Robot selling task processing method and device and storage medium

Abstract

The invention discloses a robot selling task processing method and device and a storage medium, and relates to the technical field of robots, and the method comprises the steps: obtaining an influence factor of a selling point demand based on a sensor disposed at a selling point, the influence factor comprising the current queuing number of people at the selling point and the commodity sales volume; when the current queuing number of people is greater than a preset queuing number of people threshold or the commodity sales volume is greater than a preset commodity sales volume, determining that the demand change occurs at the selling point; and determining a first adjustment duration of the corresponding task according to a preset queuing model and the influence factor, and adjusting a time window of the task according to the first adjustment duration. By dynamically adjusting the task duration of task execution of the robot, the idle or overload condition of the robot can be reduced, task interruption or insufficient service caused by demand change is avoided, and the service efficiency of the robot is improved.

Description

Technical Field

[0001] The present application relates to the field of robotics technology, and in particular to a method, device, and storage medium for processing robot vending tasks. Background Art

[0002] Robotic sales tasks involve robots performing various sales-related tasks across various locations. These tasks may include direct sales, such as processing customer purchase requests, or auxiliary tasks, such as restocking, inventory management, and equipment maintenance. However, the current low efficiency of robots performing sales tasks hinders their widespread adoption.

[0003] The above content is only used to assist in understanding the technical solution of this application and does not constitute an admission that the above content is prior art. Summary of the Invention

[0004] The main purpose of this application is to provide a robot sales task processing method, device and storage medium, aiming to solve the technical problem of how to improve the service efficiency of robots in performing sales tasks.

[0005] To achieve the above objectives, the present application proposes a robot vending task processing method, which includes:

[0006] Obtaining, based on sensors installed at the point of sale, factors influencing demand at the point of sale, the factors including the current number of people queuing at the point of sale and the volume of merchandise sold;

[0007] When the current number of people in the queue is greater than a preset queue number threshold, or the commodity sales volume is greater than a preset commodity sales volume, determining that a demand change occurs at the sales point;

[0008] A first adjustment duration of the corresponding task is determined according to a preset queuing model and the influencing factor, and the time window of the task is adjusted according to the first adjustment duration.

[0009] In one embodiment, the step of determining the first adjusted duration of the corresponding task according to the preset queuing model and the influencing factor includes:

[0010] Get the customer arrival rate based on the average number of customers in the preset time period;

[0011] Obtaining an initial service rate based on historical data, and adjusting the initial service rate based on a product attribute adjustment coefficient and an environmental parameter adjustment coefficient to obtain a target service rate;

[0012] Inputting the target service rate and the customer arrival rate into the preset queuing model to obtain an average waiting time;

[0013] The first adjustment duration is determined according to the average waiting time and the current number of people in queue.

[0014] In one embodiment, after the steps of determining a first adjusted duration of the corresponding task according to the preset queuing model and the influencing factor, and adjusting the time window of the task according to the first adjusted duration, the method further includes:

[0015] If the adjustment range of the task time window is greater than a preset adjustment range threshold, the task chain is adjusted according to a preset reinforcement learning model.

[0016] In one embodiment, the step of adjusting the task chain according to the preset reinforcement learning model includes:

[0017] Acquire sales point location information, robot resource information, and task information according to the task chain;

[0018] Inputting the sales point location information, the robot resource information, and the task information into the reinforcement learning model, and outputting a probability distribution of an adjustment action;

[0019] The adjustment action with the highest probability is determined as the target adjustment action, and the task chain is updated according to the target adjustment action.

[0020] In one embodiment, after the step of adjusting the task chain according to a preset reinforcement learning model if the adjustment amplitude of the task time window is greater than a preset adjustment amplitude threshold, the method further includes:

[0021] When the sales difference between the actual sales volume and the predicted sales volume of the target task is greater than a preset sales difference threshold, a corresponding second adjustment period is determined according to the sales difference, and the time window of the target task is adjusted according to the second adjustment period;

[0022] A loss function is defined according to the sales volume difference, an adjustment amount of the feature weight is determined based on a preset optimization algorithm and the loss function, and the feature weight of the preset reinforcement learning model is adjusted according to the adjustment amount.

[0023] In one embodiment, before the step of obtaining the influencing factor of the point-of-sale demand based on the sensor provided at the point-of-sale, the method further includes:

[0024] Obtain the historical stay time of the historical tasks corresponding to the selling point, the task start time of the task to be generated, and the type of the sold commodity;

[0025] The historical stay time, the crowd density characteristics of the sales point, and the storage characteristics of the sold goods are input into a preset prediction model, the historical stay time is adjusted to obtain the task stay time, and the time window is generated according to the task stay time.

[0026] In one embodiment, before the step of obtaining the influencing factor of the point-of-sale demand based on the sensor provided at the point-of-sale, the method further includes:

[0027] Obtaining a first distance between the position of the idle robot and the position of the sales point to be assigned the task, and the number of tasks already assigned to the idle robot;

[0028] Determine a comprehensive score based on the first distance, the number of scheduled tasks, and corresponding weight coefficients;

[0029] A first target robot is determined according to the comprehensive score, and the task to be assigned is assigned to the first target robot for execution.

[0030] In one embodiment, the robot sales task processing method further includes:

[0031] When it is detected that a scheduled task has a time conflict and / or a sales point conflict, obtaining the task type and / or task area of the conflicting task;

[0032] Merging the conflicting tasks of the same task type and / or the same task area;

[0033] Assign the merged conflicting tasks to the second target robot for execution.

[0034] In addition, to achieve the above-mentioned purpose, the present application also proposes a robot vending task processing device, which includes: a memory, a processor, and a computer program stored on the memory and runnable on the processor, and the computer program is configured to implement the steps of the robot vending task processing method described above.

[0035] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium. A computer program is stored on the storage medium, and when the computer program is executed by the processor, the steps of the robot sales task processing method described above are implemented.

[0036] This application proposes a method for processing robot vending tasks. Sensors installed at vending points detect the current queue length and product sales volume. When the current queue length exceeds a preset queue threshold, or product sales volume exceeds a preset product sales volume, the task duration for the corresponding vending task at that point is adjusted based on the queuing model. By dynamically adjusting the task duration of robot tasks, it is possible to reduce idle or overloaded robot instances, avoid task interruptions or service shortages caused by changing demand, and thus improve the robot's service efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0038] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0039] Figure 1 A flowchart of the first embodiment of the robot sales task processing method of this application is provided;

[0040] Figure 2 A flowchart of the second embodiment of the robot sales task processing method of this application is provided;

[0041] Figure 3 A flowchart of the third embodiment of the robot sales task processing method of this application is provided;

[0042] Figure 4 A flowchart illustrating a fourth embodiment of the robot sales task processing method of this application;

[0043] Figure 5 This is a schematic diagram of the device structure of the hardware operating environment involved in the robot vending task processing method in the embodiment of the present application.

[0044] The purpose, features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0045] It should be understood that the specific embodiments described herein are merely used to explain the technical solutions of the present application and are not intended to limit the present application.

[0046] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0047] The main solution of the embodiment of the present application is: based on sensors installed at the sales point, an influencing factor of the sales point demand is obtained, and the influencing factor includes the current number of people queuing at the sales point and the sales volume of goods; when the current number of people queuing is greater than a preset queue number threshold, or the sales volume of goods is greater than a preset sales volume of goods, it is determined that a demand change has occurred at the sales point; according to the preset queuing model and the influencing factor, a first adjustment duration of the corresponding task is determined, and the time window of the task is adjusted according to the first adjustment duration.

[0048] Robotic sales tasks refer to robots performing various sales-related tasks in different locations. These tasks may include direct product sales, such as processing customer purchase requests, or auxiliary tasks such as replenishment, inventory management, and equipment maintenance.

[0049] When assigning sales tasks to robots, the scheduling system typically generates a service time table based on historical sales data, allocating longer times to high-traffic areas and shorter times to low-traffic areas. This allocation method assumes that demand patterns are highly cyclical, but in real-world scenarios, unexpected events can disrupt historical patterns. When a sales point experiences a sudden change in demand, such as an increase in queues or a hot-selling product, the sales task is still executed according to the originally allocated service time. After the specified service time, the robot is forced to leave, resulting in longer waiting times for users. When certain sales points are empty for extended periods of time, this results in a waste of robot resources.

[0050] This application provides a solution that uses sensors installed at sales points to monitor the current queue length and product sales volume. When the current queue length exceeds a preset queue threshold, or product sales exceed a preset product sales volume, the task duration for the sales task corresponding to that point is adjusted based on the queuing model. By dynamically adjusting the task duration for robots to perform tasks, it is possible to reduce idle or overloaded robot situations, avoid task interruptions or service shortages caused by changes in demand, and improve the robot's service efficiency.

[0051] It should be noted that the execution subject of this embodiment can be a computing service device with network communication and program execution capabilities, such as a tablet computer, personal computer, mobile phone, server, server cluster, etc., or an electronic device or device capable of implementing the above functions. The following uses a robot vending task processing device as an example to illustrate this embodiment and the following embodiments.

[0052] Based on this, the embodiment of the present application provides a robot sales task processing method, referring to Figure 1 , Figure 1 This is a flowchart of the first embodiment of the robot sales task processing method of this application.

[0053] In this embodiment, the robot sales task processing method includes steps S10 to S30:

[0054] Step S10 : obtaining an influencing factor of sales point demand based on sensors installed at the sales point, wherein the influencing factor includes the current number of people queuing at the sales point and the sales volume of goods.

[0055] In this embodiment, factors influencing point-of-sale demand include: the current number of people in line, product sales volume, shelf inventory, and vending machine temperature. Points of sale are equipped with sensors to collect real-time data such as customer flow, inventory, and equipment status (e.g., number of people in line, shelf inventory, and machine temperature). During the robot's vending process, if it detects changes in demand at a point, such as long queues or hot sales of a particular product, the robot's dwell time at that point will be adjusted.

[0056] Optionally, the number of people in the queue can be monitored and recorded in real time by devices such as cameras or infrared sensors to reflect real-time changes in customer flow.

[0057] Optionally, for inventory data, RFID (Radio Frequency Identification) technology or weight sensors are used to monitor the number of goods on the shelves or weight changes in real time. For example, RFID technology is used to read the tag information of the remaining goods on the shelves in real time and compare it with the initial inventory data to calculate the remaining number of goods. The RFID tag is affixed to the product packaging, and a reader / writer matching the RFID tag is installed above or on the side of the shelf. The reader / writer reads the tag information in real time and obtains a list of identifiers of the currently scanned goods. The ID list read in real time is compared with the initial inventory list in the database. The current inventory quantity = initial inventory quantity - the number of IDs that have left the shelf + the number of newly added IDs.

[0058] Optionally, for device status data, sensors can be used to monitor the device's operating status, such as machine temperature and switch status, to promptly detect and address device failures and ensure normal operation. For example, a temperature sensor can collect device temperature data in real time, filter the collected data, store it in a database, and compare the current temperature with a preset threshold.

[0059] Step S20: When the current number of people in the queue is greater than a preset threshold number of people in the queue, or the sales volume of the product is greater than a preset sales volume of the product, it is determined that a demand change occurs at the sales point.

[0060] In this embodiment, preset rules for adjusting the robot's dwell time are designed based on the type and extent of demand changes. Sensor data is compared with preset thresholds to determine whether the thresholds have been exceeded. For example, if the number of queues exceeds a certain threshold, the robot's dwell time is increased; if the shelf inventory is low, the robot's dwell time is reduced. When sensor data exceeds the preset threshold, demand at the point of sale is considered to have changed, triggering the robot's dwell time adjustment process.

[0061] In another optional embodiment, a sliding window algorithm is used to dynamically update the data set within a fixed time window and calculate statistical indicators within the window (such as mean, variance, trend slope, etc.). For example, the x-coordinate (time) and y-coordinate (quantity of goods) of the data points within the window are calculated, and the slope is calculated using the least squares formula. The positive or negative sign of the slope indicates the trend direction, and the absolute value indicates the rate of change. If the trend slope of the current window remains negative for more than a certain period of time (for example, three consecutive windows) and the absolute value exceeds a set threshold, it is considered that the quantity of goods is decreasing rapidly and that a demand change has occurred at the point of sale.

[0062] Step S30 : determining a first adjustment duration of the corresponding task according to a preset queuing model and the influencing factor, and adjusting the time window of the task according to the first adjustment duration.

[0063] In this implementation, when demand changes are detected, a queuing theory model is used to determine the required delay for each additional person in the queue, based on the real-time queue count and average service time. This delay is also adjusted based on product attributes (e.g., reserving time for cabinet opening and closing for refrigerated goods) and environmental parameters (e.g., extending the delay by 10% on rainy or snowy days). Finally, the mission field of view for the point of sale is adjusted based on the adjusted delay.

[0064] In a feasible implementation, step S30 may include steps S31 to S34:

[0065] Step S31: obtaining a customer arrival rate based on the average number of customers in a preset time period.

[0066] In this implementation, the M / M / 1 queuing model (Poisson arrival / exponential service / single channel) is used as the basic framework to monitor the arrival rate λ, service rate μ, and queue length Q in real time. The arrival rate λ represents the number of customers arriving per unit time. The service rate μ is the number of people served per unit time, calculated by combining product attributes (such as the handling time of refrigerated goods) and environmental parameters (extended handling time in rainy or snowy weather). The queue length Q represents the number of people currently in line. Using cameras or sensors, the total number of customers arriving at the point of sale within a preset time period is counted in real time. The customer arrival rate is calculated by dividing the total number of customers by the length of the preset time period.

[0067] Step S32: obtaining an initial service rate based on historical data, and adjusting the initial service rate based on a product attribute adjustment coefficient and an environmental parameter adjustment coefficient to obtain a target service rate.

[0068] In this implementation, relevant information about the robot's sales tasks, including service start and end times, and the number of customers served, is collected through the robot's operating system logs and point-of-sale surveillance videos. This data is filtered to identify data from normal sales conditions, excluding abnormal sales data caused by equipment failures, special promotions, and emergencies. Time is divided into appropriate time units, and the average number of customers served per hour by the robot within each time unit is calculated as the initial service rate.

[0069] The initial service rate is adjusted based on factors such as product attributes, environmental parameters, and robot status. Product attributes can include storage conditions, size and weight, and whether packaging is required. For refrigerated items, time should be reserved for cabinet opening and closing, and a larger adjustment factor should be set. Environmental parameters can include weather conditions and the sales point environment. A larger adjustment factor should be set on rainy or snowy days. The initial service rate is multiplied by the product attribute adjustment factor and the environmental parameter adjustment factor to obtain the target service rate. By adjusting the target service rate based on different product and environmental characteristics, the robot can better adapt to various sales scenarios, respond to customer needs more quickly, reduce customer wait times, and thus improve customer satisfaction.

[0070] Step S33: input the target service rate and the customer arrival rate into the preset queuing model to obtain the average waiting time.

[0071] Step S34: determining the first adjustment duration according to the average waiting time and the current number of people in queue.

[0072] In this embodiment, the target service rate and customer arrival rate are input into the preset queuing model. In the M / M / 1 queuing model, the average waiting time W = λ / μ (μ-λ). The first adjustment time is determined based on the average waiting time and the current number of people in the queue, and the first adjustment time = average waiting time × current number of people in the queue. If a sudden long queue is detected or a certain product is hot-selling, the previous task time window is extended according to the first adjustment time calculated above. If the robot is low on power or the sales point is about to close, the dwell time is adjusted to shorten the previous task time window. The adjusted task time window information is updated to the robot's control system and the sales point's management system.

[0073] In this embodiment, sensors installed at sales points detect the current queue length and product sales volume. If the current queue length exceeds a preset threshold, or if product sales exceed a preset threshold, the duration of the sales task corresponding to that point is adjusted based on the queuing model. By dynamically adjusting the duration of robot tasks, robot idleness and overload can be reduced, preventing task interruptions or service shortages caused by fluctuating demand and improving the robot's service efficiency.

[0074] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above embodiment 1 can be referred to the above introduction and will not be described in detail later. Figure 2 After step S30, the robot sales task processing method further includes steps A10 to A30:

[0075] Step A10: If the adjustment range of the task time window is greater than a preset adjustment range threshold, the task chain is adjusted according to a preset reinforcement learning model.

[0076] In this embodiment, when the adjustment range of the task time window exceeds a preset threshold, it indicates that the operating conditions of the sales point have changed significantly. The reinforcement learning model is triggered to adjust the task chain, for example, by bringing forward or merging adjacent tasks that are less affected. Failure to adjust the task chain in a timely manner may result in a waste of robot resources (e.g., too few tasks, idle robots) or an inability to meet customer needs (e.g., too many tasks, overloaded robots). Through adjustments made by the reinforcement learning model, task requirements and robot resources can be accurately matched, improving resource utilization.

[0077] In a feasible implementation, step A10 may include steps A11 to A13:

[0078] Step A11: Acquire sales point location information, robot resource information, and task information according to the task chain.

[0079] In this embodiment, the location information of the sales point, including latitude and longitude coordinates and the region in which it is located, is obtained from the sales point's geographic information system database or related positioning services. This location information is then preprocessed through coordinate conversion and region coding. Robot resource information, including the robot's battery level and current status (idle, busy, etc.), is obtained from the robot management system. This resource information is quantified and encoded, for example, by converting the battery level into a percentage. Task information, including task type (e.g., product sales, replenishment, etc.), priority, estimated completion time, and associated products or services, is obtained from the task management system. Feature extraction and encoding are performed on this task information, for example, by representing the task type using a one-hot encoding and converting the estimated completion time into a timestamp or relative time value. Sales point location information helps the model account for the impact of geographical factors on task execution, such as customer flow distribution and traffic conditions in different regions. Robot resource information enables the model to understand the robot's actual work capacity and avoid assigning tasks that exceed its capabilities. Task information clarifies the specific requirements and priorities of tasks, ensuring that important tasks are prioritized. Different sales points, different robots, and different tasks have their own characteristics. Comprehensive consideration of the above information can make the adjustment of the task chain more personalized to meet the needs of different scenarios.

[0080] Step A12: Input the sales point location information, the robot resource information, and the task information into the reinforcement learning model, and output a probability distribution of the adjustment action.

[0081] Step A13: confirm the adjustment action with the highest probability as the target adjustment action, and update the task chain according to the target adjustment action.

[0082] In this implementation, preprocessed vending machine location information, robot resource information, and task information are integrated to construct a feature vector. Reinforcement learning models can include DQN (Deep Q-Network) and PPO (Proximal Policy Optimization). This constructed feature vector is input into the reinforcement learning model, which then calculates the input feature vector and outputs a probability distribution for adjustment actions. The action with the highest probability is selected to update the task chain. Adjustment actions can include moving adjacent tasks forward or merging them.

[0083] Optionally, the reinforcement learning model is a PPO model. The feature vector s is input into the Actor network of the PPO model to calculate the action probability distribution π(a|s). The Actor network is a deep neural network that includes an input layer, a hidden layer, and an output layer. The input layer is used to receive the feature vector s. The hidden layer has one or more fully connected layers, each of which contains an activation function (such as ReLU, Tanh, etc.). The output layer outputs a vector whose dimension matches the size of the action space m. The feature vector s is input into the Actor network f θ (s), where θ represents the parameters of the network (weights and biases), and the output vector z is obtained by forward propagation calculation, z = f θ (s), where z∈R m , m is the dimension of the action space. The output vector z is converted into a probability distribution using the Softmax function.

[0084] Step A20: When the sales difference between the actual sales volume and the predicted sales volume of the target task is greater than the preset sales difference threshold, a corresponding second adjustment duration is determined according to the sales difference, and the time window of the target task is adjusted according to the second adjustment duration.

[0085] In this embodiment, after adjusting the task chain based on the reinforcement learning model, the execution efficiency of the task area is tested. If the actual sales volume is lower than the predicted value, the task time window is adjusted to terminate the task early. The second adjustment duration can be set based on the sales volume difference interval. The sales volume difference is divided into multiple intervals, each corresponding to a second adjustment duration. When the sales volume difference is small, the adjustment range is small; when the sales volume difference is large, the adjustment range is large.

[0086] Step A30: defining a loss function according to the sales volume difference, determining an adjustment amount of feature weights based on a preset optimization algorithm and the loss function, and adjusting the feature weights of the preset reinforcement learning model according to the adjustment amount.

[0087] In this embodiment, when the actual sales volume is lower than the predicted value, the data is fed back to the reinforcement learning model to adjust the reinforcement learning model. The sales volume difference is used as the target variable, and the current feature weight of the reinforcement learning model is used. The loss function can be defined as the sum of the squares of the sales volume difference. A gradient descent method (such as stochastic gradient descent, Adam optimizer) is selected to calculate the gradient of the loss function with respect to the feature weight, and the weight is adjusted according to the gradient direction to gradually reduce the loss function. By minimizing the loss function, the optimal weight is found, and the reinforcement learning model is updated. When the actual sales volume is lower than the predicted value, future prediction errors can be reduced by learning the shortcomings of the current strategy and adjusting the parameters.

[0088] Based on the first embodiment of the present application, in the third embodiment of the present application, the same or similar contents as those in the above embodiment 1 can be referred to the above introduction and will not be described in detail later. Figure 3 , step S10 may include steps S01 to S02:

[0089] Step S01: Obtain the historical stay time of the historical tasks corresponding to the sales point, the task start time of the task to be generated, and the type of the sold commodity.

[0090] In this embodiment, all historical task records of each selling point are obtained from the historical task records, including the task type, selling point area, task start time, task end time, type of goods sold, residence time, etc. The historical residence time at the same selling point that has the same type of goods sold as the current task to be generated and has a similar task start time is filtered out. For example, in the task for which the residence time needs to be determined, the product type is selling cold drinks and the task start time is 16:00 on Saturday afternoon. The historical task records are filtered for records in which the type of goods sold is cold drinks and the task start time is between 15:00 and 17:00 on weekend afternoons. Using the filtered historical residence time data, statistical indicators such as the mean and median are calculated as a preliminary estimate of the residence time of the current task to be generated.

[0091] Step S02: Input the historical stay time, the crowd density characteristics of the sales point, and the storage characteristics of the sold goods into a preset prediction model, adjust the historical stay time to obtain the task stay time, and generate the time window according to the task stay time.

[0092] In this embodiment, when a new task is created, the task duration is automatically generated through a preset prediction model in combination with the historical stay time, the heat map of the flow of people around the sales point, and the attributes of the goods. Traffic flow characteristics such as crowd density and peak hour traffic are extracted from the surrounding crowd heat map, and the crowd flow characteristics, commodity attribute characteristics, and historical stay time characteristics are combined into a feature vector as the input of the model. The preset prediction model can select linear regression, decision tree, random forest, gradient boosting tree and other models according to the data characteristics and business needs. The feature vector is input into the trained prediction model to obtain the task stay time. By combining the historical stay time, crowd density and commodity storage characteristics, the various factors affecting the task stay time can be considered more comprehensively, thereby improving the accuracy of the prediction.

[0093] Based on the first embodiment of the present application, in the fourth embodiment of the present application, the same or similar contents as those in the above embodiment 1 can be referred to the above introduction and will not be described in detail later. Figure 4 Before step S10, the robot sales task processing method further includes steps B10 to B30:

[0094] Step B10: Obtain a first distance between the position of the idle robot and the position of the sales point to be assigned a task, and the number of tasks that have been assigned to the idle robot.

[0095] In this embodiment, when assigning tasks, a list of currently idle robots is queried from the robot management system or database. A unique identifier (e.g., robot ID) and the robot's current location are extracted for each idle robot. The task management system or database is then queried for all task points for the assigned tasks. Each task point's unique identifier, such as the task ID, is extracted, along with the corresponding vending point location information. The system then iterates through each idle robot and each task point, calculating the spatial distance between them.

[0096] Step B20: determining a comprehensive score based on the first distance, the number of scheduled tasks, and corresponding weight coefficients.

[0097] Step B30: Determine a first target robot based on the comprehensive score, and assign the task to be assigned to the first target robot for execution.

[0098] In this embodiment, weight coefficients are assigned to the first distance between the robot and the vending point to be assigned, as well as the number of tasks already assigned to the robot. Each robot's comprehensive score is determined by multiplying the first distance by the number of tasks already assigned. The comprehensive scoring function is: Score = α × Distance + β × TaskCount. α and β are weight coefficients, Distance is the first distance, and TaskCount is the number of tasks already assigned. A lower score indicates a shorter distance and a lower load. For each task to be assigned, the robot with the lowest score is selected and assigned to the first target robot.

[0099] Optionally, the weight coefficients can be determined based on business experience and experimental tests. For example, for urgent tasks, α = 0.3, β = 0.7, giving priority to load balancing. For non-urgent tasks, α = 0.7, β = 0.3, giving priority to distance factors.

[0100] Optionally, the weight coefficient is dynamically adjusted by a machine learning model (such as a regression model) by analyzing historical task allocation data. The distance between the robot and the vending point, the number of tasks currently scheduled for the robot, the task type, the robot load, the remaining power, and other data are used as input features. The optimal weight coefficient or task allocation effect (such as completion time, success rate) determined based on historical data is used as the output label. The machine learning model is trained by using the training set data to learn the mapping relationship between input features and output labels. The weight coefficient is predicted using the trained model. By analyzing historical data through a machine learning model, the optimal weight coefficient is predicted in real time, and the weight can be adjusted according to real-time data such as task type, robot load, and distance, so as to adapt to changes such as changes in task priority and robot status.

[0101] In this embodiment, during the execution process, the progress of the robot is tracked in real time and the execution status of the task is recorded. For example, whether the robot arrived at the task location on time, whether the task of each sales point was completed on time, whether the residence time met the requirements, etc. If an abnormal situation occurs, an alarm is issued and the user is notified. The user can choose to adjust the task schedule or dispatch a backup robot. When the task is completed, the completion of the task is confirmed. Check whether the task was completed as expected, including the residence time of the task, the sales point of execution, etc. If there are unfinished parts of the task, the user is reminded and corresponding handling suggestions are provided. The user can also manually confirm whether the task is completed. For completed tasks, a detailed task execution report is generated for the user to view. The report contains data such as the time when the task was completed and the efficiency of task execution.

[0102] In this embodiment, tasks are assigned based on factors such as whether each robot is currently idle, its distance from the task location, and the number of other tasks it is currently performing. By assigning tasks to the most suitable robots, we ensure that tasks can be completed on time and efficiently.

[0103] Based on the first embodiment of the present application, in the fifth embodiment of the present application, the same or similar contents as those in the first embodiment can be referred to above and will not be described in detail. On this basis, the robot vending task processing method can also include steps C10 to C30:

[0104] Step C10 : When it is detected that there is a time conflict and / or a sales point conflict among the scheduled tasks, the task type and / or task area of the conflicting task is obtained.

[0105] In this embodiment, after a task is created, all scheduled tasks are checked to ensure that the new task does not conflict with existing tasks. Conflict detection includes time conflicts and sales point conflicts. If the time of multiple tasks overlaps, or if multiple tasks need to be scheduled simultaneously for the same sales point, the task is marked as a conflicting task. The schedule of all tasks is obtained, and the nearest idle time period is found for low-priority tasks, with high-priority tasks being prioritized. For low-priority tasks, their start time is adjusted to postpone their execution, or if other robots are available, the task is reassigned to other robots. At the same time, it is determined whether multiple similar tasks can be merged to reduce repetitive work and thus improve task execution efficiency.

[0106] When checking whether there is a time conflict and / or sales point conflict for a scheduled task, obtain detailed information about the scheduled task, including: task ID, task type (such as handling, maintenance, inspection, etc.), sales point area, and task execution time window including the time when the task is scheduled to start and the time when the task is scheduled to complete. Traverse the task pairs and compare their time windows for each pair of tasks. When the end time of task 1 is earlier than the start time of task 2, and the end time of task 2 is earlier than the start time of task 1, mark the corresponding task pair as having a time conflict. When the task area of task 1 is the same as the task area of task 2, mark it as a sales point conflict.

[0107] Step C20: merging the conflicting tasks with the same task type and / or the same task area.

[0108] Step C30: assigning the merged conflicting tasks to the second target robot for execution.

[0109] In this embodiment, conflicting tasks are grouped based on their conflict type, task type, and task area. Tasks with overlapping time windows and the same task type are grouped together, while those with conflicting sales points and the same task area are grouped together. The grouped conflicting tasks are then merged to create a new merged task.

[0110] Alternatively, a genetic algorithm can be used to find the optimal task combination that minimizes the total number of conflicting tasks. In a genetic algorithm, each chromosome represents a task allocation scheme, and each gene in the chromosome represents a task. For example, chromosome = [A, B, C] indicates that tasks A, B, and C are assigned to the same group and may be merged. Multiple initial chromosomes are randomly generated to form an initial population. Task pairs within the chromosomes are examined for time conflicts and sales point conflicts. The number of conflicting task pairs is accumulated to obtain the total number of conflicts for the task combination, which is also the fitness of each chromosome. Using methods such as roulette wheel selection and tournament selection, chromosomes are selected based on fitness, and the selected chromosomes are crossovered to generate new chromosomes. For example, if parent 1 = [A, B] and parent 2 = [C, D], after a random crossover point is selected, the offspring will be [A, D]. The new chromosome is then mutated. For example, if chromosome = [A, B] randomly replaces one gene and mutates to [A, C]. The selection, crossover, and mutation operations are repeated until the maximum number of iterations is reached or the fitness value converges (i.e., the total number of conflicts no longer decreases significantly). The chromosome with the highest fitness is selected from the final population and decoded to obtain the optimal task combination solution. The merged task is generated based on the optimal solution.

[0111] In this embodiment, selecting the idle robot closest to the task to be assigned significantly reduces robot movement time. Selecting robots with fewer assigned tasks balances the robot load, preventing some robots from being overloaded while others remain idle. Dynamically adjusting the task allocation strategy based on the robots' real-time status and location ensures the rationality of task allocation.

[0112] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the robot sales task processing method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0113] The present application provides a robot vending task processing device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the robot vending task processing method in the above-mentioned embodiment one.

[0114] Reference below Figure 5, which shows a schematic diagram of the structure of a robotic vending task processing device suitable for implementing the embodiments of the present application. The robotic vending task processing device in the embodiments of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, personal digital assistants (PDAs), tablet computers (Portable Android Devices), etc., as well as fixed terminals such as digital TVs and desktop computers. Figure 5 The robot vending task processing device shown is merely an example and should not impose any limitations on the functions and scope of use of the embodiments of the present application.

[0115] like Figure 5 As shown, the robot vending task processing device may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 1002 or the program loaded from the storage device 1003 to the random access memory (RAM) 1004. In the random access memory 1004, various programs and data required for the operation of the robot vending task processing device are also stored. The processing device 1001, the read-only memory 1002, and the random access memory 1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, hard disk, etc.; and communication devices 1009. The communication device 1009 can allow the robotic vending task processing device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a robotic vending task processing device with various systems, it should be understood that it is not required to implement or have all of the systems shown. More or fewer systems may be implemented or have alternatively.

[0116] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device 1003, or installed from a read-only memory 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are performed.

[0117] The robotic vending task processing device provided in this application, employing the robotic vending task processing method described in the aforementioned embodiment, can address the technical problem of improving the service efficiency of robots performing vending tasks. Compared to the prior art, the beneficial effects of the robotic vending task processing device provided in this application are the same as those of the robotic vending task processing method described in the aforementioned embodiment. Other technical features of the robotic vending task processing device are the same as those disclosed in the aforementioned embodiment and are not further elaborated here.

[0118] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0119] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0120] The present application provides a computer-readable storage medium having computer-readable program instructions (i.e., computer program) stored thereon, and the computer-readable program instructions are used to execute the robot vending task processing method in the above-mentioned embodiment.

[0121] The computer-readable storage medium provided in the present application can be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM, Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM, CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present embodiment, the computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in combination with an instruction execution system or device. The program code contained on the computer-readable storage medium can be transmitted using any appropriate medium, including but not limited to: an electric wire, an optical cable, a radio frequency (RF, Radio Frequency), etc., or any suitable combination thereof.

[0122] The above-mentioned computer-readable storage medium may be included in the robot vending task processing device; or it may exist independently without being assembled into the robot vending task processing device.

[0123] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the robot vending task processing device, the robot vending task processing device: obtains the influencing factors of the demand at the vending point based on the sensors set at the vending point, and the influencing factors include the current number of people queuing and the sales volume of goods at the vending point; determines that a demand change has occurred at the vending point when the current number of people queuing is greater than a preset threshold number of people queuing, or the sales volume of goods is greater than a preset sales volume of goods; determines a first adjustment duration of the corresponding task according to the preset queuing model and the influencing factors, and adjusts the time window of the task according to the first adjustment duration.

[0124] Computer program code for performing the operations of the present application can be written in one or more programming languages, or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0125] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0126] The modules described in the embodiments of the present application may be implemented in software or hardware, wherein the name of a module does not necessarily limit the unit itself.

[0127] The computer-readable storage medium provided in this application stores computer-readable program instructions (i.e., a computer program) for executing the aforementioned robot vending task processing method. This computer-readable storage medium can address the technical problem of improving the service efficiency of robots performing vending tasks. Compared to the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the robot vending task processing method provided in the aforementioned embodiment, and are not further elaborated here.

[0128] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned robot vending task processing method when executed by a processor.

[0129] The computer program product provided in this application can solve the technical problem of how to improve the service efficiency of robots performing sales tasks. Compared with the existing technology, the beneficial effects of the computer program product provided in this application are the same as the beneficial effects of the robot sales task processing method provided in the above embodiment, and will not be repeated here.

[0130] The above description is only part of the embodiments of the present application and does not limit the patent scope of the present application. All equivalent structural transformations made by using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A robot vending task processing method, characterized in that: The robot sales task processing method includes: Obtaining, based on sensors installed at the point of sale, factors influencing demand at the point of sale, the factors including the current number of people queuing at the point of sale and the volume of merchandise sold; When the current number of people in the queue is greater than a preset queue number threshold, or the commodity sales volume is greater than a preset commodity sales volume, determining that a demand change occurs at the sales point; A first adjustment duration of the corresponding task is determined according to a preset queuing model and the influencing factor, and the time window of the task is adjusted according to the first adjustment duration.

2. The robot vending task processing method according to claim 1, characterized in that: The step of determining the first adjusted duration of the corresponding task according to the preset queuing model and the influencing factor includes: Get the customer arrival rate based on the average number of customers in the preset time period; Obtaining an initial service rate based on historical data, and adjusting the initial service rate based on a product attribute adjustment coefficient and an environmental parameter adjustment coefficient to obtain a target service rate; Inputting the target service rate and the customer arrival rate into the preset queuing model to obtain an average waiting time; The first adjustment duration is determined according to the average waiting time and the current number of people in queue.

3. The robot vending task processing method according to claim 1, wherein: After the step of determining a first adjusted duration of the corresponding task according to the preset queuing model and the influencing factor, and adjusting the time window of the task according to the first adjusted duration, the step further includes: If the adjustment range of the task time window is greater than a preset adjustment range threshold, the task chain is adjusted according to a preset reinforcement learning model.

4. The robot vending task processing method according to claim 3, wherein: The step of adjusting the task chain according to the preset reinforcement learning model includes: Acquire sales point location information, robot resource information, and task information according to the task chain; Inputting the sales point location information, the robot resource information, and the task information into the reinforcement learning model, and outputting a probability distribution of an adjustment action; The adjustment action with the highest probability is determined as the target adjustment action, and the task chain is updated according to the target adjustment action.

5. The robot vending task processing method according to claim 3, wherein: After the step of adjusting the task chain according to a preset reinforcement learning model if the adjustment range of the task time window is greater than a preset adjustment range threshold, the method further includes: When the sales difference between the actual sales volume and the predicted sales volume of the target task is greater than a preset sales difference threshold, a corresponding second adjustment period is determined according to the sales difference, and the time window of the target task is adjusted according to the second adjustment period; A loss function is defined according to the sales volume difference, an adjustment amount of the feature weight is determined based on a preset optimization algorithm and the loss function, and the feature weight of the preset reinforcement learning model is adjusted according to the adjustment amount.

6. The robot vending task processing method according to claim 1, wherein: Before the step of obtaining the influencing factor of the point-of-sale demand based on the sensor provided at the point-of-sale, the method further includes: Obtain the historical stay time of the historical tasks corresponding to the selling point, the task start time of the task to be generated, and the type of the sold commodity; The historical stay time, the crowd density characteristics of the sales point, and the storage characteristics of the sold goods are input into a preset prediction model, the historical stay time is adjusted to obtain the task stay time, and the time window is generated according to the task stay time.

7. The robot vending task processing method according to claim 1, wherein: Before the step of obtaining the influencing factor of the point-of-sale demand based on the sensor provided at the point-of-sale, the method further includes: Obtaining a first distance between the position of the idle robot and the position of the sales point to be assigned the task, and the number of tasks already assigned to the idle robot; Determine a comprehensive score based on the first distance, the number of scheduled tasks, and corresponding weight coefficients; A first target robot is determined according to the comprehensive score, and the task to be assigned is assigned to the first target robot for execution.

8. The robot vending task processing method according to claim 1, wherein: The robot sales task processing method further includes: When it is detected that a scheduled task has a time conflict and / or a sales point conflict, obtaining the task type and / or task area of the conflicting task; Merging the conflicting tasks of the same task type and / or the same task area; Assign the merged conflicting tasks to the second target robot for execution.

9. A robot vending task processing device, characterized in that: The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the robot vending task processing method according to any one of claims 1 to 8.

10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the robot vending task processing method according to any one of claims 1 to 8 are implemented.

Images