Multi-task priority scheduling method and device for medical robot

Abstract

The application provides a multi-task priority scheduling method and system for a medical robot, and belongs to the technical field of robots, and specifically comprises: a time period division module responsible for determining a scheduling processing method of the medical robot in a time period, using the scheduling processing method and an idle control method to determine available robots, a task acquisition module responsible for collecting task data, a priority evaluation module for calculating a total priority score of a task priority through a multi-dimensional priority evaluation model, dividing priority levels, a robot state acquisition module for collecting the current position, load, equipment health state and function adaptation state of each available robot in real time, and a scheduling and distribution module for sorting candidate robots meeting preset conditions according to priority levels based on the task priority and the state of the available robots, and selecting the robot with the smallest load and the closest distance to distribute the task, thereby improving the reliability of the distribution processing.

Description

Technical Field

[0001] This invention belongs to the field of robotics technology, and in particular relates to a multi-task priority scheduling method and apparatus for medical robots. Background Technology

[0002] With the development of medical informatization and intelligentization, medical robots have been widely used in hospital inpatient wards, undertaking auxiliary medical tasks such as health education and intelligent reminders, effectively reducing the burden on medical staff. Currently, the tasks of medical robots mainly come from two sources: manual assignment by nurses and automatic generation by the management backend based on data from the Hospital Information System (HIS). Moreover, multi-department, multi-robot collaborative operation has become the mainstream deployment mode.

[0003] Specifically, the invention patent application CN202410524815.3, "A Method and System for Path Optimization of Hospital Automated Delivery Robots Based on Deep Learning," utilizes a drug delivery robot to load nursing medications, determine whether the medications match the final medication prescription, and generate a delivery path based on patient-related information to transport the medications. This method has advantages such as improving the efficiency of medication preparation, eliminating the need for manual transport by medical staff, reducing waste of human resources, and improving nursing efficiency. However, it also has the following technical problems: During delivery processing, there may be periods with a large number of delivery tasks. Using a fixed power supply to replenish power may not be enough to meet the delivery needs during busy periods. Therefore, determining a forced idle charging control method for certain periods based on the busyness of delivery tasks in different periods, in order to ensure the reliability of delivery processing, has become an urgent technical problem to be solved.

[0004] Therefore, there is an urgent need for a multi-task priority scheduling method and device for medical robots. Summary of the Invention

[0005] To achieve the objectives of this invention, the following technical solution is adopted: Specifically, this application provides a multi-task priority scheduling device for medical robots, which includes: Time Segmentation Module: Based on historical delivery data in the idle control time period and adjacent time periods, determine the idle control method for the idle control time period; based on the interval data between the time period and different idle control time periods, and combined with the historical delivery data of the time period, determine the scheduling and processing method for medical robots in the time period; and use the scheduling and processing method and the idle control method to determine the available robots. Task Acquisition Module: Collects task data, including task identifier, task type, task objective, task trigger time, urgency level identifier, and task source; Priority assessment module: Calculates the total priority score of tasks through a multi-dimensional priority assessment model and classifies priority levels; Robot Status Acquisition Module: Collects real-time data on the current position, load, device health status, and functional compatibility status of each available robot. Scheduling and allocation module: Based on task priority and the status of available robots, sort them by priority, filter candidate robots that meet preset conditions, and select the robot with the least load and the closest distance to assign tasks.

[0006] The beneficial effects of this invention are as follows: A multi-dimensional priority assessment model (including standardized formulas and scenario-based calibration rules) ensures that urgent and high-impact tasks are executed first, reducing the risk of medical delays and improving service safety and standardization. Multi-machine collaborative scheduling optimizes resource allocation, avoiding idleness and overload. Combined with the logic of fine-tuning the time consumption of tasks with the same priority, it further improves the overall scheduling efficiency and robot utilization. Based on historical delivery data from idle control periods and adjacent periods, an idle control method is determined for each idle control period, constructing a dual evaluation framework: first, assessing the service risk of future adjacent periods; and second, assessing the cumulative extent of the system's historical use of aggressive charging strategies. By determining whether the system has accumulated sufficient power safety redundancy through numerous historical aggressive charging periods, the framework dynamically decides whether the current period should adopt high-intensity charging to further strengthen reserves, or lower-intensity charging to release more readily available robot resources. This achieves an intelligent trade-off between long-term power health and short-term scheduling flexibility.

[0007] Based on the interval data between time periods and different idle control periods, as well as the historical delivery data of the time periods, the scheduling and processing methods for medical robots in the time periods are determined. When the remaining power due to the current idle control strategy is insufficient to meet the delivery requirements, the system assesses whether there are sufficient planned charging opportunities near the current time period and the demand pressure of the current and future time periods. With real-time delivery delay as a constraint, the upper limit of the number of charging robots is dynamically adjusted to find the optimal charging solution that does not affect the current service. This method aims to ensure the absolute priority of service response, while proactively improving the cluster's power health by utilizing all possible fragmented time.

[0008] Furthermore, the available robot is one that is not forced to operate in an idle state.

[0009] Furthermore, the evaluation model includes task urgency weight, urgency coefficient, task type weight, task type coefficient, patient priority weight, patient condition coefficient, task timeliness weight, and timeliness coefficient.

[0010] Furthermore, candidate robots that meet preset criteria are selected, specifically including: Priority sorting: sorted by level 1 > level 2 > level 3 > level 4, and sorted by score for the same priority. Candidate robot selection: Must meet the following conditions: Available robots in the same department or the nearest cross-department robots; Functional adaptation to task requirements; Current workload ≤ 3 (preset threshold). Optimal allocation: Select the robot with the "lowest current load + closest to the target ward" from the candidate robots, issue task instructions and update the load status synchronously.

[0011] Furthermore, it also includes an exception handling module, which monitors the task execution status in real time. If an exception occurs, such as robot failure or task timeout, the priority assessment and scheduling allocation will be retried by calculating the urgency correction formula. Prioritize high-priority tasks: When a new high-priority task is introduced, the current low-priority task is paused (while preserving its progress), and the high-priority task is executed first. Once completed, the low-priority task is resumed.

[0012] Specifically, this application provides a multi-task priority scheduling method for medical robots, applied to the aforementioned multi-task priority scheduling device for medical robots, specifically including: S1 uses scheduling data from the scheduling strategy to determine idle medical robots in different time periods. Based on time period data where no idle medical robots exist, it determines the idle control period for medical robots. Based on the idle control period and historical delivery data in adjacent time periods, it determines the idle control method for the idle control period. S2 uses the idle control method to actively control the idle state of the medical robot. Based on the monitoring data of the remaining power of the medical robot in different time periods, and combined with the delivery data of the robot in adjacent time periods, when it is determined that the idle control period needs to be updated and identified, the process proceeds to the next step. S3 determines the scheduling method for the medical robot during the time period based on the interval data between the time period and different idle control time periods, and in combination with the historical delivery data of the time period.

[0013] Furthermore, the idle medical robot refers to a medical robot that is in an idle state, specifically a medical robot whose idle time is longer than a preset idle time threshold.

[0014] Furthermore, the method for determining the idle control period of the medical robot is as follows: Time periods when there are no idle medical robots are defined as busy time periods, and time periods that are busy time periods on different dates are defined as filtered busy time periods; The busyness coefficient of the time period is determined based on the percentage of dates that fall within the busy period in different time periods; Based on the selected busy periods and the busy coefficients in different periods, the idle control periods of the medical robot are determined.

[0015] Other features and advantages will be set forth in the following description, and the objects and other advantages of the invention are realized and obtained through the structures particularly pointed out in the description and the drawings.

[0016] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0017] The above and other features and advantages of the present invention will become more apparent from a detailed description of exemplary embodiments thereof with reference to the accompanying drawings.

[0018] Figure 1 This is a framework diagram of a multi-task priority scheduling device for medical robots; Figure 2 This is a flowchart of a multi-task priority scheduling method for medical robots; Figure 3 This is a flowchart illustrating the method for determining the idle control period of a medical robot; Figure 4 This is a flowchart illustrating the method for determining the idle control method during the idle control period. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0020] Example 1 like Figure 1 As shown, this application provides a multi-task priority scheduling device for medical robots, specifically including: Time Segmentation Module: Based on historical delivery data in the idle control time period and adjacent time periods, determine the idle control method for the idle control time period; based on the interval data between the time period and different idle control time periods, and combined with the historical delivery data of the time period, determine the scheduling and processing method for medical robots in the time period; and use the scheduling and processing method and the idle control method to determine the available robots. Task Acquisition Module: Collects task data, including task identifier, task type, task objective, task trigger time, urgency level identifier, and task source; Priority assessment module: Calculates the total priority score of tasks through a multi-dimensional priority assessment model and classifies priority levels; Robot Status Acquisition Module: Collects real-time data on the current position, load, device health status, and functional compatibility status of each available robot. Scheduling and allocation module: Based on task priority and the status of available robots, sort them by priority, filter candidate robots that meet preset conditions, and select the robot with the least load and the closest distance to assign tasks.

[0021] Furthermore, the available robot is one that is not forced to operate in an idle state.

[0022] Furthermore, the evaluation model includes task urgency weight, urgency coefficient, task type weight, task type coefficient, patient priority weight, patient condition coefficient, task timeliness weight, and timeliness coefficient, and calculates the total score using the following core formula:

[0023] Where S is the total score of task priority, w1, w2, w3, w4 are the weights of task urgency, task type, patient priority, and task timeliness, respectively, and c1, c2, c3, c4 are the corresponding coefficients. Grading: The total score calculated according to the core formula is divided into four levels (Level 1 is the highest): ① Level 1: score ≥ 0.9; ② Level 2: 0.7-0.89; ③ Level 3: 0.5-0.69; ④ Level 4: < 0.5.

[0024] Detailed parameter definitions and value specifications: Table 1 Weight Values

[0025] 2. Formula constraints (to ensure the rationality of calculations): Weight normalization constraint: (Ensure that the weighting allocation conforms to the "importance ratio" logic, and avoid duplication or omission); Coefficient value constraints: (To avoid scoring distortion due to abnormal coefficients); Scoring interval constraints: (The lowest score corresponds to the "low urgency, low importance, ordinary patients, long time-sensitive" task, and the highest score corresponds to the "high urgency, high importance, critical patients, immediate time-sensitive" task.)

[0026] Furthermore, candidate robots that meet preset criteria are selected, specifically including: Priority sorting: sorted by level 1 > level 2 > level 3 > level 4, and sorted by score for the same priority. Candidate robot selection: Must meet the following conditions: Available robots in the same department or the nearest cross-department robots; Functional adaptation to task requirements; Current workload ≤ 3 (preset threshold). Optimal allocation: Select the robot "closest to the target ward" from the candidate robots, issue task instructions and update the load status synchronously.

[0027] Furthermore, it also includes an exception handling module, which monitors the task execution status in real time. If an exception occurs, such as robot failure or task timeout, the priority assessment and scheduling allocation will be retried by calculating the urgency correction formula. Urgency adjustment formula after task timeout / abnormality: When a task times out or fails to execute, or when a robot malfunctions, its urgency needs to be increased, and the score is adjusted as follows:

[0028] Where: Sadj is the score after anomaly correction (may cross priority levels, triggering rescheduling); λ is the anomaly urgency coefficient (set according to timeout duration: timeout ≤ 10 minutes: λ=0.1, timeout 10-30 minutes: λ=0.2, timeout > 30 minutes: λ=0.3). Logic: The longer the timeout, the higher the score after correction, and the higher the priority for reassignment to avoid medical delays.

[0029] Grading: Total score calculated according to the core formula It is divided into four levels (Level 1 is the highest): ① Level 1: score ≥ 0.9; ② Level 2: 0.7-0.89; ③ Level 3: 0.5-0.69; ④ Level 4: < 0.5.

[0030] High-priority task priority: When a new high-priority task is added, it is executed first. The new task and the existing tasks in the task queue are re-scored and sorted according to the priority model. Then, candidate robots are selected (current load <= 3), and the robot closest to the task is selected. The task instruction is issued and the load status is updated synchronously. After the robot receives the task instruction, it reorders the tasks in the group according to the distance to the ward. The closer task is executed first. After completion, the low-priority task is resumed.

[0031] Example 2 Specifically, such as Figure 2 As shown, this application provides a multi-task priority scheduling method for medical robots, applied to the aforementioned multi-task priority scheduling device for medical robots, specifically including: S1 uses scheduling data from the scheduling strategy to determine idle medical robots in different time periods. Based on time period data where no idle medical robots exist, it determines the idle control period for medical robots. Based on the idle control period and historical delivery data in adjacent time periods, it determines the idle control method for the idle control period. S2 uses the idle control method to actively control the idle state of the medical robot. Based on the monitoring data of the remaining power of the medical robot in different time periods, and combined with the delivery data of the robot in the time periods, when it is determined that the idle control period needs to be updated and identified, the next step is performed. S3 determines the scheduling method for the medical robot during the time period based on the interval data between the time period and different idle control time periods, and in combination with the monitoring results of the remaining power of different robots during the time period.

[0032] Furthermore, the idle medical robot refers to a medical robot that is in an idle state, specifically a medical robot whose idle time is longer than a preset idle time threshold.

[0033] Specifically, such as Figure 3 As shown, the method for determining the idle control period of the medical robot is as follows: The core objective of this embodiment is to address the contradiction between "immediate service availability" and "continuous battery life" in a medical robot swarm under dynamic and uneven delivery demands. Its decision-making logic involves intelligently identifying the overall system's busy patterns through historical data analysis, rather than relying solely on single-day conditions. Based on this, the system proactively and predictively forces some robots into idle charging states during "relatively idle but necessary charging" periods, thereby ensuring that sufficiently charged robots are available during "absolutely busy" periods. This represents a shift in scheduling strategy from "passive response" to "proactive planning," aiming to optimize the long-term stable operation of the system.

[0034] S11 defines the time period when there are no idle medical robots as the busy time period, and the time period that is a busy time period in different dates as the filtered busy time period; "Busy period" refers to a specific date when all medical robots are on duty and there are no "idle medical robots" (i.e., robots idle for more than a preset threshold, such as 30 minutes). "Filtering busy periods" extracts the periods from historical data (such as the past 14 days) that fall within the "busy period" every day. For example, listing the period from 09:00 to 10:00 every day separately and checking whether it lacks idle robots every day for the past 14 days.

[0035] This step aims to filter out occasional busy periods and extract systemic, recurring peak demand points. Focusing only on periods that are "busy every day" allows for precise identification of the core, unchanging peak times in hospital operations (such as the period after daily ward rounds or peak times for sample delivery). This provides the most reliable and essential set of "critical periods" for subsequent decision-making, serving as the cornerstone for all subsequent strategy formulation. Ignoring busy periods caused solely by special circumstances on a single day makes the strategy more universal and stable.

[0036] Specific example: By analyzing last week's data, the system found that in the two time periods of 9:00-10:00 AM and 2:00-3:00 PM each day, there were no idle robots for more than 6 out of 7 days. After "filtering," it was confirmed that 9:00-10:00 AM was the "filtered busy period" with all 7 days being busy, while there were idle robots on 2:00-3:00 PM only 1 day, so it was not included.

[0037] S12 determines the busy coefficient of the time period based on the proportion of dates belonging to the busy time period in different time periods; The "busy period coefficient" is calculated for each individual time unit (e.g., every 30 minutes), out of the total number of dates in the historical statistics, based on the proportion of dates marked as "busy periods". The formula is: Busy Period Coefficient = (Number of days that were busy periods) / (Total number of days in the statistics). It is a value between 0 and 1, quantifying the "probability" or "frequency" of a period being busy.

[0038] Simply "filtering busy periods" can only identify absolute peaks, but it cannot measure the busyness of other periods. Introducing a "busyness coefficient" provides a more refined global perspective. The higher the coefficient, the busier the period is, and the more robot resources need to be reserved; the lower the coefficient, the less busy the period is, and the more suitable it is to schedule robot charging. This quantitative indicator provides a data basis for subsequent comparisons and threshold judgments.

[0039] Specific example: Continuing from the previous example, the statistical period from 08:30 to 09:00 was a busy period on 5 out of the past 7 days, so its busyness coefficient is 5 / 7≈0.71. However, the period from 13:00 to 13:30 was busy on only 2 out of the 7 days, so its busyness coefficient is 2 / 7≈0.29.

[0040] S13 determines the idle control period of the medical robot based on the selected busy periods and the busy coefficients in different periods.

[0041] Understandably, if the proportion of the number of busy periods in different dates exceeds the preset threshold for the proportion of busy periods, then the periods with a busy coefficient less than the first busy coefficient threshold will be designated as idle control periods. This means that the medical robots in the control section will be kept idle and charged, thereby ensuring reliable delivery during the busy periods.

[0042] The system first determines the density of "filtered busy periods". If the density is high, it indicates that there are very fixed peak periods, which need to be given special attention. The system then adopts a strict method to determine the idle control periods (using a stricter first busy coefficient threshold, such as 0.3).

[0043] When the proportion of "screened busy periods" to the total number of time periods in a day is greater than the "preset busy period proportion threshold" (e.g., 20%), it indicates that the system has significant concentrated fixed peaks. At this time, the strategy tends to be conservative, and those time periods with a busy coefficient lower than the first busy coefficient threshold (e.g., 0.3) are identified as "idle control periods".

[0044] A high percentage of peak periods indicates very clear and concentrated pressure on critical services each day. The system must prioritize ensuring uninterrupted resource availability during these peak times. Therefore, "safe periods" (coefficient < 0.3) that have historically been relatively idle are allocated for robot charging. This is a robust strategy of "prioritizing key areas and charging at the edge."

[0045] Specific example: Continuing from the previous example, suppose a day is divided into 48 time slots (each half-hour slot). If the selected "busy time slots" (e.g., 9:00-10:00) comprise 6 time slots, accounting for 12.5%, which is less than the 20% threshold, this situation will not be triggered. However, if the system analyzes data from another hospital and finds 15 "busy time slots," accounting for 31%, then this situation will be triggered. Subsequently, the system will classify all time slots with a busy coefficient <0.3 as idle control time slots.

[0046] Additionally, it is understandable that if the percentage of peak hours in different dates does not exceed a preset peak hour percentage threshold, the following content will also be included: Scenario 1: If the average busy coefficient in different time periods is greater than the preset busy coefficient threshold, then the time periods with busy coefficients less than the first busy coefficient threshold will be designated as idle control periods. In other words, the medical robots in the control section will be forced to be idle and charged, thereby ensuring reliable delivery during the selected busy periods.

[0047] When the percentage of "screened busy periods" is not greater than the threshold, but the average busy coefficient of all periods is greater than the preset busy coefficient threshold (e.g., 0.5), it means that although there is no concentrated "absolute peak", the entire hospital is under high load. In this case, all periods with a busy coefficient < 0.3 are classified as idle control periods.

[0048] Scenario 2: If the average busy coefficient in different time periods is not greater than the preset busy coefficient threshold, obtain the percentage of time periods on different dates when there are no idle medical robots. If the percentage of time periods on different dates when there are idle medical robots is not greater than the preset percentage threshold for time periods, then the time periods with a busy coefficient less than the second busy coefficient threshold are designated as idle control time periods. That is, the medical robots in the control section are forced to be idle and charged, thereby ensuring reliable delivery during the selected busy time periods. The second busy coefficient threshold is less than the first busy coefficient threshold.

[0049] If the average busy coefficient across different time periods is not greater than the preset busy coefficient threshold, the system then checks if there are any dates where the percentage of time periods with no idle medical robots is greater than the preset percentage threshold (e.g., 80%) (i.e., "extremely busy days" with almost no idle robots throughout the day). If no such dates exist, time periods with a busy coefficient less than a lower second busy coefficient threshold (e.g., 0.15) are designated as idle control periods.

[0050] Since the overall workload is not heavy and there are no "extremely busy days," a lower threshold (0.15) is adopted to deal with this potential risk, allowing charging only during those "almost never busy" periods to reduce the impact on delivery time. This is a defensive strategy of "dealing with extremes and finding opportunities to fill gaps."

[0051] Specific example: Continuing from example S12, the percentage of busy periods did not exceed the threshold, but the average busy coefficient for all periods throughout the day was calculated to be 0.4 (less than 0.5). Further investigation revealed that there were 30% (less than 50%) of the periods with no idle robots, which falls under the category of "generally busy days". Therefore, the system adopted a general strategy, setting only the very few periods with a busy coefficient <0.15 as idle control periods.

[0052] Case 3: If the proportion of days with no idle medical robots in different dates is not greater than the preset threshold for the proportion of the number of time periods, determine whether the proportion of the number of the selected busy time periods is within the preset proportion range. If yes, then the time periods with a busy coefficient less than the first busy coefficient threshold are taken as idle control time periods. If no, then the time periods with a busy coefficient less than the second busy coefficient threshold are taken as idle control time periods.

[0053] When the percentage of days without available medical robots is no greater than 0.5, the decision-making criterion reverts to the concentration of "busy periods." It is determined whether this percentage falls within a preset range (e.g., 10%-20%). If so, it indicates the existence of a manageable fixed peak, and a first threshold (0.3) is applied; if not (i.e., percentage <10%, fixed peaks are very rare), a more stringent second threshold (0.15) is applied.

[0054] This is a refined adjustment for a relatively smooth operating mode. If there are some moderately concentrated peaks (accounting for 10%-20%), the system still needs to reserve resources for these peaks, so a relatively lenient threshold (0.3) is adopted. If even moderately concentrated peaks are rare (accounting for <10%), it indicates that the demand is very dispersed, and the system redundancy may be high. In order to further improve the robot's battery health and charging efficiency, a stricter threshold (0.15) can be adopted to schedule charging in more time periods, achieving "balanced charging and optimized state". This reflects the flexibility of the strategy.

[0055] Specific example: Finally, the system evaluates the complete data of the current hospital: the proportion of busy periods is 12% (falling in the range of 10%-20%). Therefore, the system finally decides to adopt the first busy coefficient threshold of 0.3, identify all periods with busy coefficients below 0.3 as idle control periods, and issue a forced idle charging instruction.

[0056] Specifically, such as Figure 4 As shown, the method for determining the idle control period is as follows: The core objective of this embodiment is to make refined decisions regarding the "percentage of robots to be charged" within a defined "idle control period." Its logic transcends single-dimensional demand analysis, constructing a dual evaluation framework: first, assessing the service risk of adjacent future periods; and second, evaluating the cumulative extent of the system's historical use of aggressive charging strategies. By determining whether the system has accumulated sufficient power safety redundancy through numerous historical aggressive charging periods, it dynamically decides whether the current period should involve high-intensity charging (a preset percentage, such as 30%) to further strengthen reserves, or lower-intensity charging (a second preset percentage, such as 15%) to release more readily available robot resources. This achieves an intelligent trade-off between long-term power health and short-term scheduling flexibility.

[0057] S21 determines the proportion of the idle control period in all time periods based on the idle control period data, and uses it as the idle control period proportion; "Idle Control Period Percentage" refers to the average proportion of all time periods planned for mandatory charging, derived from historical data analysis, to the total number of time periods throughout the day (e.g., 48 half-hour periods). This indicator reflects the sufficiency of "time resources" available for centralized charging at a macro level.

[0058] This forms the basis for strategy adjustments. A lower percentage (e.g., <10%) means that charging opportunities are precious, and the system tends to take a more proactive approach during any available charging period to compensate for the limited time window. A higher percentage means that the system has more scheduling flexibility, allowing for adjustments to charging intensity based on more granular rules (such as risks in adjacent time periods and historical strategy accumulation), avoiding unnecessary resource idleness.

[0059] S22 determines the busy coefficient in the adjacent time periods of the idle control period based on historical delivery data in the adjacent time periods of the idle control period; This step focuses on the upcoming "adjacent period" (i.e., the period immediately following the current idle control period). By analyzing the busy coefficient (historical busy frequency) of this period and determining whether it belongs to a consistently "filtered busy period," the service pressure following it is quantified.

[0060] This is a direct risk input for decision-making. If adjacent time periods are either absolute peaks or highly likely to be busy, the charging strategy for the current period must prioritize "preparatory work." Even if high-intensity charging may temporarily reduce the number of standby robots, it ensures that the robots participating in charging can meet peak demand with higher battery levels. This reflects the forward-looking and safeguarding nature of the decision-making process.

[0061] S23 determines the idle control method for the idle control period based on the idle control period ratio and the busy coefficient in the adjacent periods of the idle control period.

[0062] It should be noted that the percentage of idle control periods is determined based on the percentage of idle control periods on different dates in all time periods.

[0063] Specifically, the idle control method for determining the idle control period includes: Case 1: If the proportion of the idle control period is less than the preset idle period proportion threshold, then the idle control method for the idle control period is determined to be that during the scheduling process, the robot with the least remaining power is kept in an idle state and charged. When the "percentage of idle control time" is less than a preset threshold, it indicates that charging opportunities are scarce in the system. At this time, the decision-making logic is simplified: a high-intensity charging strategy (preset ratio, such as 30%) is adopted in all cases, prioritizing the replenishment of energy to the robot with the lowest battery level.

[0064] When resources (time) are extremely limited, the primary decision-making objective is to quickly raise the minimum power level of the entire robot swarm to prevent a systemic low-power crisis. There is no room for considering historical accumulation or minor future risks; the most direct and effective power replenishment method must be adopted. This is a bottom-line strategy of "resource scarcity, efficiency first."

[0065] Specific example: In a certain hospital, only three time periods at night are designated as idle control periods, accounting for only 6%, less than 10%. During these idle control periods, the system unconditionally executes the instruction: immediately force the robots in the bottom 30% of battery level to enter charging mode.

[0066] Case 2: If the proportion of the idle control period is not less than the preset idle period proportion threshold, and if there is a busy period among the adjacent periods of the idle control period, then the idle control method of the idle control period is to control the power of the robot with the least remaining power in the idle state and charge it during the scheduling process. When the charging window is ample (the percentage is not less than the threshold), but the adjacent time period is a "busy period" (e.g., the current time is 8:30, and the next time is 9:00, which is the daily peak time for ward rounds and medication delivery), a high-intensity charging strategy (preset ratio, such as 30%) is also adopted.

[0067] This is a response to a risk with extremely high certainty. Faced with an inevitable surge in activity, recharging as many low-battery robots as possible is a high-priority replenishment solution, and the intensity must be sufficient.

[0068] Case 3: If there is no busy period to be selected in the adjacent periods of the idle control period, obtain the number of idle control periods in which the power of the robots with the preset proportion is controlled at the rated power. If the number of idle control periods in which the power of the robots with the preset proportion is controlled at the rated power is greater than the preset idle period number threshold, then the idle control method of the idle control period is determined to be that, during the scheduling process, the power of the robot with the least remaining power of the second preset proportion is controlled in an idle state and charged, wherein the second preset proportion is less than the preset proportion. Scenario 4: If the number of idle control periods with a preset proportion of robot power at rated power does not exceed a preset idle period number threshold, determine whether the busy coefficient of the adjacent periods of the idle period is greater than a preset busy coefficient threshold. If yes, the idle control method for the idle control period is to keep the robot with the least remaining power at a preset proportion in an idle state and charge it during the scheduling process. If no, the idle control method for the idle control period is determined to be to keep the robot with the least remaining power at a second preset proportion in an idle state and charge it during the scheduling process.

[0069] When the charging window is ample and there is no highest risk of "selecting busy periods" in adjacent time slots, the system enters a fine-tuning mode. At this point, a key evaluation dimension is introduced: "the number of idle control periods during which a preset proportion of the robot's battery power is controlled within the rated power range." This metric does not measure the success of charging, but rather the accumulated number of historical periods during which the system has executed a "high-intensity charging strategy."

[0070] Scenario 3: If the historical quantity > preset threshold: This indicates that the system has frequently employed aggressive charging strategies in the past. This usually means that the average power level or power safety buffer of the entire robot swarm has been well maintained and consolidated. Therefore, under conditions where the current risk is not high, the system can "take a break" and adopt a lower-intensity charging strategy (second preset ratio, such as 15%), releasing more robots (15% more than under the high-intensity charging strategy) to maintain a schedulable state, thereby enhancing the system's flexibility in responding to temporary tasks.

[0071] Scenario 4: If the historical number is less than or equal to the preset threshold: This indicates that the system has historically adopted aggressive charging tactics infrequently, and the power safety redundancy may not be sufficient. In this case, decision-making needs to cautiously rely on predictions of specific risks in adjacent time periods. If the busy period coefficient is high in adjacent time slots (e.g., >0.6): Although it is not always busy, the probability of being busy is high. In order to cope with this high probability risk and at the same time continue to accumulate power reserves, a high-intensity charging strategy (30%) is adopted.

[0072] If the busy coefficient is low in adjacent time periods (e.g., ≤0.6), the future risk is relatively low. Considering the insufficient accumulation of historical aggressive charging, but also the low current risk, a compromise and conservative strategy can be adopted, namely, to use low-intensity charging (15%), which provides some replenishment without excessively consuming resources.

[0073] This is the essence of the intelligence in this embodiment. It extends decision-making from simply "looking to the future" to "looking to the past," forming a closed loop. The number of historically aggressive charging periods serves as a proxy indicator for the investment in the system's "battery health." Sufficient investment (Scenario 3) allows for greater composure and flexibility in the current situation; insufficient investment (Scenario 4) necessitates a decision based on subtle differences in current risks, whether to increase investment or maintain a moderate level of replenishment. This reflects the continuity and adaptability of the strategy.

[0074] Example of Scenario 3: The period from 3:00 PM to 3:30 PM is an idle control period. The adjacent period from 3:30 PM to 4:00 PM has a busy coefficient of 0.4, which is not a selected busy period. The system queries the historical records and finds that there have been 8 periods where "30% high-intensity charging" has been performed, exceeding the threshold (e.g., 5 periods). The system determines that the historical accumulation of high-intensity charging is sufficient and the power reserve is good. Therefore, in the current period, only the robot with the lowest power level (15%) will be scheduled to charge, allowing more robots to be on standby.

[0075] Example of Scenario 4: The scenario is the same as above, but the system query finds only 3 historical high-intensity charging periods, which is below the threshold. In this case, since the busy coefficient of adjacent periods is 0.4 (≤0.6), it is considered low risk. The system's final decision: adopt a lower-intensity charging strategy (15%). This is partly due to the low future risk, and partly because of insufficient historical accumulation, it is not advisable to excessively consume scheduling resources during low-risk periods. Instead, a moderate and continuous replenishment strategy is adopted.

[0076] This embodiment achieves dynamic optimization and self-adaptation of charging intensity control by constructing a three-layer decision-making model of "global charging resources - future delivery risks - historical strategy accumulation". The decision-making not only considers the upcoming demand, but also evaluates the "progress" of the system's long-term power health construction through historical strategy data, making charging management a planning process with memory and sustainability.

[0077] It achieves an optimal balance between long-term reserves and short-term response: by judging whether historical aggressive charging is "sufficient," it intelligently switches between "continuing to strengthen power reserves" and "releasing resources to enhance immediate response," enabling the system to cope with long-term risks while optimizing daily efficiency and avoiding rigidity in charging strategies. When safety redundancy is sufficient, reducing the charging ratio increases the standby robot pool, directly improving the ability to handle sudden and scattered tasks and optimizing the overall utilization rate of robot assets.

[0078] It should be noted that the adjacent time period refers to the time period that is adjacent to the time period and belongs to the period after the time period.

[0079] Furthermore, it is determined that an update identification process for idle control periods is required, specifically including: The core objective of this embodiment is to establish an intelligent early warning and update mechanism to determine whether the currently set "idle control period" scheme is still effective and whether it needs to be re-identified and updated. Its decision-making logic is based on multi-dimensional risk transmission analysis: it not only focuses on insufficient power within a single period but also on the temporal coupling relationship between insufficient power and delivery pressure. By analyzing the combined risk patterns of "insufficient power periods" and "subsequent busy periods," the system can identify potential systemic risk points under the current charging strategy. Therefore, before delivery reliability is affected, it proactively triggers the replanning of idle control periods, achieving system self-optimization and continuous adaptation.

[0080] S31 uses monitoring data of the remaining power of medical robots at different time periods to determine the medical robots whose remaining power is within a preset remaining power range, and classifies them as robots with insufficient power. It should be noted that the "low battery robot" refers to a medical robot whose remaining battery power is within a preset range after completing the delivery task for the specified time period.

[0081] It is understood that the preset remaining power range is determined based on the distance the robot can transport within different remaining power ranges, and the specific range is determined based on the user's settings.

[0082] The system first defines a "low-battery robot"—a robot whose remaining battery level falls into a preset danger zone (e.g., less than 10%) after completing a delivery task for a certain period. This zone is set based on the robot's travel distance on a single charge and the typical task distance, indicating that the battery is insufficient to reliably perform the next routine task. The system iterates through all historical time periods and statistically analyzes the occurrence of such robots in each period.

[0083] Directly monitoring the remaining battery power after a task is the most fundamental indicator for evaluating the effectiveness of the current charging strategy. If the robot frequently finds itself in a low-battery state after completing a task, it indicates that the current charging schedule (both the charging time and intensity) may not be able to meet the actual workload, posing a risk of delivery disruptions. This is the basic data source that triggers updates.

[0084] Specifically, the above steps include the following: S311 determines whether there are no robots with insufficient power in all time periods. If so, it is determined that no update identification process for idle control periods is needed. If not, proceed to step S312. If no robot with insufficient power is found in the historical records for any time period, it means that the existing charging strategy fully meets the requirements and the system is in an ideal state. Therefore, it is determined that no update is needed. A fast rejection channel is set up to avoid unnecessary complex calculations in a risk-free situation and improve system efficiency.

[0085] S312 takes the time period excluding the idle control period as other time periods, and determines whether there are robots with insufficient power on different dates in other time periods. If so, it is determined that the idle control period needs to be updated and identified. If not, proceed to step S313. The analysis focuses on "other time periods" (i.e., non-idle control periods). If a robot with insufficient power appears on every analysis date within these time periods, it indicates a systemic flaw in the existing charging schedule—the charging periods completely fail to cover the actual high-consumption periods. This necessitates a direct update, representing a fundamental rejection of the charging strategy. It identifies a complete disconnect between the charging period settings and actual power consumption patterns, requiring immediate replanning.

[0086] S313 uses the average percentage of robots with insufficient power on different dates in other time periods to determine the power impact coefficient of the other time periods, and determines whether there are other time periods with a power impact coefficient greater than a preset impact coefficient threshold. If so, it is determined that the idle control period needs to be updated and identified. If not, proceed to step S32.

[0087] If S312 is not met, calculate the power impact coefficient for each "other time period" (the average percentage of days in which the robot experiences insufficient power during that time period). If the power impact coefficient for any "other time period" exceeds a preset threshold (e.g., 0.6), it indicates that power crises frequently occur during that time period, making it a definite high-risk point. The judgment needs to be updated.

[0088] Even if the problem doesn't occur daily, frequently occurring risks are equally dangerous. This step quantifies the "frequency" to identify periods where the problem repeatedly occurs, indicating that these periods may need to be added as idle control periods or that the charging strategies for adjacent periods need to be adjusted.

[0089] S32 determines the busy coefficient of the adjacent time periods based on the delivery data of the robots in the adjacent time periods of the stated time period; This step introduces risk transmission analysis over time. For a period of insufficient power, the risk is not limited to the current period but will also propagate to the immediately following adjacent period (i.e., the next period). The system calculates the "adjacent period busy coefficient" for each period and combines it with its own "power impact coefficient" to calculate the "delivery impact value".

[0090] The above steps include the following: S321 determines the delivery impact value of adjacent time periods based on the busy coefficient and power impact coefficient of the adjacent time period, and determines whether there are adjacent time periods with a delivery impact value greater than a preset impact threshold. If so, it is determined that the idle control time period needs to be updated and identified. If not, proceed to step S33. For each time period, the delivery impact value is jointly determined by the "battery impact coefficient" and the "adjacent time period busy coefficient" (e.g., delivery impact value = battery impact coefficient × adjacent time period busy coefficient). The higher this value, the greater the risk that "the robot has just finished its work and is running out of power, and then immediately faces a busy task." If any delivery impact value exceeds a preset high threshold (e.g., 0.3), the decision needs to be updated.

[0091] This captures the most dangerous scenario. It assesses the risk of a "perfect storm"—the direct overlap of a power shortage and delivery pressure in time. This combination implies an extremely high probability of delivery failure, representing a top-tier risk that must be mitigated by updating charging strategies.

[0092] S33 determines whether an update identification process for idle control periods is needed based on the low-power robot during the specified time period and the busy coefficient of adjacent time periods.

[0093] It is understandable that the higher the busyness coefficient of adjacent time periods, the higher the power consumption impact coefficient, and therefore the higher the delivery impact value.

[0094] Specifically, in the above steps, if there are multiple delivery impact values ​​in adjacent time periods within the preset impact value range, it is determined that an update identification process for idle control time periods is required; otherwise, it is determined that an update identification process for idle control time periods is not required.

[0095] If none of the above steps trigger an update, the system proceeds to the final assessment. This step checks if there are multiple time periods whose delivery impact values, while not reaching the standard for a single extremely high-risk scenario, all fall within the preset medium-risk range (e.g., more than three time periods with delivery impact values ​​between 0.1 and 0.3). If so, an update is required; otherwise, the existing plan is maintained.

[0096] This step aims to identify systemic vulnerabilities, rather than single points of failure. The simultaneous presence of moderate risk across multiple time periods indicates that the current charging strategy is putting the system under strain. While no serious problems have yet surfaced, redundancy is insufficient and resilience is poor. Any minor fluctuation (such as a slight increase in workload or a temporary decrease in charging efficiency) could lead to a cascading failure. Triggering an update is to proactively improve the overall robustness and resilience of the system.

[0097] This embodiment constructs a three-tiered progressive analysis framework of "individual risk frequency screening -> time-series risk transmission assessment -> system vulnerability pattern identification," realizing a forward-looking and adaptive idle control period update triggering mechanism. It achieves a leap from "post-event remediation" to "pre-event prevention": by analyzing risk patterns in historical data, it proactively issues warnings and adjusts strategies before delivery reliability is actually compromised, ensuring continuous stability of service quality. It innovatively correlates "battery status" with "demand pressure" on a timeline, accurately locating key risk links that could lead to task failure due to improper charging arrangements. It can identify and handle extreme high-risk individual points (step S321) and also detect and improve widely distributed systemic moderate risks (step S33), making optimization decisions more comprehensive.

[0098] Specifically, the method for determining the scheduling and processing method of the medical robots during the aforementioned time period is as follows: The core objective of this embodiment is to address the nuanced decision-making process of whether and how many robots should be proactively charged during any non-idle control period. Its core logic involves constructing a three-layer decision filter based on "prioritization, risk anticipation, and dynamic balancing": First, it assesses whether there are sufficient planned charging opportunities in the vicinity of the current time period; if so, it prioritizes ensuring current service. Second, when charging opportunities are insufficient, it comprehensively assesses the demand pressure of the current and future periods to decide whether to intervene in charging. Finally, when deciding to charge, it dynamically adjusts the upper limit of the number of charging robots, constrained by real-time delivery delays, to find the optimal charging solution that does not affect current service. This method aims to ensure absolute priority in service response while proactively improving the cluster's battery health by utilizing all possible fragmented time.

[0099] S41 uses the interval data between the time period and different idle control time periods to determine the idle control time period whose interval length is within a preset interval length range, and takes the idle control time period whose interval length is within the preset interval length range as the adjacent control time period. "Adjacent control periods" are the core concept of this step, specifically referring to idle control periods on a future timeline whose time interval between them and the currently decided period (referred to as the "current period") falls within a preset range (e.g., 30 minutes to 2 hours). The system identifies all planned mandatory charging windows that are about to arrive in the near future by calculating the time difference.

[0100] This step is designed to implement the primary principle of "trading time for space and ensuring immediate service." Assessing the number of "adjacent control periods" essentially measures whether the system has sufficient opportunities for centralized, planned power replenishment in the near future. If the answer is yes, then the "opportunity cost" of proactively reducing the number of available robots for charging during the current period, when there may still be temporary delivery needs, is too high and unnecessary. This judgment, acting as the first filter, effectively avoids ineffective or negative scheduling interventions when charging resources are generally abundant, ensuring that robot resources are prioritized to meet immediate and unpredictable delivery needs.

[0101] Specific example: Suppose the system is making a decision regarding the non-idle control period of 10:30 AM. It calculates and finds that there are two preset idle control periods in the future, 11:00-11:30 and 12:00-12:30, and the interval between them and 10:30 is within 1.5 hours. Therefore, these two periods are identified as "adjacent control periods".

[0102] S42 determines the busyness coefficient of the time period after the stated time period based on historical delivery data of the time period after the stated time period; The "busyness factor of subsequent time periods" is the basis for evaluation in this step. It refers to the historical average busyness factor of multiple consecutive time periods (e.g., the next four time periods) starting from the current time period. The system quantifies the expected delivery demand pressure in the medium term by calculating the average (or weighted average) of the busyness factors of these subsequent time periods.

[0103] The significance of this step lies in injecting a forward-looking strategic perspective into scheduling decisions. The impact of charging behavior is continuous; robots currently being charged will be unavailable for some time in the future. Therefore, decisions must consider the impact of charging behavior on future service capacity. If a very busy period is expected in the future, then actively charging during the current (relatively idle) period has high strategic value; conversely, if idleness is expected in the future, the urgency of charging is lower. This assessment provides a crucial risk judgment basis for subsequently selecting charging strategies of different intensities.

[0104] Specific example: Continuing from the previous example, the system calculates the historical average busy coefficient for the four time periods after 10:30 (11:00, 11:30, 12:00, 12:30), and the result is 0.7, indicating that the next two hours are expected to be relatively busy.

[0105] S43 determines the scheduling method for the medical robot in the time period based on the adjacent control time periods of the time period, the busy coefficient of the time period after the time period, and the busy coefficient of the time period.

[0106] Furthermore, if the number of adjacent control periods in the time period is greater than a preset threshold for the number of adjacent control periods, then the scheduling method for the medical robot in the time period is determined to be no scheduling process required.

[0107] Judging the adequacy of charging: Judgment and Operation: If the number of identified "adjacent control periods" is greater than the preset threshold for the number of adjacent control periods (e.g., 1), the system directly determines that the current period "does not require scheduling processing".

[0108] This is a direct application of the S41 conclusion, embodying the principle of minimizing opportunity cost. When at least one definite charging window exists in the near future, the system chooses to postpone charging demands to the planned time slot, thereby reserving all robot resources in the current time slot to ensure maximum service responsiveness. This is the most robust strategy, ensuring complete decoupling of planned charging and on-demand delivery services in terms of resource consumption.

[0109] For example, in the decision-making process at 10:30, since two "adjacent control periods" (number > threshold 1) were identified, the system directly decided not to execute any active charging instructions during that period, and all robots remained in standby mode.

[0110] Furthermore, if the number of adjacent control periods is not greater than a preset threshold for the number of adjacent control periods, the following is also included: S431 determines whether the frequency coefficient of the time period is greater than the preset value of the busy coefficient. If yes, it determines that the scheduling processing method of the medical robot in the time period is no scheduling processing required. If no, it proceeds to step S432. When the number of "adjacent control periods" is insufficient (≤ threshold), the system first checks whether the "busy coefficient" of the current period itself is greater than a preset busy coefficient threshold. If so, it is determined that "no scheduling processing is required".

[0111] This judgment establishes the ironclad rule that "service assurance takes absolute priority over power supply maintenance." Even if future charging opportunities are scarce, as long as the current period is historically busy, the system will never proactively reduce available capacity at present. This prevents the system from making erroneous decisions that create a certain service crisis in the present to cope with future risks, and is the bottom-line logic for ensuring service reliability.

[0112] Specific example: Suppose the system is making a decision about 2:00 PM, and there are no adjacent control periods within the next 2 hours. However, the busy coefficient of 2:00 PM itself is as high as 0.5 (> the threshold of 0.4), so the system determines that no charging scheduling will be performed during this period.

[0113] S432 determines whether there are adjacent control time periods in the time period. If yes, proceed to step S433. If no, determine that the scheduling and processing method of the medical robot in the time period is the preset scheduling and processing method. S433 uses the busy coefficient of the time period after the stated time period to determine whether the average busy coefficient of the time period after the stated time period is greater than the preset busy coefficient value. If so, the scheduling method of the medical robot in the stated time period is determined to be the preset scheduling method. If not, the scheduling method of the medical robot in the stated time period is determined to be the second preset scheduling method.

[0114] After passing through the first two layers of filtering (i.e., when there are few charging opportunities and the system is not busy), the system enters the stage of refined strategy selection.

[0115] Path A (S432): If there are no adjacent control periods in the current time period, it means that there is a complete lack of planned charging opportunities in the near future. At this time, the system adopts the "preset scheduling processing method" (i.e., aggressive charging strategy).

[0116] Path B (S433): If there is at least one “adjacent control period”, the average busy coefficient of the future period obtained from S42 needs to be evaluated.

[0117] If the coefficient is high (>threshold): it indicates that a busy period is about to begin, and active preparation is needed, adopting the "pre-set scheduling processing method" (aggressive charging).

[0118] If the coefficient is low (≤ threshold): it indicates that the recent pressure is small, and the "second preset scheduling method" (gentle charging) can be used.

[0119] This layer achieves a fine balance between risk and reward. The "pre-set scheduling method" is used for high-risk or high-urgency scenarios (no future opportunities or very busy future), aiming to seize limited opportunity windows and replenish as much power as possible, with a stronger "investing in the future" attribute. The "second pre-set scheduling method" is used for low-risk scenarios (future charging opportunities and not busy future), aiming to provide gentle, maintenance-oriented power replenishment while ensuring minimal impact on current service capacity. Both methods control charging intensity through different "target quantities."

[0120] Specific examples: Suppose the system makes a decision for time period A, which has a busy coefficient of 0.4 and no adjacent control time periods (path A). Due to the lack of charging opportunities in the near future and the high risk, the system decides to adopt a "preset scheduling method" (aggressive charging), that is, idle control is implemented as long as there is remaining power within a preset remaining power range. Suppose the system makes a decision for time period B, which is not busy, has an adjacent control time period at 5:00, and has a future average busy coefficient of 0.5 (path B, greater than the threshold of 0.4). The system then adopts the "preset scheduling method" (aggressive charging), that is, idle control is implemented as long as there is remaining power within a preset remaining power range.

[0121] Furthermore, the preset scheduling processing method is to put any medical robot with remaining power within a preset remaining power range into an idle state for charging. It should also be noted that during the idle control, at least a preset proportion (e.g., 0.5) of the medical robots are in working state at different times.

[0122] Furthermore, when there are medical robots with remaining power within a preset remaining power range, the second preset scheduling method uses a target number as a constraint and only performs idle control processing on the target number of medical robots with remaining power within the preset remaining power range during the specified time period, so that they are in an idle state for charging. It should also be noted that during the idle control, at least a preset proportion (e.g., 0.7) of medical robots are in working state at different times.

[0123] Furthermore, when the target number of medical robots with remaining power within the preset remaining power range are idled during the specified time period, if there are more than the target number of medical robots with remaining power within the preset remaining power range at the same time, the target number of medical robots with the smallest remaining power within the preset remaining power range will be idled. If there are not more than the target number of medical robots with remaining power within the preset remaining power range at the same time, the idled medical robots with remaining power within the preset remaining power range will be idled gradually until the target number is reached.

[0124] It should be noted that when performing scheduling processing during the time period, the delivery delay of the time period under the current scheduling strategy is determined. When the delivery delay does not meet the requirements, the number of medical robots under idle control is gradually reduced until the number that meets the delivery delay requirements is obtained, and the scheduling processing is carried out using the number that meets the delivery delay requirements.

[0125] Both methods share the same execution logic, with the core difference being the different preset values ​​for the target number (the second preset method has a target number of N2, while the third preset method has no limit). The operation process is as follows: 1) Identify all robots with current battery levels in a preset danger range (e.g., <10%); 2) Sort them by battery level from low to high; 3) Execute "limited priority charging": if the total number of low-battery robots is ≤ the target number N2, then all are charged until the number of idle medical robots during the specified time period reaches the target number N2; if the total number is > N2, then only the N robots with the lowest battery levels are charged.

[0126] This design embodies the principle of combining "prioritizing the most needed" with "acting within our means." Prioritizing charging the robot with the lowest battery level quickly raises the overall battery safety threshold of the robot cluster, representing the optimal solution for risk mitigation. Simultaneously, setting a clear "target number" upper limit establishes a safety margin for delivery service capacity during the current period, ensuring a minimum number of available robots are always maintained, thus preventing service disruptions due to overcharging. This mechanism establishes a quantifiable and controllable balance between "replenishing battery power" and "ensuring service."

[0127] Specific example: In an aggressive charging scenario, the target number N² = 2. Assuming that 5 robots have a battery level below 10%, the system will sort them by battery level and only control the 2 robots with the lowest battery levels to enter the charging state, while the remaining medical robots will remain in the available state.

[0128] If only one robot is below 10% at any given time, then determine whether the number of robots undergoing idle control processing in the current time period has reached 2. If it has reached 2, no idle control processing is required; if it has not reached 2, then idle control processing will continue.

[0129] "Delivery delay" refers to the waiting time from when a delivery task is issued to when a robot begins to execute that task. This mechanism monitors the current average delivery delay in real time while implementing charging control according to the above decisions. If the delay exceeds a preset satisfactory threshold, the system considers the current charging control intensity (i.e., the target number N) to be affecting immediate service and initiates adjustment: it gradually reduces the target number (e.g., decreasing it by 1 each time) and reassesses the delay, repeating this process until a new, smaller target number N' is found, such that when controlling N' robots to charge, the delivery delay returns to a satisfactory level.

[0130] This is the ultimate security guarantee and the core of adaptive optimization in this method. All the aforementioned decisions are based on historical data and future predictions, while real-time "delivery delay" is the most authentic and direct feedback signal of the current system's service capacity. Introducing this feedback loop enables the system to have the ability to "learn online" and "correct deviations instantly." It can dynamically detect deviations between theoretical decisions and actual situations (such as a sudden increase in temporary tasks) and immediately adjust the charging strategy to ensure that the highest principle of "not affecting the current delivery service" is strictly adhered to in every execution, thereby achieving a perfect unity between the global optimization goal (charging) and local constraints (instant service).

[0131] Specific example: During time period A, the system aggressively controls 5 robots to charge. However, after execution, it detects that the average delivery delay has suddenly increased from the normal 30 seconds to 3 minutes. The system then automatically reduces the target number from 5 to 4, releasing one robot, and the delay drops to 2 minutes, but is still too high. The system further reduces it to 2, at which point the delay returns to a satisfactory level of 40 seconds. The system then locks this new target number N'=2 and executes subsequent charging control accordingly.

[0132] This embodiment constructs a complete closed loop of "proximity opportunity assessment -> self and future demand assessment -> hierarchical decision-making and dynamic feedback," realizing a highly intelligent and adaptive unplanned charging scheduling method. Through three layers of filters and real-time feedback, it ensures that no charging behavior will sacrifice the reliability of current and foreseeable future delivery services. It can proactively identify risk windows and utilize fragmented time to carry out preventative charging, nipping power crises in the bud and improving the long-term robustness of the system.

[0133] Example 2 In a second aspect, the present invention provides a computer system comprising: a memory and a processor connected in communication, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the above-described multi-task priority scheduling method for a medical robot when running the computer program.

[0134] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0135] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0136] The above description is merely one or more embodiments of this specification and is not intended to render this specification useless. For those skilled in the art, various modifications and variations can be made to the one or more embodiments of this specification. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of one or more embodiments of this specification should be included within the scope of the claims of this specification.

Claims

1. A multi-task priority scheduling device for medical robots, characterized in that, Specifically, it includes: Time Segmentation Module: Based on historical delivery data in the idle control time period and adjacent time periods, determine the idle control method for the idle control time period; based on the interval data between the time period and different idle control time periods, and combined with the historical delivery data of the time period, determine the scheduling and processing method for medical robots in the time period; and use the scheduling and processing method and the idle control method to determine the available robots. Task Acquisition Module: Collects task data, including task identifier, task type, task objective, task trigger time, urgency level identifier, and task source; Priority assessment module: Calculates the total priority score of tasks through a multi-dimensional priority assessment model and classifies priority levels; Robot Status Acquisition Module: Collects real-time data on the current position, load, device health status, and functional compatibility status of each available robot. Scheduling and allocation module: Based on task priority and the status of available robots, sort them by priority, filter candidate robots that meet preset conditions, and select the robot with the least load and the closest distance to assign tasks.

2. The multi-task priority scheduling device for medical robots as described in claim 1, characterized in that, The available robots are those that are not subject to mandatory control and are kept in an idle state.

3. The multi-task priority scheduling device for medical robots as described in claim 1, characterized in that, The evaluation model includes task urgency weight, urgency coefficient, task type weight, task type coefficient, patient priority weight, patient condition coefficient, task timeliness weight, and timeliness coefficient.

4. The multi-task priority scheduling device for medical robots as described in claim 1, characterized in that, The process involves filtering candidate robots that meet preset criteria, specifically including: Priority sorting: sorted by level 1 > level 2 > level 3 > level 4, and sorted by score for the same priority. Candidate robot selection: Must meet the following conditions: Available robots in the same department or the nearest cross-department robots; Functional adaptation to task requirements; Current workload ≤ 3 (preset threshold). Optimal allocation: Select the robot with the "lowest current load + closest to the target ward" from the candidate robots, issue task instructions and update the load status synchronously.

5. The multi-task priority scheduling device for medical robots as described in claim 1, characterized in that, It also includes an exception handling module that monitors the task execution status in real time. If an exception occurs, such as robot failure or task timeout, the priority assessment and scheduling allocation will be retried by calculating the urgency correction formula.

6. A multi-task priority scheduling method for a medical robot, applied to the multi-task priority scheduling device for a medical robot as described in any one of claims 1-5, characterized in that, Specifically, it includes: S1 uses scheduling data from the scheduling strategy to determine idle medical robots in different time periods. Based on time period data where no idle medical robots exist, it determines the idle control period for medical robots. Based on the idle control period and historical delivery data in adjacent time periods, it determines the idle control method for the idle control period. S2 uses the idle control method to actively control the idle state of the medical robot. Based on the monitoring data of the remaining power of the medical robot in different time periods, and combined with the delivery data of the robot in adjacent time periods, when it is determined that the idle control period needs to be updated and identified, the process proceeds to the next step. S3 determines the scheduling method for the medical robot during the time period based on the interval data between the time period and different idle control time periods, and in combination with the historical delivery data of the time period.

7. The multi-task priority scheduling method for medical robots as described in claim 6, characterized in that, The idle medical robot is a medical robot that is in an idle state. Specifically, a medical robot whose idle time is longer than a preset idle time threshold is considered an idle medical robot.

8. The multi-task priority scheduling method for medical robots as described in claim 6, characterized in that, The method for determining the idle control period of the medical robot is as follows: Time periods when there are no idle medical robots are defined as busy time periods, and time periods that are busy time periods on different dates are defined as filtered busy time periods; The busyness coefficient of the time period is determined based on the percentage of dates that fall within the busy period in different time periods; Based on the selected busy periods and the busy coefficients in different periods, the idle control periods of the medical robot are determined.

9. The multi-task priority scheduling method for medical robots as described in claim 8, characterized in that, If the proportion of the number of busy periods in different dates is greater than the preset threshold for the proportion of busy periods, then the periods with a busy coefficient less than the first busy coefficient threshold will be designated as idle control periods. In other words, the medical robots in the control section will be forced to be idle and charged, thereby ensuring reliable delivery during the selected busy periods.

10. The multi-task priority scheduling method for medical robots as described in claim 6, characterized in that, The method for determining the scheduling and processing method of medical robots during the aforementioned time period is as follows: Based on the interval data between the time period and different idle control time periods, determine the idle control time periods whose interval length is within a preset interval length range from the time period, and take the idle control time periods whose interval length is within the preset interval length range from the time period as adjacent control time periods. Based on historical delivery data for periods following the stated time period, determine the busyness factor for periods following the stated time period; Based on the adjacent control periods of the time period, the busy coefficient of the time period after the time period, and the busy coefficient of the time period, the scheduling and processing method of the medical robot in the time period is determined.

Images