed, the task type of each task, the task location, the task time limit, and the task priority; The status data of a wireless charging station should include at least: the location of the charging station, the working status of the transmitter, the current occupancy status, and the queue. The status data of the energy storage buffer unit should include at least: the current SOC value of the energy storage unit, the available charging and discharging power, and the grid connection status.

[0015] In this embodiment, in a smart warehouse scenario, the scenario includes 50 embodied smart mobile robots, 5 wireless charging stations (each charging station is equipped with 2 charging positions) and 1 centralized energy storage buffer unit (capacity of 500kWh).

[0016] Step S1 mainly involves data acquisition and preprocessing: Each embodied intelligent mobile robot collects status data in real time through onboard sensors and control systems, including: current position coordinates (obtained through laser SLAM or UWB positioning), current SOC value (i.e., state of charge, read through BMS), current running speed, and current drive motor power consumption.

[0017] Historical task data is obtained from the warehouse management system (WMS), including the task type (such as picking, handling, and replenishment) for each task in the past 30 days, the task duration, and the power consumption curve during task execution.

[0018] The current task queue data is obtained from the scheduling system in real time.

[0019] The status data of the wireless charging station is obtained through the reporting of the charging station controller.

[0020] The status data of the energy storage buffer unit is read through the energy storage management system (EMS).

[0021] All data is uploaded to the cloud scheduling platform in real time via the 5G network, and undergoes data cleaning and normalization preprocessing.

[0022] S2. Use the energy consumption prediction model to predict the task energy consumption requirements of each robot in the future time window. The energy consumption prediction model adopts a Long Short-Term Memory (LSTM) or Transformer architecture. The input features include at least the task type code, task duration, task path length, task load weight, average power consumption of historical similar tasks, robot current charge state and running state. The output is the predicted power consumption value at each time step within the future time window, and the predicted power consumption value is integrated to obtain the predicted total energy consumption demand value.

[0023] In this embodiment, an energy consumption prediction model based on the Transformer architecture is constructed.

[0024] The model's input features include: Task type encoding (one-hot encoding, 10 task types in total); Estimated duration of the task (provided by WMS); Task path length (provided by the path planning module); Task load weight (provided by WMS); Average power consumption (statistical value) for similar historical tasks. Current SOC value of the embodied intelligent mobile robot; Current operating status of the embodied intelligent mobile robot (running / standby / charging).

[0025] The model uses historical data from the past 7 days as a training set to predict the power consumption curve for the next hour.

[0026] For the i-th embodied intelligent mobile robot, the model outputs the predicted power consumption value for each time step (1-minute interval) within the next hour. Then the predicted total energy demand is: .

[0027] S3. Dynamically adjust the target value of the state of charge according to the urgency of the task, and determine the robots with charging needs and their charging urgency and charging demand indicators by combining the task energy consumption requirements and the current state of charge. The criteria for determining whether a robot has a charging requirement include: setting a safe charge state limit; when the robot's current remaining available power is lower than the safe reserved power determined by the safe charge state limit, or when the value of the current remaining available power minus the predicted task energy consumption requirement is lower than the safe reserved power, it is determined that the robot has a charging requirement.

[0028] Specifically, setting a safety charge state limit for the operation of embodied intelligent mobile robots. The safe state of charge limit refers to the minimum electrical energy reserve that an intelligent mobile robot must maintain during task execution to cope with uncertainties such as path changes, load fluctuations, and temporary detours. The value is set to 25%.

[0029] The corresponding safety reserve power is ,in, The rated battery capacity of the i-th embodied intelligent mobile robot.

[0030] Charging demand determination criteria: The i-th embodied intelligent mobile robot is determined to have a charging demand when it meets any of the following conditions: Condition A (Safety threshold triggered): Current remaining available electrical energy Below the safety reserve power , This condition ensures that the embodied intelligent mobile robot never falls below the safe power limit under any circumstances.

[0031] Condition B (Task Energy Consumption Trigger): The current remaining available electrical energy is insufficient to support the predicted future energy consumption demand and cannot reserve a safety margin, i.e. This condition ensures that the embodied intelligent mobile robot retains a safe reserve of power after completing its task.

[0032] Charging urgency index The calculation formula is: ; in, This represents the current state of charge of the i-th embodied intelligent mobile robot (expressed as a decimal, e.g., 0.35 represents 35%). The safe state of charge limit is set at 0.25. It represents the urgency of the current battery level approaching the safe threshold. The closer When the urgency level increases, the charging urgency level rises; , , Preset weighting coefficients; Task time limit (unit: minutes). As a task priority weight; Charging demand indicators The calculation formula is: ; in, The target value for the state of charge is set to 90% when performing emergency tasks and 80% when performing routine tasks, and is subsequently dynamically determined by S7. The rated battery capacity of the i-th embodied intelligent mobile robot; The net electrical energy (in kWh) required to charge the battery of the i-th embodied intelligent mobile robot is always a non-negative value. After charging, the state of charge should be restored to at least 10% of the safe state of charge limit.

[0033] It should be noted that the target value of state of charge used in this step is... The SOC target value is dynamically determined by step S7. In actual system operation, the SOC target value adjustment in step S7 can also be used as a pre-emptive strategy, executed before the charging demand assessment in step S3, to provide for S3. The basis for its value.

[0034] At the same time, the charging demand must also meet the minimum charging guarantee constraint: after charging, the state of charge of the embodied intelligent mobile robot should be restored to at least [value missing]. (i.e., not less than 35%), to ensure sufficient safety margin for operation after charging. This constraint is expressed as: ; If calculated If the constraint is not met, the charging amount will be automatically increased to the minimum value that satisfies the constraint.

[0035] In this embodiment, a safe charging state limit is set. Corresponding safety reserve power .

[0036] Taking a certain moment as an example, the scheduling platform receives status data from three embodied intelligent mobile robots, such as... Figure 2 ; Charging demand assessment: Embossed Intelligent Mobile Robot-01:E remain (3.5kWh) > E safe (2.5kWh), condition A is not met; E remain -Eneed =3.5-4.2=-0.7kWh <E safe (2.5kWh), condition B is met → there is a charging demand.

[0037] Embossed Intelligent Mobile Robot-02:E remain (2.8kWh) > E safe (2.5kWh), condition A is not met; E remain -E need =2.8-3.8=-1.0kWh <E safe (2.5kWh), condition B is met → there is a charging demand.

[0038] Embossed Intelligent Mobile Robot-03:E remain (5.5kWh) > E safe (2.5kWh), condition A is not met; E remain -E need =5.5 - 5.0 = 0.5kWh <E safe (2.5kWh), condition B is met → there is a charging demand (25% safety margin cannot be retained after the task is completed).

[0039] Charging urgency index U i Calculate (taking α=0.05, β=0.3, γ=0.3, SOC) min =0.25): Embossed Intelligent Mobile Robot-01: U = 0.05 × [1 / (0.35 - 0.25)] + 0.3 × (1 / 30) + 0.3 × 5 = 0.05 × 10 + 0.01 + 1.5 = 0.5 + 0.01 + 1.5 = 2.01 Embossed Intelligent Mobile Robot-02: U = 0.05 × [1 / (0.28-0.25)] + 0.3 × (1 / 60) + 0.3 × 3 = 0.05 × 33.33 + 0.005 + 0.9 = 1.667 + 0.005 + 0.9 = 2.572 Embossed Intelligent Mobile Robot-03: U = 0.05 × [1 / (0.55-0.25)] + 0.3 × (1 / 20) + 0.3 × 10 = 0.05 × 3.33 + 0.015 + 3 = 0.167 + 0.015 + 3 = 3.182 Urgency ranking: Intelligent Mobile Robot-03 (3.182) > Intelligent Mobile Robot-02 (2.572) > Intelligent Mobile Robot-01 (2.01). Intelligent Mobile Robot-03 is the most urgent because the task is urgent and there is no safety margin after completion. Intelligent Mobile Robot-02 is the second most urgent because its current battery level is approaching the 25% safety line.

[0040] Charging demand Q i Computation (SOC) target As dynamically determined in step S7, in this example, the embodied intelligent mobile robot-01 has an emergency task target of 90%, the embodied intelligent mobile robot-02 has a regular task target of 80%, and the embodied intelligent mobile robot-03 has an emergency task target of 90%. Embossed Intelligent Mobile Robot-01: Q = max(0,(0.90-0.35)×10) =5.5kWh Embossed Intelligent Mobile Robot-02: Q = max(0,(0.80-0.28)×10) =5.2kWh Embossed Intelligent Mobile Robot-03: Q = max(0,(0.90-0.55)×10) =3.5kWh Minimum charging guarantee constraint verification (post-charge not lower than SOC_min + 10% = 35%): The SOC of the embodied intelligent mobile robot-01 after charging is 35% + 5.5 / 10 = 90% ≥ 35%, which meets the requirement.

[0041] The SOC of the embodied intelligent mobile robot-02 after charging is 28% + 5.2 / 10 = 80% ≥ 35%, which meets the requirements.

[0042] The SOC of the embodied intelligent mobile robot-03 after charging is 55% + 3.5 / 10 = 90% ≥ 35%, which meets the requirements.

[0043] S4. With the goal of minimizing the total downtime, charging cost and battery health deterioration of the embodied intelligent mobile robot cluster, the charging timing and charging amount of each robot are jointly optimized under the condition of satisfying the system operation constraints. To minimize the total downtime, charging cost, and battery health deterioration of the embodied intelligent mobile robot cluster, a multi-objective optimization model is established to obtain the optimal charging time and optimal charging amount for each embodied intelligent mobile robot. The objective function of the multi-objective optimization model is: ; in, Let i be the downtime of the i-th embodied intelligent mobile robot due to charging. For charging costs (considering time-of-use electricity pricing). The charging process degrades battery health. These are preset weighting coefficients. System operation constraints include: charging station capacity constraints, charging power constraints, state of charge safety constraints, and task completion constraints; The capacity constraint for charging stations is: the number of charging stations serving at the same time cannot exceed the number of available charging stations. The charging power constraint is: the charging power of a single robot shall not exceed the rated maximum power; The state of charge safety constraint is: the real-time state of charge of the robot during operation shall not fall below the safety lower limit. (Take 25%), the state of charge after charging should not exceed the upper limit. (Take 95%), and recover to at least 10% above the safety lower limit, that is (i.e., no less than 35%); ensure sufficient safety margin before the next mission is executed; The task completion constraint is: the robot must complete charging and arrive at the task location before the task deadline.

[0044] In this embodiment, each charging station uses 22kW wireless charging power with a charging efficiency of 93%.

[0045] Constraints: Five charging stations with a total of 10 charging spots; The maximum charging power of a single embodied intelligent mobile robot is 22kW; SOC safety constraints: The SOC during operation must not be lower than 25%, and the SOC after charging must not exceed 95%, and must be restored to at least 35% after charging; The embodied intelligent mobile robot-03 has a 20-minute time limit and must complete charging and arrive at the mission location before the time limit.

[0046] The solution is obtained using a genetic algorithm, with the population size, number of iterations, crossover probability, and mutation probability dynamically configured based on the number of embodied intelligent mobile robots and the number of charging stations.

[0047] Optimization results: Embossed Intelligent Mobile Robot-03 (Highest Urgency): Recharges immediately with a capacity of 3.5kWh (from 55% to 90%), charging time is approximately 10 minutes, and performs emergency tasks immediately after charging; Embossed Intelligent Mobile Robot-02 (Second in urgency): Charges immediately, with a charging capacity of 5.2kWh (from 28% to 80%), and a charging time of approximately 15 minutes. It is located at a different charging position than Embossed Intelligent Mobile Robot-03. Embossed Intelligent Mobile Robot-01 (Urgency Level Again): Delay charging for 20 minutes (after Embossed Intelligent Mobile Robot-03 is completed), charging capacity 5.5kWh (from 35% to 90%), charging time approximately 16 minutes.

[0048] S5. When the charging time of multiple robots conflicts, the charging station resources will be competitively and collaboratively scheduled based on the charging urgency index. The competitive and cooperative scheduling of charging station resources includes: Charging time slots are allocated in descending order of charging urgency. For robots that do not receive priority charging time, alternative charging plans are generated based on their charging urgency index and task time limit. Alternative charging options include a standby charging option or a partial charging option; The waiting charging plan involves postponing the charging start time to the earliest available idle time at the charging station, recalculating the state of charge after charging is completed and whether it can meet the task time limit. If it can, it will be used as an alternative charging plan. Some charging solutions stipulate that if the waiting time is unacceptable, the charging amount will be reduced to the minimum value that meets the minimum charging guarantee constraint; that is, the SOC after charging will not be lower than the minimum SOC. min The minimum value of +10% is used to shorten the charging time and make the charging position available earlier.

[0049] If none of the alternative charging options can meet the task time limit, a system alarm will be triggered. If the alternative solution adjusts the optimal charging amount or optimal charging timing, the adjusted parameters will be fed back to S4 for re-evaluation and optimization.

[0050] Specifically, S51, the charging urgency index calculated based on step S3. Prioritize all embodied intelligent mobile robots that request charging. S52. Allocate charging station resources in order of priority from high to low, with the high-priority embodied intelligent mobile robot receiving the charging time first. S53. For embodied intelligent mobile robots that do not receive priority charging time, generate alternative charging schemes based on their charging urgency index and task time limit. If neither of the two alternative solutions can meet the task deadline, a system-level alarm will be triggered, requiring manual intervention from the dispatcher. S54. Dynamically update the charging station queue and send the scheduling results to each intelligent mobile robot. S55. If the alternative scheme generated in S53 adjusts the optimal charging amount or optimal charging timing determined in S4, the adjusted parameters are fed back to the objective function of step S4 for re-evaluation. If the adjusted scheme does not meet the task completion constraints or the battery health deteriorates beyond the preset tolerance, such as if the objective function deteriorates by more than 5%, S4 is re-optimized, and the alternative time window in S53 is used as a new constraint for re-solving. Otherwise, the adjusted scheme is directly adopted. If the new scheme is adopted after re-optimization, the issuance step of S54 is repeated to ensure that the scheduling command is received by each execution end.

[0051] In this embodiment, since both the embodied intelligent mobile robot-02 and the embodied intelligent mobile robot-03 request immediate charging, a competitive scheduling process is executed: According to the charging urgency ranking calculated in step S3: Intelligent Mobile Robot-03 (3.182) > Intelligent Mobile Robot-02 (2.572) > Intelligent Mobile Robot-01 (2.01), Intelligent Mobile Robot-03 will be given priority to charge at charging station 1, position 1. Intelligent Mobile Robot-02 will be given priority to charge at charging station 1, position 2. Intelligent Mobile Robot-01 will start charging at charging station 1, position 1 after Intelligent Mobile Robot-03 has finished charging (10 minutes later).

[0052] S6. Based on time-of-use electricity pricing and robot charging load forecasts, coordinate the charging and discharging of the energy storage buffer unit at the charging station. In step S6, the coordinated scheduling of the energy storage buffer unit specifically includes: S61. Obtain time-of-use electricity price information from the power grid and predict the trend of electricity price changes within a preset time window in the future. S62. Based on the charging load curve of the embodied intelligent mobile robot predicted by S4, formulate a charging and discharging plan for the energy storage buffer unit: during periods of low electricity price, the grid charges the energy storage buffer unit, and during periods of high electricity price or high robot charging load, the energy storage buffer unit prioritizes power supply. S63. Monitor the SOC status of the energy storage buffer unit in real time. When the SOC status of the energy storage buffer unit is lower than the preset lower limit, switch to direct power supply from the grid. S64. When the mains power is interrupted, the energy storage buffer unit switches to emergency power supply mode to ensure the charging needs of the critical embodied intelligent mobile robot.

[0053] In this embodiment, the time-of-use electricity price information of the power grid is as follows: The electricity price during peak hours (10:00-15:00, 18:00-21:00) is 1.2 yuan / kWh; The electricity price during normal hours (07:00-10:00, 15:00-18:00, 21:00-23:00) is 0.8 yuan / kWh; The electricity price during off-peak hours (23:00-07:00) is 0.4 yuan / kWh.

[0054] The current time is 9:30 AM (normal time). The predicted charging load for the embodied intelligent mobile robot in the next hour is approximately 50kW. The energy storage buffer unit's current SOC is 80%.

[0055] Energy storage coordinated control strategy: Since the current period is normal and the energy storage SOC is sufficient, the energy storage buffer unit will prioritize charging and powering the embodied intelligent mobile robot, while reserving 20% ​​capacity as an emergency backup. Compared to direct grid power supply, this period can save approximately (0.8-0.4)×50 = 20 yuan / hour in electricity costs.

[0056] S7. Dynamically adjust the target value of the state of charge of each robot according to the urgency of the task, the joint optimization results and the competitive scheduling results, and feed back the adjusted target value of the state of charge to S3 to form a dynamic adjustment cycle. Based on the optimization results of step S4 and the scheduling results of step S5, the SOC target value of each embodied intelligent mobile robot is dynamically adjusted: When the embodied intelligent mobile robot is about to perform an emergency or high-energy-consuming task, increase the SOC target value, such as from 80% to 90%, to ensure sufficient power reserves. When the embodied intelligent mobile robot is about to perform a low-intensity task or is in standby mode, reduce the target SOC value, such as from 80% to 70%, to reduce unnecessary deep charging and extend battery life.

[0057] In this embodiment, the embodied intelligent mobile robot-03 is about to perform an emergency task. The task has an emergency priority, and its SOC target value is dynamically adjusted from the usual 80% to 90% to ensure sufficient power reserves.

[0058] The embodied intelligent mobile robot-01 will also perform high-priority tasks, with a target value set at 90%. The embodied intelligent mobile robot-02 will perform routine tasks, maintaining a target value of 80%.

[0059] S8. Send the scheduling instructions to each robot and the energy storage controller of the charging station.

[0060] The scheduling instructions generated by S4-S7 are sent to each robot and the energy storage controller of the charging station through the wireless communication network. Each robot goes to the designated charging station for wireless charging within the specified time according to the scheduling instructions. The status data during the execution process is collected in real time and fed back to the scheduling center for model iteration and optimization. The execution feedback data includes: real-time collection of the actual charging start time, actual charging duration, actual power consumption curve and task completion status of each robot. The collected data is compared with the predicted data and the prediction error (MAPE) is calculated. When the cumulative error exceeds the preset threshold (e.g., MAPE>15%), the incremental retraining of the energy consumption prediction model is triggered - the model is fine-tuned for 3 epochs using the data from the most recent 7 days.

[0061] In this embodiment, scheduling instructions are sent via the 5G network to the embodied intelligent mobile robot-01, embodied intelligent mobile robot-02, embodied intelligent mobile robot-03, and the energy storage controller of charging station No. 1. Each execution terminal executes the instructions, and the status data during the execution process is transmitted back to the cloud scheduling platform in real time for online learning and continuous optimization of the model.

[0062] The system for wireless charging scheduling and energy storage collaborative optimization of embodied intelligent mobile robots based on task prediction is characterized by the application of the task prediction-based wireless charging scheduling and energy storage collaborative optimization method for embodied intelligent mobile robots as described above.

[0063] This invention provides a method and system for wireless charging scheduling and energy storage coordination optimization of embodied intelligent mobile robots based on task prediction. It realizes the optimization of the charging mechanism from passively triggered charging to active predictive charging. Based on the prediction results of the energy consumption demand of future tasks of embodied intelligent mobile robots, the optimal charging time and charging amount are planned in advance. In multi-embodied intelligent mobile robot application scenarios, realize intelligent competitive and cooperative scheduling of charging station resources and minimize the total downtime of the cluster; By incorporating the energy storage buffer units at charging stations into a unified scheduling framework, peak shaving and valley filling of the power grid load and optimization of electricity costs can be achieved. The target value of the state of charge (SOC) of the energy storage battery of the embodied intelligent mobile robot is dynamically adjusted to balance the urgency of the task and the health of the battery. The engineering constraint of reserving sufficient safe power is explicitly incorporated into the charging demand judgment and optimization model to ensure the robustness and safety of system operation.

[0064] Example 2 The difference between this embodiment and Embodiment 1 is that the energy consumption prediction model uses a Long Short-Term Memory (LSTM) network instead of the Transformer architecture, making it suitable for edge deployment scenarios with limited computing resources.

[0065] The LSTM model takes a 60-time-step input sequence (corresponding to 60 minutes) as input, with each time-step having a 12-dimensional feature dimension. The model contains two LSTM hidden layers, each with 128 neurons, and a fully connected output layer that outputs the predicted power consumption values ​​for the next 60 time steps.

[0066] The model was trained on a historical dataset using the Adam optimizer with a learning rate of 0.001, a batch size of 32, and 200 training epochs. After training, the model was deployed on an edge gateway to achieve low-latency real-time predictions.

[0067] For other parts not described in Example 2, please refer to Example 1, and will not be repeated here.

[0068] Example 3 The difference between this embodiment and Embodiment 1 is that the multi-objective optimization model is solved using the mixed integer linear programming (MILP) method, which is suitable for scenarios with a small number of embodied intelligent mobile robots (e.g., less than 20), and can guarantee the global optimal solution.

[0069] The time axis is discretized into time slots with 5-minute intervals, and the decision variables are binary variables. (This indicates whether the i-th embodied intelligent mobile robot is charging at the k-th charging station in the t-th time slot). The objective function and constraints are both linearized and solved using commercial solvers (such as Gurobi or CPLEX).

[0070] For other parts not described in Example 3, please refer to Example 1, and will not be repeated here.

[0071] Based on the disclosure and teachings of the foregoing specification, those skilled in the art can make changes and modifications to the above embodiments. Therefore, the present invention is not limited to the specific embodiments disclosed and described above, and some modifications and changes to the present invention should also fall within the protection scope of the claims of the present invention. Furthermore, although some specific terms are used in this specification, these terms are only for convenience of explanation and do not constitute any limitation on the present invention.

Claims

1. A method for wireless charging scheduling and energy storage collaborative optimization of embodied intelligent mobile robots based on task prediction, characterized in that, Includes the following steps: S1. Obtain the status data of each robot in the embodied intelligent mobile robot cluster, the task queue data, the wireless charging station status data, and the energy storage buffer unit status data. S2. Use the energy consumption prediction model to predict the task energy consumption requirements of each robot in the future time window. S3. Dynamically adjust the target value of the state of charge according to the urgency of the task, and determine the robots with charging needs and their charging urgency and charging demand indicators by combining the task energy consumption requirements and the current state of charge. S4. With the goal of minimizing the total downtime, charging cost and battery health deterioration of the embodied intelligent mobile robot cluster, the charging timing and charging amount of each robot are jointly optimized under the condition of satisfying the system operation constraints. S5. When the charging time of multiple robots conflicts, the charging station resources will be competitively and collaboratively scheduled based on the charging urgency index. S6. Based on time-of-use electricity pricing and robot charging load forecasts, coordinate the charging and discharging of the energy storage buffer unit at the charging station. S7. Dynamically adjust the target value of the state of charge of each robot according to the urgency of the task, the joint optimization results and the competitive scheduling results, and feed back the adjusted target value of the state of charge to S3 to form a dynamic adjustment cycle. S8. Send the scheduling instructions to each robot and the energy storage controller of the charging station.

2. The method for wireless charging scheduling and energy storage collaborative optimization of embodied intelligent mobile robots based on task prediction as described in claim 1, characterized in that, In step S2, the energy consumption prediction model adopts a long short-term memory network or a Transformer architecture. The input features include at least the task type code, task duration, task path length, task load weight, average power consumption of historical similar tasks, robot current charge state and running state. The output is the predicted power consumption value for each time step within the future time window, and the predicted power consumption value is integrated to obtain the predicted value of total energy consumption demand.

3. The method for wireless charging scheduling and energy storage collaborative optimization of embodied intelligent mobile robots based on task prediction as described in claim 1, characterized in that, In step S3, determining that a robot has a charging requirement includes: setting a safe charge state limit; when the robot's current remaining available power is lower than the safe reserved power determined by the safe charge state limit, or when the value of the current remaining available power minus the predicted task energy consumption requirement is lower than the safe reserved power, it is determined that the robot has a charging requirement.

4. The method for wireless charging scheduling and energy storage collaborative optimization of embodied intelligent mobile robots based on task prediction as described in claim 3, characterized in that, In S3, the charging urgency index The calculation formula is: ； in, The current state of charge of the i-th embodied intelligent mobile robot. To the limit of safe state of charge, It represents the urgency of the current battery level approaching the safe threshold. The closer When the urgency level increases, the charging urgency level rises; , , Preset weighting coefficients; For the task time limit, As a task priority weight; Charging demand indicators The calculation formula is: ； in, The target value for the state of charge is set to 90% when performing emergency tasks and 80% when performing routine tasks, and is subsequently dynamically determined by S7. The rated battery capacity of the i-th embodied intelligent mobile robot; The net energy value that the i-th embodied intelligent mobile robot needs to charge its battery is always a non-negative value; After charging, the state of charge should be restored to at least 10% of the safe state of charge limit.

5. The method for wireless charging scheduling and energy storage collaborative optimization of embodied intelligent mobile robots based on task prediction according to claim 1, characterized in that, In step S4, a multi-objective optimization model is established to minimize the total downtime, charging cost, and battery health deterioration of the embodied intelligent mobile robot cluster, thereby obtaining the optimal charging time and optimal charging amount for each embodied intelligent mobile robot. The objective function of the multi-objective optimization model is: ； in, Let i be the downtime of the i-th embodied intelligent mobile robot due to charging. For charging costs, The charging process degrades battery health. These are preset weighting coefficients.

6. The method for wireless charging scheduling and energy storage collaborative optimization of embodied intelligent mobile robots based on task prediction according to claim 1, characterized in that, In S4, the system operation constraints include: charging station capacity constraints, charging power constraints, state of charge safety constraints, and task completion constraints. The capacity constraint for charging stations is: the number of charging stations serving at the same time cannot exceed the number of available charging stations. The charging power constraint is: the charging power of a single robot shall not exceed the rated maximum power; The state of charge safety constraints are as follows: during robot operation, the real-time state of charge shall not be lower than the safety lower limit, and after charging, the state of charge shall not exceed the upper limit and shall be restored to at least 10% of the level exceeding the safety lower limit; The task completion constraint is: the robot must complete charging and arrive at the task location before the task deadline.

7. The method for wireless charging scheduling and energy storage collaborative optimization of embodied intelligent mobile robots based on task prediction according to claim 1, characterized in that, In S5, the competitive scheduling of charging station resources includes: allocating charging time slots in descending order of charging urgency index; For robots that do not receive priority charging time, alternative charging plans are generated based on their charging urgency index and task time limit. Alternative charging options include a standby charging option or a partial charging option; The waiting charging plan involves postponing the charging start time to the earliest available idle time at the charging station, recalculating the state of charge after charging is completed and whether it can meet the task time limit. If it can, it will be used as an alternative charging plan. Some charging solutions involve reducing the charging amount to the minimum value that meets the minimum charging guarantee constraint if the waiting time is unacceptable. If none of the alternative charging options can meet the task time limit, a system alarm will be triggered. If the alternative solution adjusts the optimal charging amount or optimal charging timing, the adjusted parameters will be fed back to S4 for re-evaluation and optimization.

8. The method for wireless charging scheduling and energy storage collaborative optimization of embodied intelligent mobile robots based on task prediction according to claim 1, characterized in that, In step S6, the coordinated scheduling of the energy storage buffer unit specifically includes: Obtain time-of-use electricity pricing information from the power grid; During periods of low electricity prices, the grid charges the energy storage buffer unit; during periods of high electricity prices or peak periods of robot charging load, the energy storage buffer unit prioritizes power supply. When the state of charge of the energy storage buffer unit is lower than the preset lower limit, the power supply is switched to the grid directly. When the mains power is interrupted, the energy storage buffer unit switches to emergency power supply mode.

9. The method for wireless charging scheduling and energy storage collaborative optimization of embodied intelligent mobile robots based on task prediction according to claim 1, characterized in that, In step S8, the scheduling instructions generated in steps S4-S7 are sent to each robot and the energy storage controller of the charging station through a wireless communication network. Each robot goes to the designated charging station for charging within a specified time according to the scheduling instructions. Real-time collection of status data during execution and feedback to the scheduling center for iterative optimization; The execution feedback data includes: real-time collection of the actual charging start time, actual charging duration, actual power consumption curve and task completion status of each robot; comparison of the collected data with the predicted data and calculation of the prediction error; when the accumulated error exceeds the preset threshold, the incremental retraining of the energy consumption prediction model is triggered.

10. A wireless charging scheduling and energy storage collaborative optimization system for embodied intelligent mobile robots based on task prediction, characterized in that, The method for wireless charging scheduling and energy storage collaborative optimization of embodied intelligent mobile robots based on task prediction, as described in any one of claims 1 to 9, is applied.

Images