WO2022052543A1 - 一种递送机器人的云端调度方法、装置和服务器 - Google Patents

一种递送机器人的云端调度方法、装置和服务器 Download PDF

Info

Publication number
WO2022052543A1
WO2022052543A1 PCT/CN2021/100231 CN2021100231W WO2022052543A1 WO 2022052543 A1 WO2022052543 A1 WO 2022052543A1 CN 2021100231 W CN2021100231 W CN 2021100231W WO 2022052543 A1 WO2022052543 A1 WO 2022052543A1
Authority
WO
WIPO (PCT)
Prior art keywords
waybill
package
optimized
packages
delivery
Prior art date
Application number
PCT/CN2021/100231
Other languages
English (en)
French (fr)
Inventor
王超
Original Assignee
上海有个机器人有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 上海有个机器人有限公司 filed Critical 上海有个机器人有限公司
Publication of WO2022052543A1 publication Critical patent/WO2022052543A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06316—Sequencing of tasks or work
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B25—HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25J—MANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J9/00—Programme-controlled manipulators
    • B25J9/16—Programme controls
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/083—Shipping
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network

Definitions

  • the present invention relates to the field of robots, in particular to a cloud scheduling method, device and server for delivering robots.
  • the current building delivery robots have warehousing capacity and the ability to move within the building, and can undertake the delivery tasks of food delivery and express delivery within the building.
  • the number of takeaways and express delivery in buildings is increasing.
  • building delivery robots are also facing challenges in service efficiency and scale.
  • the demand for delivery increases , which means that the number of robots in demand will increase, and secondly, the increase in distribution demand, especially in the centralized distribution situation (such as takeaway afternoon peak hours), will cause the risk of prolonging the delivery time, which will lead to a decline in service experience. Therefore, how to improve the overall carrying capacity of the robot cluster, and how to ensure the delivery time of the robot while the distribution scale increases, is of great significance to the cost optimization and efficiency improvement of the delivery robot.
  • the present invention provides a cloud scheduling method, device and server for a delivery robot, and solves the technical problem of how to schedule the waybill and the transport capacity of the robot, thereby improving the delivery efficiency of the robot.
  • a cloud scheduling method for a delivery robot comprising the following steps:
  • Step 1 Aggregate the waybills in the waybill pool according to the similarity to form a set of waybill packages including at least one optimized waybill package;
  • Step 2 scheduling and sorting all optimized waybill packages in the set of waybill packages
  • Step 3 according to the sorting result of the optimized waybill package, sequentially assign the optimized waybill package to the delivery robot.
  • the aggregation of the waybills in the waybill pool according to the similarity to form a set of waybill packages including at least one optimized waybill package specifically includes the following steps:
  • S103 select two target waybill packages with the smallest similarity and the similarity less than a preset threshold in the similarity list, and calculate the total number of waybills after the two target waybill packages are combined, if the total number of waybills is less than or equal to If the delivery capacity is preset, the two target waybill packages are merged into one optimized waybill package, and the similarity list is updated;
  • S104 Repeat S103 until the similarity of any two waybill packages in the similarity list is greater than or equal to the preset threshold or the total number of waybills after the combination of the two target waybill packages is greater than the preset delivery capacity, and the waybill
  • the package merging process ends, and a set of waybill packages is formed, and the set of waybill packages includes at least one optimized waybill package.
  • the preset similarity formula is:
  • S(A, B) is the similarity between the waybill package A and the waybill package B
  • F is the sum of the number of stairs climbed by the robot after the waybill package A and the waybill package B are combined
  • w is the waybill package A and waybill package B.
  • the sum of the numbers, d0 is the total moving distance of the delivery robot on the same floor after the waybill package A and the waybill package B are combined
  • ⁇ is the weight coefficient.
  • the scheduling and sorting of all optimized waybill packages in the waybill package set specifically includes the following steps:
  • S202 define a binary group Z(t*,n*) used to represent the score of the optimized waybill package.
  • the t * t_max-t, otherwise the t* value is 0;
  • n* is the value of n; otherwise, the value of n* is 0, so
  • the t_max is the preset maximum remaining delivery time;
  • it also includes an active order pressing step, specifically: obtaining a target optimized waybill package whose binary group Z is 0, setting a time field for each target waybill in the target optimized waybill package, and assigning all The target waybill is returned to the waybill pool, and the time field is the time point when the target waybill first participated in the combined order.
  • it also includes the step of forcibly outputting the waybill, specifically: obtaining the time field of each target waybill, and calculating the order pressing time corresponding to the target waybill according to the current time, when the order pressing time is greater than the preset time
  • the optimized waybill package containing the target waybill is preferentially output.
  • the optimized waybill package is sequentially allocated to the delivery robot, which specifically includes the following steps:
  • S301 Obtain a list of candidate robots, where the robots on the candidate robot list have the following characteristics: the number of real-time waybills of the robot is less than the preset delivery capacity, and the remaining delivery time of any real-time waybill of the robot is greater than the preset delivery time Minimum remaining delivery time;
  • S302 output the optimized waybill package according to the sorting result, and obtain a robot that carries at least one real-time waybill package in the candidate robot list, calculate the similarity between the optimized waybill package and all real-time waybill packages in turn, and obtain the similarity At least one target robot that satisfies the preset merging conditions, and assigns the optimized waybill package to the optimal target robot according to the preset tracking principle;
  • a second aspect of the embodiments of the present invention provides a cloud scheduling device for a delivery robot, including an aggregation module, a sorting module, and an allocation module,
  • the aggregation module is configured to aggregate the waybills in the waybill pool according to the similarity to form a set of waybill packages including at least one optimized waybill package;
  • the sorting module is used for scheduling and sorting all the optimized waybill packages in the set of waybill packages;
  • the assigning module is configured to sequentially assign the optimized waybill packages to the delivery robot according to the sorting result of the optimized waybill packages.
  • the aggregation module specifically includes:
  • the first calculation unit is used to calculate the similarity between any two waybill packages by using a preset similarity formula, and establish a similarity list. Low;
  • a merging unit configured to select two target waybill packages with the smallest similarity and less than a preset threshold in the similarity list, and calculate the total number of waybills after the two target waybill packages are merged, if the total number of waybills is less than or equal to the preset delivery capacity, merge the two target waybill packages into one optimized waybill package, and update the similarity list;
  • a set generating unit is used to repeatedly drive the merging unit until the similarity of any two waybill packages in the similarity list is greater than or equal to the preset threshold or the combined total number of waybills of the two target waybill packages is greater than or equal to
  • the process of merging the waybill packages ends, and a set of waybill packages is formed, and the set of waybill packages includes at least one optimized waybill package.
  • the preset similarity formula is:
  • S(A, B) is the similarity between the waybill package A and the waybill package B
  • F is the sum of the number of stairs climbed by the robot after the waybill package A and the waybill package B are combined
  • w is the waybill package A and waybill package B.
  • the sum of the numbers, d0 is the total moving distance of the delivery robot on the same floor after the waybill package A and the waybill package B are combined
  • ⁇ is the weight coefficient.
  • the sorting module specifically includes:
  • a first obtaining unit configured to obtain the remaining delivery time t of each waybill in the optimized waybill package and the number of waybill n of each optimized waybill package;
  • the sorting unit is used to schedule and sort all the optimized waybill packages according to the size of the binary group Z.
  • the cloud dispatching device of the delivery robot further includes an order pressing module, and the order pressing module is used to obtain a target optimized waybill package with the binary group Z value of 0, which is the target optimized waybill package.
  • a time field is set for each target waybill, and the target waybill is returned to the waybill pool, and the time field is the time point when the target waybill first participated in the merger.
  • the cloud scheduling device of the delivery robot further includes a forced output module, and the forced output module is configured to obtain the time field of each target waybill, and calculate the press order corresponding to the target waybill according to the current time time, when the order pressing time is greater than the preset order pressing time threshold, the optimized waybill package containing the target waybill is preferentially output.
  • the distribution module specifically includes:
  • the second obtaining unit is configured to obtain a list of candidate robots, and the robots on the list of candidate robots have the following characteristics: the number of real-time waybills of the robot is less than the preset delivery capacity, and the remaining delivery of any real-time waybill of the robot is The duration is greater than the preset minimum remaining delivery duration;
  • the tracking unit is used to output the optimized waybill package according to the sorting result, and obtain the robot that carries at least one real-time waybill package in the list of candidate robots, and sequentially calculate the similarity between the optimized waybill package and all real-time waybill packages , obtain at least one target robot whose similarity satisfies the preset merging condition, and allocate the optimized waybill package to the optimal target robot according to the preset order tracking principle;
  • the allocation unit is used for allocating the optimized waybill package to any idle robot when the similarity between the real-time waybill package and the optimized waybill package of all robots in the list of alternative robots does not meet the preset merging condition, so
  • the idle robot is a robot that does not currently carry any waybills.
  • a third aspect of the embodiments of the present invention provides a server, including a memory, a processor, and a computer program stored in the memory and executable on the processor, which is implemented when the processor executes the computer program The steps of the cloud scheduling method for the delivery robot described above.
  • a fourth aspect of the embodiments of the present invention provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, implements the above-mentioned cloud scheduling method for a delivery robot step.
  • the present invention provides a cloud scheduling method, device and server for delivery robots, which not only perform multi-delivery orders, consolidate orders, and sort order packages for multiple delivery orders from the perspective of dispatching orders, but also screen and sort multiple robots from the perspective of transportation capacity scheduling.
  • Pursue and divide orders, and chase orders first and then divide orders, so as to maximize the results of waybill aggregation, increase the number of orders backed by robots, reduce the operating cost of robot distribution in practice, and improve distribution efficiency and service experience.
  • Embodiment 1 is a schematic flowchart of a cloud scheduling method for a delivery robot provided in Embodiment 1;
  • Fig. 2 is the similarity calculation schematic diagram of two waybill packages in embodiment 1;
  • Embodiment 3 is a schematic structural diagram of a cloud scheduling device for a delivery robot provided in Embodiment 2;
  • FIG. 4 is a schematic structural diagram of a server provided in Embodiment 3.
  • FIG. 4 is a schematic structural diagram of a server provided in Embodiment 3.
  • the cloud scheduling method of the present invention first assumes that there is no individual difference between delivery robots, that is, the maximum number of waybills that can be carried by all delivery robots is the same. At the same time, the present invention adopts the method of cloud server assignment to assign delivery tasks to the robot, rather than the method of autonomously grabbing orders by the robot. The reason is that the cloud server perspective has the capacity of all robots and the situation of all waybills. Compared with the local perspective of the robot, it is easier to achieve the global optimal allocation result.
  • the cloud scheduling method of the present invention is compatible with the situation of sending orders and taking orders.
  • a delivery mode that only includes delivery orders means that all robots start from a fixed delivery starting point, and then return to this fixed delivery starting point after all orders are delivered to accept the next task assignment.
  • order picking the robot can accept a delivery task with a delivery starting point, and the robot can only deliver after going to the delivery starting point to pick up the order.
  • FIG. 1 is a schematic flowchart of a cloud scheduling method for a delivery robot according to Embodiment 1 of the present invention. As shown in FIG. 1 , the method includes the following steps:
  • Step 1 Aggregate the waybills in the waybill pool according to the similarity to form a set of waybill packages including at least one optimized waybill package, which is the waybill scheduling process.
  • the waybill is the abbreviation of the delivery task, and a waybill contains the starting point, the destination point, the contact information of the delivery recipient, and so on.
  • the overall delivery time of the waybill is also called the time limit of the waybill, and the goal of the present invention is that the lower the time limit of the waybill, the better.
  • the step 1 specifically includes the following steps:
  • S101 Obtain a list of waybills corresponding to the waybill pool, and create a corresponding waybill package for each waybill in the list of waybills.
  • the calculation method of the similarity is defined as: the sum of the times the robot needs to climb the stairs after the two waybill packages are merged is higher than the total number of orders, that is, the average delivery difficulty of the orders in the combined waybill package, and the corresponding preset similarity
  • the sum, w is the sum of the waybill numbers of waybill package A and waybill package B.
  • a waybill is represented by a vertical arrow.
  • a waybill from 1F to 8F is represented as an arrow starting from line 1F and ending on line 8F, so that the above preset similarities can be used.
  • the degree formula calculates the similarity between any two waybill packages in Figure 2.
  • the above similarity definition method only considers the situation of different floors, but the actual test data shows that the time consumed by the robot to go up and down the floor is the highest proportion of the entire delivery time of the waybill, because the robot needs to take the elevator to go up and down the floor, and there is an elevator during the period. , Entering and exiting the elevator, taking the elevator and other steps.
  • the factor of the same floor can still be further refined, that is, the definition of similarity can be further expanded, and the distance factor within the same floor can be added.
  • the extended definition of similarity is as follows: the sum of the distances the robot needs to move after the two waybill packages are merged is higher than the total odds.
  • the shortest distance that the robot needs to move can be obtained by using the Dijkstra algorithm to solve the shortest path on the plane.
  • w is the sum of the waybill numbers of the waybill package A and the waybill package B
  • d0 is the total moving distance of the delivery robot on the same floor after the waybill package A and the waybill package B are combined
  • ⁇ is the weight coefficient.
  • step S103 is performed, two target waybill packages with the smallest similarity and the similarity less than a preset threshold in the similarity list are selected, and the total number of waybills after the two target waybill packages are merged is calculated, if the total number of waybills is If it is less than or equal to the preset delivery capacity N, the two target waybill packages are merged into one optimized waybill package, and the similarity list is updated.
  • the package merging process ends, and a set of waybill packages is formed, and the set of waybill packages includes at least one optimized waybill package.
  • Step 2 scheduling and sorting all optimized waybill packages in the set of waybill packages.
  • the remaining delivery time t of the waybill in the waybill package is defined as the difference between the expected delivery time and the current time. If there is no expected delivery time point on the waybill, the expected delivery time point is defined as the delivery initiation time point plus a fixed time T0, which is generally the delivery time constraint.
  • the size n of the waybill package that is, the number of waybills in the waybill package.
  • scheduling and sorting all optimized waybill packages in the waybill package set specifically includes the following steps:
  • S202 define a binary group Z(t*,n*) used to represent the score of the optimized waybill package, where t* represents the reverse feature of the remaining delivery time, and the longer the remaining delivery time, the smaller the value of t* .
  • n* represents the forward feature of the waybill package. The larger the value of n* is, the higher the score of the two-tuple is.
  • S203 schedule and sort all the optimized waybill packages according to the size of the binary group Z.
  • the size comparison rule of Z is that the first t* is compared first, and the second n* is compared second. At this point, the overall sequence of the optimized waybill package can be output.
  • the cloud scheduling method further includes a step of actively placing orders.
  • Order pressing refers to the phenomenon that the waybill is delayed and dispatched by the system, that is, the waybill is pressed and does not participate in the process of merging orders for the time being.
  • Billing is generally carried out before merging, and its purpose is to provide a large enough waybill pool for the merging process. Pressing the order will prolong the overall timeliness of the waybill, but a good merging effect will reduce the delivery timeliness of the overall waybill. Therefore, the control of pressing the order should be considered more as a balance.
  • the active order pressing step is specifically: obtaining the target optimized waybill package with the binary group Z value of 0, and the value of Z as 0 indicates that all the waybill in the waybill package have no remaining time critical situation, and the waybill package is small enough, these waybills.
  • the package will participate in the order pressing process, that is, it will not output these waybill packages directly, but only output the waybill package with Z>0.
  • a time field is set for each target waybill in the target optimization waybill package, and the time field is the time point when the target waybill first participates in the merging.
  • the order pressing time of the target waybill can be calculated according to the time field, and when the order pressing time is greater than the preset order pressing time threshold, the entire optimized waybill package containing the target waybill is forcibly output.
  • step 3 is performed, and the optimized waybill packages are sequentially allocated to the delivery robot according to the sorting result of the optimized waybill packages, that is, the capacity scheduling is completed.
  • the robot In the robot distribution system in the building, the robot has the ability to move, also called the capacity.
  • Capacity scheduling refers to selecting a robot according to a certain method for the optimized waybill package in turn, and issuing a distribution task to achieve the distribution of the waybill package to the robot.
  • the purpose of the present invention is that the higher the human efficiency of the transport capacity, the better.
  • the optimized waybill package is sequentially allocated to the delivery robot, which specifically includes the following steps:
  • S301 Obtain a list of candidate robots, where the robots on the candidate robot list have the following characteristics: the number of real-time waybills of the robot is less than the preset delivery capacity, and the remaining delivery time of any real-time waybill of the robot is greater than the preset delivery time The minimum remaining delivery time, that is, the robot is neither fully loaded nor carrying any waybills that are about to expire or have expired.
  • S302 output the optimized waybill package according to the sorting result, and obtain a robot that carries at least one real-time waybill package in the candidate robot list, calculate the similarity between the optimized waybill package and all real-time waybill packages in turn, and obtain the similarity At least one target robot that satisfies the above preset merging conditions, and distributes the optimized waybill package to the optimal target robot according to the preset order tracking principle.
  • the waybill package A is directly allocated to the current robot R. This process is called the order splitting process. The difference from the chasing process is whether the robot has already backed the order when the allocation occurs.
  • the process of chasing orders is given priority, followed by the process of direct order distribution. Moreover, if there are multiple robots such as R1 and R2, all of the currently considered waybill package A can be tracked, and the robot with a large amount of back orders will be given priority. For example, suppose that after chasing an order, the amount of the back order of R1 is N1, and the amount of back order of R2 is N2, if N1 ⁇ N2, choose R2, otherwise choose R1.
  • the passive ordering process means that the waybill package cannot be temporarily assigned to a suitable robot for transportation, and has to enter the ordering process.
  • the waybill in the waybill package that is passively pressed will be released and returned to the waybill pool, and will be merged with the new waybill that will flow into the system in the future to form the waybill package again.
  • the passive order placement process will not be controlled by the order placement time threshold, because its essence is that the available capacity is insufficient, and the system does not have an active strategy for placing orders, so there is no control of the order placement time threshold.
  • the above cloud scheduling method of delivery robots not only performs order pressing, merging and ordering of delivery orders for multiple delivery orders from the perspective of waybill scheduling, but also screens, chases and divides orders from the perspective of capacity scheduling, and chases orders first. Orders are distributed afterward, thereby maximizing the aggregation result of the waybill, increasing the amount of back orders of the robot, reducing the operating cost of robot distribution in practice, and improving the distribution efficiency and service experience at the same time.
  • the embodiment of the present invention further provides a cloud scheduling apparatus for a delivery robot.
  • the cloud scheduling device of the delivery robot may be a software module, and the software module includes several instructions, which are stored in a memory, and the processor can access the memory and call the instructions for execution, so as to complete the delivery robot described in the above embodiments. cloud scheduling method.
  • the cloud scheduling device of the delivery robot can also be constructed by hardware devices.
  • the cloud scheduling device of the delivery robot can be constructed by one or more chips, and the chips can work in coordination with each other to The cloud scheduling method for the delivery robot described in the above embodiments is completed.
  • the cloud dispatching device of the delivery robot can also be constructed of various logic devices, such as general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), Microcontrollers, ARM (AcornRISCMachine) or other programmable logic devices, discrete gate or transistor logic, discrete hardware components, or any combination of these components.
  • DSPs digital signal processors
  • ASICs application-specific integrated circuits
  • FPGAs field programmable gate arrays
  • Microcontrollers ARM (AcornRISCMachine) or other programmable logic devices, discrete gate or transistor logic, discrete hardware components, or any combination of these components.
  • FIG. 3 is a schematic structural diagram of a cloud scheduling device for a delivery robot provided in Embodiment 2 of the present invention.
  • the cloud scheduling device for a delivery robot includes an aggregation module 100, a sorting module 200, and an allocation module 300.
  • the aggregation module 100 is configured to aggregate the waybills in the waybill pool according to similarity to form a set of waybill packages including at least one optimized waybill package;
  • the sorting module 200 is configured to schedule and sort all the optimized waybill packages in the set of waybill packages;
  • the assigning module 300 is configured to sequentially assign the optimized waybill packages to the delivery robot according to the sorting result of the optimized waybill packages.
  • the aggregation module 100 specifically includes:
  • the new unit 101 is used to obtain a list of waybills corresponding to the waybill pool, and create a corresponding waybill package for each waybill in the list of waybill;
  • the first calculation unit 102 is used to calculate the similarity between any two waybill packages by using a preset similarity formula, and establish a similarity list. the lower;
  • the merging unit 103 is configured to select two target waybill packages with the smallest similarity and the similarity less than a preset threshold in the similarity list, and calculate the total number of waybills after the two target waybill packages are merged. If the number is less than or equal to the preset delivery capacity, the two target waybill packages are merged into one optimized waybill package, and the similarity list is updated;
  • the set generating unit 104 is used to repeatedly drive the merging unit until the similarity of any two waybill packages in the similarity list is greater than or equal to the preset threshold or the total number of waybills after the two target waybill packages are merged
  • the delivery capacity is greater than the preset delivery capacity
  • the process of merging the waybill packages is completed, and a set of waybill packages is formed, and the set of waybill packages includes at least one optimized waybill package.
  • the preset similarity formula is:
  • S(A, B) is the similarity between the waybill package A and the waybill package B
  • F is the sum of the number of stairs climbed by the robot after the waybill package A and the waybill package B are combined
  • w is the waybill package A and waybill package B.
  • the sum of the numbers, d0 is the total moving distance of the delivery robot on the same floor after the waybill package A and the waybill package B are combined
  • ⁇ is the weight coefficient.
  • the sorting module 200 specifically includes:
  • a first obtaining unit 201 configured to obtain the remaining delivery time t of each waybill in the optimized waybill package and the number of waybill n of each optimized waybill package;
  • the sorting unit 203 is configured to schedule and sort all the optimized waybill packages according to the size of the binary group Z. The larger the binary group Z, the higher the ranking of the corresponding optimized waybill package.
  • the cloud scheduling device of the delivery robot further includes an order pressing module 400, and the order pressing module 400 is configured to obtain a target optimized waybill package whose binary group Z is 0, and optimize the waybill package for the target Set a time field for each target waybill in , and return the target waybill to the waybill pool, and the time field is the time point when the target waybill first participated in the merger.
  • the cloud scheduling device of the delivery robot further includes a forced output module 500, which is configured to obtain the time field of each target waybill, and calculate the corresponding value of the target waybill according to the current time.
  • the order pressing time when the order pressing time is greater than the preset order pressing time threshold, the optimized waybill package containing the target waybill is preferentially output.
  • the distribution module 300 specifically includes:
  • the second obtaining unit 301 is configured to obtain a list of candidate robots, where the robots on the list of candidates have the following characteristics: the number of real-time waybills of the robot is less than the preset delivery capacity, and the remaining number of any real-time waybill of the robot is The delivery time is greater than the preset minimum remaining delivery time;
  • the tracking unit 302 is configured to output the optimized waybill package according to the sorting result, and obtain the robot that carries at least one real-time waybill package in the candidate robot list, and sequentially calculate the similarity between the optimized waybill package and all real-time waybill packages degree, obtain at least one target robot whose similarity satisfies the preset merging conditions, and allocate the optimized waybill package to the optimal target robot according to the preset tracking principle;
  • the allocation unit 303 is configured to allocate the optimized waybill package to any idle robot when the similarity between the real-time waybill package of all robots in the candidate robot list and the optimized waybill package does not meet the preset merging condition,
  • the idle robot is a robot that does not currently carry any waybill.
  • FIG. 4 is a schematic structural diagram of a server according to Embodiment 3 of the present invention.
  • the server 600 includes one or more processors 61 and a memory 62 .
  • one processor 61 is taken as an example in FIG. 4 .
  • the processor 61 and the memory 62 may be connected by a bus or other means.
  • the memory 62 can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as corresponding to the cloud scheduling method of the delivery robot in the embodiment of the present invention. program instructions/modules.
  • the processor 61 executes various functional applications and data processing of the cloud scheduling device of the delivery robot by running the non-volatile software programs, instructions and modules stored in the memory 62, that is, to realize the delivery robot provided by the above method embodiments.
  • Memory 62 may include high speed random access memory, and may also include nonvolatile memory, such as at least one magnetic disk storage device, flash memory device, or other nonvolatile solid state storage device. In some embodiments, memory 62 may optionally include memory located remotely from processor 61, which may be connected to processor 61 via a network. Examples of such networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
  • the program instructions/modules are stored in the memory 62, and when executed by the one or more processors 61, execute the cloud scheduling method for a delivery robot in any of the above method embodiments.
  • Embodiments of the present invention also provide a non-volatile computer storage medium, where the computer storage medium stores computer-executable instructions, and the computer-executable instructions are executed by one or more processors, for example, a process in FIG. 4
  • the controller 61 can cause the above one or more processors to execute the cloud scheduling method for the delivery robot in any of the above method embodiments.
  • An embodiment of the present invention also provides a computer program product, the computer program product includes a computer program stored on a non-volatile computer-readable storage medium, the computer program includes program instructions, and when the program instructions are electronically stored When the device executes, the electronic device is made to execute any one of the cloud scheduling methods for the delivery robot.
  • each embodiment can be implemented by means of software plus a general hardware platform, and certainly can also be implemented by hardware.
  • the above-mentioned technical solutions can be embodied in the form of software products in essence, or the parts that make contributions to related technologies, and the computer software products can be stored in computer-readable storage media, such as ROM/RAM, magnetic disks , optical disc, etc., including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform the methods described in various embodiments or some parts of the embodiments.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Development Economics (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Game Theory and Decision Science (AREA)
  • Educational Administration (AREA)
  • Data Mining & Analysis (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Robotics (AREA)
  • Mechanical Engineering (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Factory Administration (AREA)

Abstract

本发明公开了一种递送机器人的云端调度方法、装置和服务器，方法包括以下步骤：根据相似性对运单池中的运单进行聚合，形成包括至少一个优化运单包的运单包集合；对运单包集合中的所有优化运单包进行调度排序；按照优化运单包的排序结果依次将优化运单包分配给递送机器人。本发明不仅从运单调度的角度对多配送单进行压单、并单和运单包排序，而且从运力调度的角度对多机器人进行筛选、追单和分单，且先追单后分单，从而最大化运单聚合结果，提高机器人的背单量，降低实际中机器人配送的运营成本，同时提高了配送效率和服务体验。

Description

一种递送机器人的云端调度方法、装置和服务器 技术领域
本发明涉及机器人领域，尤其涉及一种递送机器人的云端调度方法、装置和服务器。
背景技术
目前的楼宇递送机器人，具备仓储能力和楼宇内移动的能力，可以承担楼宇内的外卖、快递的配送任务。现代生活中，随着外卖平台、电商购物平台的兴起，楼宇内的外卖和快递的数量日益攀升，与此同时，楼宇递送机器人也面临着服务效率和规模上的挑战，首先配送需求的增加，意味着需求的机器人数量会增加，其次，配送需求的增加，尤其是集中的配送情况(例如外卖午高峰时段)会造成配送时效的延长风险，进而造成服务体验的下降。因此，如何提高机器人集群的整体承运能力，如何在配送规模增大的同时保障机器人的配送时间，对递送机器人的成本优化和效率提升意义重大。
发明内容
本发明提供了一种递送机器人的云端调度方法、装置和服务器，解决了如何对运单以及机器人的运力进行调度，从而提高机器人递送效率的技术问题。
本发明解决上述技术问题的技术方案如下：一种递送机器人的云端 调度方法，包括以下步骤：
步骤1，根据相似性对运单池中的运单进行聚合，形成包括至少一个优化运单包的运单包集合；
步骤2，对所述运单包集合中的所有优化运单包进行调度排序；
步骤3，按照所述优化运单包的排序结果依次将所述优化运单包分配给递送机器人。
在一个优选实施方式中，所述根据相似性对运单池中的运单进行聚合，形成包括至少一个优化运单包的运单包集合，具体包括以下步骤：
S101，获取运单池对应的运单列表，对运单列表中的每个运单新建一个对应的运单包；
S102，采用预设相似度公式计算任意两个运单包的相似度，并建立相似度列表，所述相似度取值越小，两个运单包合并后的单均配送难度越低；
S103，选择所述相似度列表中相似度最小且相似度小于预设阈值的两个目标运单包，计算所述两个目标运单包合并后的总运单数，若所述总运单数小于或等于预设配送容量，则将两个目标运单包合并为一个优化运单包，并更新所述相似度列表；
S104，重复S103，直至所述相似度列表中任意两个运单包的相似度均大于或等于所述预设阈值或者两个目标运单包合并后的总运单数大于所述预设配送容量，运单包合并过程结束，形成运单包集合，所述运单包集合中包括至少一个优化运单包。
在一个优选实施方式中，所述预设相似度公式为：
S(A，B)＝F/w或者S(A，B)＝(d0+β*F)/w，
其中，S(A，B)为运单包A和运单包B的相似度，F为运单包A和 运单包B合并后机器人的爬楼次数之和，w为运单包A和运单包B的运单数之和，d0为运单包A和运单包B合并后递送机器人在同楼层的总移动距离，β为权重系数。
在一个优选实施方式中，所述对运单包集合中的所有优化运单包进行调度排序，具体包括以下步骤：
S201，获取所述优化运单包内每个运单的剩余配送时长t以及每个优化运单包的运单数量n；
S202，定义用于表示所述优化运单包得分的二元组Z(t*,n*)，当优化运单包中任一运单的剩余配送时长t小于预设最小剩余配送时长时，所述t*＝t_max–t，否则所述t*取值为0；当所述优化运单包的运单数量n大于预设最小运单数，则n*取值为n，否则n*取值为0，所述t_max为预设最大剩余配送时长；
S203，根据二元组Z的大小对所有优化运单包进行调度排序，二元组Z越大，对应优化运单包的排名越靠前。
在一个优选实施方式中，还包括主动压单步骤，具体为：获取二元组Z取值为0的目标优化运单包，为目标优化运单包中的每张目标运单设置时间字段，并将所述目标运单返回至运单池，所述时间字段为所述目标运单第一次参与并单的时间点。
在一个优选实施方式中，还包括运单强制输出步骤，具体为：获取每张目标运单的时间字段，并根据当前时间计算所述目标运单对应的压单时间，当所述压单时间大于预设压单时间阈值时，优先输出包含所述目标运单的优化运单包。
在一个优选实施方式中，所述按照优化运单包的排序结果依次将优化运单包分配给递送机器人，具体包括以下步骤：
S301，获取备选机器人列表，所述备选机器人列表上的机器人具有以下特征：所述机器人的实时运单数小于预设配送容量，且所述机器人任一实时运单的剩余配送时长均大于预设最小剩余配送时长；
S302，按照所述排序结果输出所述优化运单包，并获取备选机器人列表中承载有至少一个实时运单包的机器人，依次计算所述优化运单包和所有实时运单包的相似度，获取相似度满足预设合并条件的至少一个目标机器人，并按照预设追单原则将所述优化运单包分配给最优的目标机器人；
S303，若备选机器人列表中所有机器人的实时运单包与所述优化运单包的相似度均不满足预设合并条件，则将所述优化运单包分配给任一空闲机器人，所述空闲机器人为当前没有承载任何运单的机器人。
本发明实施例的第二方面提供了一种递送机器人的云端调度装置，包括聚合模块、排序模块和分配模块，
所述聚合模块用于根据相似性对运单池中的运单进行聚合，形成包括至少一个优化运单包的运单包集合；
所述排序模块用于对所述运单包集合中的所有优化运单包进行调度排序；
所述分配模块用于按照所述优化运单包的排序结果依次将所述优化运单包分配给递送机器人。
在一个优选实施方式中，所述聚合模块具体包括：
新建单元，用于获取运单池对应的运单列表，对运单列表中的每个运单新建一个对应的运单包；
第一计算单元，用于采用预设相似度公式计算任意两个运单包的相似度，并建立相似度列表，所述相似度取值越小，两个运单包合并后的 单均配送难度越低；
合并单元，用于选择所述相似度列表中相似度最小且相似度小于预设阈值的两个目标运单包，计算所述两个目标运单包合并后的总运单数，若所述总运单数小于或等于预设配送容量，则将两个目标运单包合并为一个优化运单包，并更新所述相似度列表；
集合生成单元，用于重复驱动所述合并单元，直至所述相似度列表中任意两个运单包的相似度均大于或等于所述预设阈值或者两个目标运单包合并后的总运单数大于所述预设配送容量，运单包合并过程结束，形成运单包集合，所述运单包集合中包括至少一个优化运单包。
在一个优选实施方式中，所述预设相似度公式为：
S(A，B)＝F/w或者S(A，B)＝(d0+β*F)/w，
其中，S(A，B)为运单包A和运单包B的相似度，F为运单包A和运单包B合并后机器人的爬楼次数之和，w为运单包A和运单包B的运单数之和，d0为运单包A和运单包B合并后递送机器人在同楼层的总移动距离，β为权重系数。
在一个优选实施方式中，所述排序模块具体包括：
第一获取单元，用于获取所述优化运单包内每个运单的剩余配送时长t以及每个优化运单包的运单数量n；
第二计算单元，用于定义用于表示所述优化运单包得分的二元组Z(t*,n*)，当优化运单包中任一运单的剩余配送时长t小于预设最小剩余配送时长时，所述t*＝t_max–t，否则所述t*取值为0；当所述优化运单包的运单数量n大于预设最小运单数，则n*取值为n，否则n*取值为0，所述t_max为预设最大剩余配送时长；
排序单元，用于根据二元组Z的大小对所有优化运单包进行调度排 序，二元组Z越大，对应优化运单包的排名越靠前。
在一个优选实施方式中，所述递送机器人的云端调度装置还包括压单模块，所述压单模块用于获取二元组Z取值为0的目标优化运单包，为目标优化运单包中的每张目标运单设置时间字段，并将所述目标运单返回至运单池，所述时间字段为所述目标运单第一次参与并单的时间点。
在一个优选实施方式中，所述递送机器人的云端调度装置还包括强制输出模块，所述强制输出模块用于获取每张目标运单的时间字段，并根据当前时间计算所述目标运单对应的压单时间，当所述压单时间大于预设压单时间阈值时，优先输出包含所述目标运单的优化运单包。
在一个优选实施方式中，所述分配模块具体包括：
第二获取单元，用于获取备选机器人列表，所述备选机器人列表上的机器人具有以下特征：所述机器人的实时运单数小于预设配送容量，且所述机器人任一实时运单的剩余配送时长均大于预设最小剩余配送时长；
追单单元，用于按照所述排序结果输出所述优化运单包，并获取备选机器人列表中承载有至少一个实时运单包的机器人，依次计算所述优化运单包和所有实时运单包的相似度，获取相似度满足预设合并条件的至少一个目标机器人，并按照预设追单原则将所述优化运单包分配给最优的目标机器人；
分配单元，用于当备选机器人列表中所有机器人的实时运单包与所述优化运单包的相似度均不满足预设合并条件时，则将所述优化运单包分配给任一空闲机器人，所述空闲机器人为当前没有承载任何运单的机器人。
本发明实施例的第三方面提供了一种服务器，包括存储器、处理器 以及存储在所述存储器中并可在所述处理器上运行的计算机程序，所述处理器执行所述计算机程序时实现以上所述递送机器人的云端调度方法的步骤。
本发明实施例的第四方面提供了一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，所述计算机程序被处理器执行时实现以上所述递送机器人的云端调度方法的步骤。
本发明提供了一种递送机器人的云端调度方法、装置和服务器，不仅从运单调度的角度对多配送单进行压单、并单和运单包排序，而且从运力调度的角度对多机器人进行筛选、追单和分单，且先追单后分单，从而最大化运单聚合结果，提高机器人的背单量，降低实际中机器人配送的运营成本，同时提高了配送效率和服务体验。
为使发明的上述目的、特征和优点能更明显易懂，下文特举本发明较佳实施例，并配合所附附图，作详细说明如下。
附图说明
为了更清楚地说明本发明实施例的技术方案，下面将对实施例中所需要使用的附图作简单地介绍，应当理解，以下附图仅示出了本发明的某些实施例，因此不应被看作是对范围的限定，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他相关的附图。
图1是实施例1提供的递送机器人的云端调度方法的流程示意图；
图2是实施例1中两运单包的相似度计算示意图；
图3是实施例2提供的递送机器人的云端调度装置的结构示意图；
图4是实施例3提供的一种服务器的结构示意图。
具体实施方式
为了使本发明的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本发明进行进一步详细说明。应当理解，此处所描述的具体实施例仅用以解释本发明，并不用于限定本发明。基于本发明中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
需要说明的是，如果不冲突，本发明实施例中的各个特征可以相互结合，均在本发明的保护范围之内。另外，虽然在装置示意图中进行了功能模块划分，在流程图中示出了逻辑顺序，但是在某些情况下，可以以不同于装置中的模块划分，或流程图中的顺序执行所示出或描述的步骤。再者，本发明所采用的“第一”、“第二”、“第三”等字样并不对数据和执行次序进行限定，仅是对功能和作用基本相同的相同项或相似项进行区分。
本发明的云端调度方法，首先假设递送机器人之间没有个体上的差别，即所有递送机器人所能承载的运单的最大数量都是一样的。同时，本发明采用云端服务器指派的方式来分配配送任务到机器人，而非机器人自主抢单的方式。其原因在于，云端服务器视角拥有所有机器人的运力情况、所有运单的情况，相比机器人本地的局部视角，更容易到达全局最优的分配结果。
同时，本发明的云端调度方法，兼容送单和取单的情况。一个只包含送单的配送模式是指，所有机器人从一个固定的配送起点出发，送完所有单子后再回到这个固定的配送起点，以接受下一次任务分配。包含取单的情况则是，机器人可以接受一个带配送起点的配送任务，机器人 要前往配送起点取单后才可以进行配送。
请参阅图1，为本发明实施例1提供一种递送机器人的云端调度方法的流程示意图，如图1所示，方法包括以下步骤：
步骤1，根据相似性对运单池中的运单进行聚合，形成包括至少一个优化运单包的运单包集合，这个即为运单调度过程。具体来说，运单即配送任务的简称，一张运单包含配送任务的起始点位、目标点位、配送收件人的联系方式等等。更进一步地，运单的整体配送时长，也叫做运单时效，本发明的目标是运单的时效越低越好。
优选实施例中，所述步骤1具体包括以下步骤：
S101，获取运单池对应的运单列表，对运单列表中的每个运单新建一个对应的运单包。
S102，采用预设相似度公式计算任意两个运单包之间的相似度，并建立相似度列表。优选实施例中，相似度的计算方式定义为：两个运单包合并后机器人需要爬楼的次数之和比上总单数，即合并后运单包内的单均配送难度，对应的预设相似度公式为：S(A，B)＝F/w，其中，S(A，B)为运单包A和运单包B的相似度，F为运单包A和运单包B合并后机器人的爬楼次数之和，w为运单包A和运单包B的运单数之和。所述相似性越好，相似度取值越小，两个运单包合并后的单均配送难度越低，两个运单包可以聚合进一步形成一个大的运单包的可能越高。
如图2所示，用一个垂直方向的箭头表示一张运单，例如从1F出发配送到8F的运单，表示为一个以1F线为起点、8F线为终点的箭头，从而可以采用以上预设相似度公式计算出图2中任意两个运单包之间的相似度。
上述相似度定义方式，只考虑了楼层不同的情况，而实际测试数据 表示，机器人上下楼层所消耗的时间是整个运单配送时间里面占比最高的，因为上下楼层需要机器人搭乘电梯，期间存在等电梯、进出电梯、乘梯等步骤。不过，仍然可以进一步细化到同楼层的因素，即可以对相似度的定义做进一步扩展，加入同楼层内的距离因素。例如，扩展地定义相似度如下：两个运单包合并后机器人需要移动的距离的之和比上总单数。其中机器人需要移动的最短距离，在平面上可以采用Dijkstra(戴克斯特拉)算法进行最短路径求解后获得，其结果假设为d0，再加上楼层因素(带一个较大的权重值β)，即S(A，B)＝(d0+β*F)/w，S(A，B)为运单包A和运单包B的相似度，F为运单包A和运单包B合并后机器人的爬楼次数之和，w为运单包A和运单包B的运单数之和，d0为运单包A和运单包B合并后递送机器人在同楼层的总移动距离，β为权重系数。如此说明，运单包相似度的定义是可以进一步扩展的，扩展后的相似度将仍然适合本云端调度方法的并单过程。
假设每个机器人的预设配送容量是N，即每个机器人一次性最多配送N单，并且每个机器人个体之间容量无差异。然后执行步骤S103，选择所述相似度列表中相似度最小且相似度小于预设阈值的两个目标运单包，计算所述两个目标运单包合并后的总运单数，若所述总运单数小于或等于预设配送容量N，则将两个目标运单包合并为一个优化运单包，并更新所述相似度列表。重复以上步骤,直至所述相似度列表中任意两个运单包的相似度均大于或等于所述预设阈值或者两个目标运单包合并后的总运单数大于所述预设配送容量N，运单包合并过程结束，形成运单包集合，所述运单包集合中包括至少一个优化运单包。
步骤2，对所述运单包集合中的所有优化运单包进行调度排序。排序时考虑以下两个因素：
第一，运单包内运单的剩余配送时长t。一般地，配送任务一旦开始，总会存在一个配送时效的约束，例如外卖配送属于即时配送，收件客户一定要在预期时间内用餐才可以认为此次配送是有效的。运单的剩余配送时长定义为预期配送时间点对当前时间的差值。如果运单不存在预期配送时间点，则定义预期配送时间点为配送发起时间点加上一个固定的时间T0，这个T0一般是配送的时效约束。
第二，运单包的大小n，即运单包内的运单数量。
综上，优选实施例中，对运单包集合中的所有优化运单包进行调度排序，具体包括以下步骤：
S201，获取所述优化运单包内每个运单的剩余配送时长t以及每个优化运单包的运单数量n。
S202，定义用于表示所述优化运单包得分的二元组Z(t*,n*)，其中t*表示剩余配送时长的反向特征，所述剩余配送时长越长，t*值越小。n*表示运单包的正向特征，n*取值越大，二元组的得分越高。
如果运单包内存在运单，其剩余配送时长t已经小于一个预设最小剩余配送时长，即临界剩余配送时间t0，则此张运单即将超时或者已经超时，设置t*＝t_max–t。如果运单包内所有运单都无此情况，直接设置Z二元组的第一位t*＝0，所述t_max为预设最大剩余配送时长，是一个足够大的且和剩余配送时长t同在时间单位的正数。
同时，预设一个运单包的大小阈值n_min(即预设最小运单数)，用于表示压单的包的大小临界值，如果优化运单包的大小n大于阈值n_min，则设置Z二元组的第二位n*＝n，否则直接设置n＝0。
S203，根据二元组Z的大小对所有优化运单包进行调度排序，二元组Z越大，对应优化运单包的排名越靠前。具体地，Z的大小比较规则 是，优先比较第一位t*，其次比较第二位n*。至此，可以输出优化运单包的整体顺序。
优选实施例中，所述云端调度方法还包括主动压单步骤。压单是指运单被系统延迟调度的现象，即运单被压着暂不参与并单过程。压单一般是在并单前进行，其目的是为并单过程提供足够大的运单池。压单会造成运单整体时效的延长，但是好的并单效果则会减少整体运单的配送时效，因此压单的控制更多地考虑是一种平衡。
所述主动压单步骤具体为：获取二元组Z取值为0的目标优化运单包，Z取值为0表示运单包内所有运单没有剩余时间临界的情况，且运单包足够小，这些运单包将参与到压单过程，即不会直接输出这些运单包，而是只输出Z>0的运单包，对于Z＝0的运单包，其中的运单将返还到运单池，参与下一次并单过程。
同时为目标优化运单包中的每张目标运单设置时间字段，所述时间字段为所述目标运单第一次参与并单的时间点。这样根据所述时间字段可以计算出所述目标运单的压单时间，当所述压单时间大于预设压单时间阈值时，强制输出包含所述目标运单的整个优化运单包。
然后执行步骤3，按照所述优化运单包的排序结果依次将所述优化运单包分配给递送机器人，即完成运力调度。在楼宇内机器人配送系统中，机器人具备移动的能力，也叫做运力。运力调度是指依次对所述优化运单包按一定方法挑选一个机器人，下发配送任务，达成运单包到机器人的分配。本发明的目的是运力的人效越高越好。
优选实施例中，按照优化运单包的排序结果依次将优化运单包分配给递送机器人，具体包括以下步骤：
S301，获取备选机器人列表，所述备选机器人列表上的机器人具有 以下特征：所述机器人的实时运单数小于预设配送容量，且所述机器人任一实时运单的剩余配送时长均大于预设最小剩余配送时长，即机器人既没有满载，也没有承运任一即将超时运单或者已超时运单。
S302，按照所述排序结果输出所述优化运单包，并获取备选机器人列表中承载有至少一个实时运单包的机器人，依次计算所述优化运单包和所有实时运单包的相似度，获取相似度满足以上预设合并条件的至少一个目标机器人，并按照预设追单原则将所述优化运单包分配给最优的目标机器人。
S303，若备选机器人列表中所有机器人的实时运单包与所述优化运单包的相似度均不满足预设合并条件，则将所述优化运单包分配给任一空闲机器人，所述空闲机器人为当前没有承载任何运单的机器人。
比如，考虑当前要分配的运单包A，依次考察每个机器人R，如果机器人R身上已经承运一个运单包B，则考虑A和B的相似度S(A,B)，如果A和B可以合并，即S(A,B)小于相似度预设阈值且合并后的大小没有达到预设配送容量N，则可以把A分配给当前机器人R。此过程，叫做追单过程。
如果机器人R身上没有任何运单包，即处于空闲状态，则直接将运单包A分配给当前机器人R，此过程叫做分单过程。区别于追单过程的地方，在于机器人在分配发生时是否已经背单。
对于以上追单和分单两个过程，优先考虑追单过程，其次考虑直接分单过程。而且，如果存在多个机器人例如R1和R2，对于当前考虑的运单包A都可以进行追单，则优先考虑追单后背单量大的机器人。举例来说，假设追单后，R1的背单量是N1，R2的背单量是N2，如果N1<N2，则选择R2，否则选择R1。
上述追单和分单步骤完成后，如仍然存在运单包未进行分配，则进入被动压单过程。此过程区别于前面所述的主动压单过程，被动压单过程是指运单包暂时无法分配到合适的机器人承运，而不得已进入压单过程。被动压单的运单包中的运单将会释放返还回运单池，和系统未来流入的新运单一起并单，重新形成运单包。此外，被动压单过程，将不受压单时间阈值的控制，因为其本质是可用运力不足，系统并无主动策略进行压单，也就没有压单时间阈值的控制的说法。
以上递送机器人的云端调度方法不仅从运单调度的角度对多配送单进行压单、并单和运单包排序，而且从运力调度的角度对多机器人进行筛选、追单和分单，且先追单后分单，从而最大化运单聚合结果，提高机器人的背单量，降低实际中机器人配送的运营成本，同时提高了配送效率和服务体验。
需要说明的是，在上述各个实施例中，上述各步骤之间并不必然存在一定的先后顺序，本领域普通技术人员，根据本发明实施例的描述可以理解，不同实施例中，上述各步骤可以有不同的执行顺序，亦即，可以并行执行，亦可以交换执行等等。
作为本发明实施例的另一方面，本发明实施例还提供一种递送机器人的云端调度装置。其中，递送机器人的云端调度装置可以为软件模块，所述软件模块包括若干指令，其存储在存储器内，处理器可以访问该存储器，调用指令进行执行，以完成上述各个实施例所阐述的递送机器人的云端调度方法。
在一些实施例中，递送机器人的云端调度装置亦可以由硬件器件搭建成的，例如，递送机器人的云端调度装置可以由一个或两个以上的芯片搭建而成，各个芯片可以互相协调工作，以完成上述各个实施例所阐 述的递送机器人的云端调度方法。再例如，递送机器人的云端调度装置还可以由各类逻辑器件搭建而成，诸如由通用处理器、数字信号处理器(DSP)、专用集成电路(ASIC)、现场可编程门阵列(FPGA)、单片机、ARM(AcornRISCMachine)或其它可编程逻辑器件、分立门或晶体管逻辑、分立的硬件组件或者这些部件的任何组合而搭建成。
图3是本发明实施例2提供一种递送机器人的云端调度装置的结构示意图，该递送机器人的云端调度装置包括聚合模块100、排序模块200和分配模块300，
所述聚合模块100用于根据相似性对运单池中的运单进行聚合，形成包括至少一个优化运单包的运单包集合；
所述排序模块200用于对所述运单包集合中的所有优化运单包进行调度排序；
所述分配模块300用于按照所述优化运单包的排序结果依次将所述优化运单包分配给递送机器人。
在一个优选实施方式中，所述聚合模块100具体包括：
新建单元101，用于获取运单池对应的运单列表，对运单列表中的每个运单新建一个对应的运单包；
第一计算单元102，用于采用预设相似度公式计算任意两个运单包的相似度，并建立相似度列表，所述相似度取值越小，两个运单包合并后的单均配送难度越低；
合并单元103，用于选择所述相似度列表中相似度最小且相似度小于预设阈值的两个目标运单包，计算所述两个目标运单包合并后的总运单数，若所述总运单数小于或等于预设配送容量，则将两个目标运单包合并为一个优化运单包，并更新所述相似度列表；
集合生成单元104，用于重复驱动所述合并单元，直至所述相似度列表中任意两个运单包的相似度均大于或等于所述预设阈值或者两个目标运单包合并后的总运单数大于所述预设配送容量，运单包合并过程结束，形成运单包集合，所述运单包集合中包括至少一个优化运单包。
在一个优选实施方式中，所述预设相似度公式为：
S(A，B)＝F/w或者S(A，B)＝(d0+β*F)/w，
其中，S(A，B)为运单包A和运单包B的相似度，F为运单包A和运单包B合并后机器人的爬楼次数之和，w为运单包A和运单包B的运单数之和，d0为运单包A和运单包B合并后递送机器人在同楼层的总移动距离，β为权重系数。
在一个优选实施方式中，所述排序模块200具体包括：
第一获取单元201，用于获取所述优化运单包内每个运单的剩余配送时长t以及每个优化运单包的运单数量n；
第二计算单元202，用于定义用于表示所述优化运单包得分的二元组Z(t*,n*)，当优化运单包中任一运单的剩余配送时长t小于预设最小剩余配送时长时，所述t*＝t_max–t，否则所述t*取值为0；当所述优化运单包的运单数量n大于预设最小运单数，则n*取值为n，否则n*取值为0，所述t_max为预设最大剩余配送时长；
排序单元203，用于根据二元组Z的大小对所有优化运单包进行调度排序，二元组Z越大，对应优化运单包的排名越靠前。
在一个优选实施方式中，所述递送机器人的云端调度装置还包括压单模块400，所述压单模块400用于获取二元组Z取值为0的目标优化运单包，为目标优化运单包中的每张目标运单设置时间字段，并将所述目标运单返回至运单池，所述时间字段为所述目标运单第一次参与并单 的时间点。
在一个优选实施方式中，所述递送机器人的云端调度装置还包括强制输出模块500，所述强制输出模块500用于获取每张目标运单的时间字段，并根据当前时间计算所述目标运单对应的压单时间，当所述压单时间大于预设压单时间阈值时，优先输出包含所述目标运单的优化运单包。
在一个优选实施方式中，所述分配模块300具体包括：
第二获取单元301，用于获取备选机器人列表，所述备选机器人列表上的机器人具有以下特征：所述机器人的实时运单数小于预设配送容量，且所述机器人任一实时运单的剩余配送时长均大于预设最小剩余配送时长；
追单单元302，用于按照所述排序结果输出所述优化运单包，并获取备选机器人列表中承载有至少一个实时运单包的机器人，依次计算所述优化运单包和所有实时运单包的相似度，获取相似度满足预设合并条件的至少一个目标机器人，并按照预设追单原则将所述优化运单包分配给最优的目标机器人；
分配单元303，用于当备选机器人列表中所有机器人的实时运单包与所述优化运单包的相似度均不满足预设合并条件时，则将所述优化运单包分配给任一空闲机器人，所述空闲机器人为当前没有承载任何运单的机器人。
图4是本发明实施例3提供的一种服务器的结构示意图，如图4所示，该服务器600包括一个或多个处理器61以及存储器62。其中，图4中以一个处理器61为例。
处理器61和存储器62可以通过总线或者其他方式连接。存储器62 作为一种非易失性计算机可读存储介质，可用于存储非易失性软件程序、非易失性计算机可执行程序以及模块，如本发明实施例中的递送机器人的云端调度方法对应的程序指令/模块。处理器61通过运行存储在存储器62中的非易失性软件程序、指令以及模块，从而执行递送机器人的云端调度装置的各种功能应用以及数据处理，即实现上述方法实施例提供的递送机器人的云端调度方法以及上述装置实施例的各个模块或单元的功能。
存储器62可以包括高速随机存取存储器，还可以包括非易失性存储器，例如至少一个磁盘存储器件、闪存器件、或其他非易失性固态存储器件。在一些实施例中，存储器62可选包括相对于处理器61远程设置的存储器，这些远程存储器可以通过网络连接至处理器61。上述网络的实例包括但不限于互联网、企业内部网、局域网、移动通信网及其组合。
所述程序指令/模块存储在所述存储器62中，当被所述一个或者多个处理器61执行时，执行上述任意方法实施例中的递送机器人的云端调度方法。
本发明实施例还提供了一种非易失性计算机存储介质，所述计算机存储介质存储有计算机可执行指令，该计算机可执行指令被一个或多个处理器执行，例如图4中的一个处理器61，可使得上述一个或多个处理器可执行上述任意方法实施例中的递送机器人的云端调度方法。
本发明实施例还提供了一种计算机程序产品，所述计算机程序产品包括存储在非易失性计算机可读存储介质上的计算机程序，所述计算机程序包括程序指令，当所述程序指令被电子设备执行时，使所述电子设备执行任一项所述的递送机器人的云端调度方法。
以上所描述的装置或设备实施例仅仅是示意性的，其中所述作为分离部件说明的单元模块可以是或者也可以不是物理上分开的，作为模块单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络模块单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到各实施方式可借助软件加通用硬件平台的方式来实现，当然也可以通过硬件。基于这样的理解，上述技术方案本质上或者说对相关技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品可以存储在计算机可读存储介质中，如ROM/RAM、磁碟、光盘等，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行各个实施例或者实施例的某些部分所述的方法。
最后应说明的是：以上实施例仅用以说明本发明的技术方案，而非对其限制；在本发明的思路下，以上实施例或者不同实施例中的技术特征之间也可以进行组合，步骤可以以任意顺序实现，并存在如上所述的本发明的不同方面的许多其它变化，为了简明，它们没有在细节中提供；尽管参照前述实施例对本发明进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本发明各实施例技术方案的范围。

Claims (10)

  1. 一种递送机器人的云端调度方法，其特征在于，包括以下步骤：
    步骤1，根据相似性对运单池中的运单进行聚合，形成包括至少一个优化运单包的运单包集合；
    步骤2，对所述运单包集合中的所有优化运单包进行调度排序；
    步骤3，按照所述优化运单包的排序结果依次将所述优化运单包分配给递送机器人。
  2. 根据权利要求1所述递送机器人的云端调度方法，其特征在于，所述根据相似性对运单池中的运单进行聚合，形成包括至少一个优化运单包的运单包集合，具体包括以下步骤：
    S101，获取运单池对应的运单列表，对运单列表中的每个运单新建一个对应的运单包；
    S102，采用预设相似度公式计算任意两个运单包的相似度，并建立相似度列表，所述相似度取值越小，两个运单包合并后的单均配送难度越低；
    S103，选择所述相似度列表中相似度最小且相似度小于预设阈值的两个目标运单包，计算所述两个目标运单包合并后的总运单数，若所述总运单数小于或等于预设配送容量，则将两个目标运单包合并为一个优化运单包，并更新所述相似度列表；
    S104，重复S103，直至所述相似度列表中任意两个运单包的相似度均大于或等于所述预设阈值或者两个目标运单包合并后的总运单数大于所述预设配送容量，运单包合并过程结束，形成运单包集合，所述运 单包集合中包括至少一个优化运单包。
  3. 根据权利要求2所述递送机器人的云端调度方法，其特征在于，所述预设相似度公式为：
    S(A，B)＝F/w或者S(A，B)＝(d0+β*F)/w，
    其中，S(A，B)为运单包A和运单包B的相似度，F为运单包A和运单包B合并后机器人的爬楼次数之和，w为运单包A和运单包B的运单数之和，d0为运单包A和运单包B合并后递送机器人在同楼层的总移动距离，β为权重系数。
  4. 根据权利要求1-3任一所述递送机器人的云端调度方法，其特征在于，所述对运单包集合中的所有优化运单包进行调度排序，具体包括以下步骤：
    S201，获取所述优化运单包内每个运单的剩余配送时长t以及每个优化运单包的运单数量n；
    S202，定义用于表示所述优化运单包得分的二元组Z(t*,n*)，当优化运单包中任一运单的剩余配送时长t小于预设最小剩余配送时长时，所述t*＝t_max–t，否则所述t*取值为0；当所述优化运单包的运单数量n大于预设最小运单数，则n*取值为n，否则n*取值为0，所述t_max为预设最大剩余配送时长；
    S203，根据二元组Z的大小对所有优化运单包进行调度排序，二元组Z越大，对应优化运单包的排名越靠前。
  5. 根据权利要求4所述递送机器人的云端调度方法，其特征在于，还包括主动压单步骤，具体为：获取二元组Z取值为0的目标优化运单 包，为目标优化运单包中的每张目标运单设置时间字段，并将所述目标运单返回至运单池，所述时间字段为所述目标运单第一次参与并单的时间点。
  6. 根据权利要求5所述递送机器人的云端调度方法，其特征在于，还包括运单强制输出步骤，具体为：获取每张目标运单的时间字段，并根据当前时间计算所述目标运单对应的压单时间，当所述压单时间大于预设压单时间阈值时，优先输出包含所述目标运单的优化运单包。
  7. 根据权利要求6所述递送机器人的云端调度方法，其特征在于，所述按照优化运单包的排序结果依次将优化运单包分配给递送机器人，具体包括以下步骤：
    S301，获取备选机器人列表，所述备选机器人列表上的机器人具有以下特征：所述机器人的实时运单数小于预设配送容量，且所述机器人任一实时运单的剩余配送时长均大于预设最小剩余配送时长；
    S302，按照所述排序结果输出所述优化运单包，并获取备选机器人列表中承载有至少一个实时运单包的机器人，依次计算所述优化运单包和所有实时运单包的相似度，获取相似度满足预设合并条件的至少一个目标机器人，并按照预设追单原则将所述优化运单包分配给最优的目标机器人；
    S303，若备选机器人列表中所有机器人的实时运单包与所述优化运单包的相似度均不满足预设合并条件，则将所述优化运单包分配给任一空闲机器人，所述空闲机器人为当前没有承载任何运单的机器人。
  8. 一种递送机器人的云端调度装置，其特征在于，包括聚合模块、 排序模块和分配模块，
    所述聚合模块用于根据相似性对运单池中的运单进行聚合，形成包括至少一个优化运单包的运单包集合；
    所述排序模块用于对所述运单包集合中的所有优化运单包进行调度排序；
    所述分配模块用于按照所述优化运单包的排序结果依次将所述优化运单包分配给递送机器人。
  9. 一种计算机可读存储介质，存储有计算机程序，其特征在于，所述计算机程序被处理器执行时，实现权利要求1-7任一项所述递送机器人的云端调度方法。
  10. 一种服务器，其特征在于，包括权利要求9所述的计算机可读存储介质和处理器，所述处理器执行所述计算机可读存储介质上的计算机程序时实现如权利要求1-7任一项所述递送机器人的云端调度方法的步骤。
PCT/CN2021/100231 2020-09-09 2021-06-16 一种递送机器人的云端调度方法、装置和服务器 WO2022052543A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010939433.9A CN112053068B (zh) 2020-09-09 2020-09-09 一种递送机器人的云端调度方法、装置和服务器
CN202010939433.9 2020-09-09

Publications (1)

Publication Number Publication Date
WO2022052543A1 true WO2022052543A1 (zh) 2022-03-17

Family

ID=73611724

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/100231 WO2022052543A1 (zh) 2020-09-09 2021-06-16 一种递送机器人的云端调度方法、装置和服务器

Country Status (2)

Country Link
CN (1) CN112053068B (zh)
WO (1) WO2022052543A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116362844A (zh) * 2023-04-03 2023-06-30 大连理工大学 一种“先并单-再派单”的外卖订单配送方案生成方法

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112053068B (zh) * 2020-09-09 2021-08-24 上海有个机器人有限公司 一种递送机器人的云端调度方法、装置和服务器
CN112631310A (zh) * 2021-01-29 2021-04-09 泗阳县人民医院 一种多机器人协作的楼内物件配送系统
CN112836987B (zh) * 2021-03-01 2022-12-09 拉扎斯网络科技（上海）有限公司 一种运单调度方法和装置，计算机存储介质和电子设备
CN113033971B (zh) * 2021-03-05 2024-03-22 上海有个机器人有限公司 一种服务器及递送机器人的运单调度方法、介质和装置
CN113172625A (zh) * 2021-04-29 2021-07-27 塔米智能科技（北京）有限公司 一种机器人调度方法、装置、设备及存储介质
CN113642959B (zh) * 2021-08-06 2024-07-09 上海有个机器人有限公司 物品配送任务分配方法、装置、计算机设备及存储介质
CN113887817A (zh) * 2021-10-19 2022-01-04 上海擎朗智能科技有限公司 配送机器人的控制方法、配送系统及计算机可读存储介质
CN114493406B (zh) * 2021-12-21 2024-05-28 重庆特斯联智慧科技股份有限公司 物流机器人运行大数据分析调度系统及其调度方法

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2017208026A (ja) * 2016-05-20 2017-11-24 株式会社神戸製鋼所 在庫引当装置及び在庫引当方法
CN108694636A (zh) * 2017-04-10 2018-10-23 北京京东尚科信息技术有限公司 一种优化组单的方法和装置
CN109242184A (zh) * 2018-09-06 2019-01-18 泉州装备制造研究所 一种基于层次聚类的订单拣货优化方法
US20190228458A1 (en) * 2018-01-19 2019-07-25 International Business Machines Coporation Combining orders from multiple customers to a seller in e-commerce
CN111582612A (zh) * 2019-02-15 2020-08-25 拉扎斯网络科技（上海）有限公司 订单分配方法、装置、电子设备及计算机可读存储介质
CN112053068A (zh) * 2020-09-09 2020-12-08 上海有个机器人有限公司 一种递送机器人的云端调度方法、装置和服务器

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105450684B (zh) * 2014-08-15 2019-01-01 中国电信股份有限公司 云计算资源调度方法和系统
CN107133752B (zh) * 2016-02-29 2022-01-28 菜鸟智能物流控股有限公司 用于物流配送的数据处理、基于配送方的移动终端进行物流配送的方法和装置
CN107977763B (zh) * 2016-10-24 2022-05-06 菜鸟智能物流控股有限公司 一种资源分配方法和相关装置
CN107392512B (zh) * 2016-11-25 2018-06-01 北京小度信息科技有限公司 任务分组方法和装置
CN106897852A (zh) * 2017-01-27 2017-06-27 华南理工大学 用于物流的订单排序优化方法
CN107292701A (zh) * 2017-05-25 2017-10-24 北京小度信息科技有限公司 订单分组方法和装置
CN111612385B (zh) * 2019-02-22 2024-04-16 北京京东振世信息技术有限公司 一种对待配送物品进行聚类的方法和装置
CN111256718B (zh) * 2020-02-12 2023-04-18 上海东普信息科技有限公司 派件路线规划方法、装置、系统、设备及存储介质

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2017208026A (ja) * 2016-05-20 2017-11-24 株式会社神戸製鋼所 在庫引当装置及び在庫引当方法
CN108694636A (zh) * 2017-04-10 2018-10-23 北京京东尚科信息技术有限公司 一种优化组单的方法和装置
US20190228458A1 (en) * 2018-01-19 2019-07-25 International Business Machines Coporation Combining orders from multiple customers to a seller in e-commerce
CN109242184A (zh) * 2018-09-06 2019-01-18 泉州装备制造研究所 一种基于层次聚类的订单拣货优化方法
CN111582612A (zh) * 2019-02-15 2020-08-25 拉扎斯网络科技（上海）有限公司 订单分配方法、装置、电子设备及计算机可读存储介质
CN112053068A (zh) * 2020-09-09 2020-12-08 上海有个机器人有限公司 一种递送机器人的云端调度方法、装置和服务器

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116362844A (zh) * 2023-04-03 2023-06-30 大连理工大学 一种“先并单-再派单”的外卖订单配送方案生成方法
CN116362844B (zh) * 2023-04-03 2023-11-03 大连理工大学 一种“先并单-再派单”的外卖订单配送方案生成方法

Also Published As

Publication number Publication date
CN112053068A (zh) 2020-12-08
CN112053068B (zh) 2021-08-24

Similar Documents

Publication Publication Date Title
WO2022052543A1 (zh) 一种递送机器人的云端调度方法、装置和服务器
US11544645B2 (en) Inventory scheduling method and device and non-transitory computer readable storage medium
CN109544054B (zh) 库存调度方法、装置以及计算机可读存储介质
US20220063915A1 (en) Goods sorting method and goods sorting system
CN108596540B (zh) 补货方法及装置、计算机可读存储介质、电子设备
CN110197351B (zh) 一种物品出库方法和装置
WO2019120158A1 (zh) 物品拣选方法及相关设备
CN110390499B (zh) 订单分配方法和装置
CN109767151A (zh) 仓储管理方法、装置、介质及电子设备
WO2019000780A1 (zh) 订单调度方法、装置、电子设备及计算机可读存储介质
CN113044462B (zh) 机器人调度的方法、装置、系统、存储介质及程序产品
CN110390498B (zh) 订单分配方法和装置
CN114202266B (zh) 一种烟草仓储调度方法及装置
CN112198880A (zh) 一种agv任务分配方法、物流分拣方法及系统
CN112278674A (zh) 调度方法、装置、设备及存储介质
CN110390497B (zh) 物品入库方法和装置
TW202225065A (zh) 倉儲機器人的控制方法、裝置、設備及儲存媒體
CN110134985A (zh) 多层穿梭车系统输送机缓存长度建模及优化方法
CN112101831A (zh) 货物的出库方法、装置、介质及电子设备
CN114620387A (zh) 空单元容器处理方法和装置
CN112214024A (zh) 一种agv任务分配方法、物流分拣方法及系统
WO2024199517A1 (zh) 订单处理方法、任务调度方法、装置及电子设备
WO2020144879A1 (ja) 入出庫管理装置、入出庫管理システム、入出庫管理方法およびプログラム
WO2021104524A1 (zh) 一种agv调度方法和装置
CN117408780A (zh) 订单处理方法、装置、电子设备及机器可读存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21865602

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21865602

Country of ref document: EP

Kind code of ref document: A1