Self-service robot system and control method thereof

Abstract

The invention provides a self-service robot system and a control method thereof, and relates to the technical field of artificial intelligence, the system comprises at least one robot and a robot management platform, and the robot comprises a motion control module, a business module and an interaction module; the business module is integrated with all business services in a preset application scene; the interaction module generates a robot scheduling request based on the identified client intention; the robot management platform responds to a robot scheduling request initiated by the interaction module or the customer-oriented business system, a scheduling task is determined based on the robot scheduling request and state information and movement information of the robot, and the scheduling task represents a target robot, a destination and a target business service; and the motion control module of the target robot responds to the scheduling task and drives the target robot to arrive at the destination to execute the target business service. According to the invention, there is no need to configure a plurality of robots and respective exclusive scheduling processes, and the problems of deployment cost increase and resource redundancy are avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, and in particular to a self-service robot system and a control method thereof. BACKGROUND

[0002] With the rapid development of artificial intelligence and robot technology, self-service robots are equipped in public places of finance, medical treatment, government affairs, transportation and the like. Since the self-service robots usually only have part of functions, such as the self-service robots of financial sites only having one or two functions of business handling, business consultation and answer, queuing, and welcoming reception, a plurality of different types of self-service robots are usually equipped, and exclusive scheduling processes for various robots are configured for collaborative operation, which not only increases deployment cost, but also causes resource redundancy. SUMMARY

[0003] The present application provides a self-service robot system and a control method thereof, to solve the defects that a plurality of different types of self-service robots are equipped in the prior art, and exclusive scheduling processes for various robots are configured for collaborative operation, which not only increases deployment cost, but also causes resource redundancy.

[0004] The present application provides a self-service robot system, comprising at least one robot and a robot management platform, the robot comprising a motion control module, a business module and an interaction module; The business module integrates all business services in a preset application scenario; The interaction module is used for identifying a customer intention based on voice information of the customer, or based on the voice information of the customer and environmental information in the application scenario, and generating a robot scheduling request based on the customer intention; The robot management platform is used for determining a scheduling task based on the robot scheduling request and state information and movement information of all the robots in response to the robot scheduling request initiated by the interaction module or a business system facing customers, the scheduling task containing a target robot, a destination and a target business service, and the robot management platform transmitting the scheduling task to the target robot; The motion control module of the target robot is used for driving the target robot to reach the destination in response to the scheduling task, so that the target robot performs the target business service.

[0005] According to a self-service robot system provided by the present invention, the business services include human-computer interaction services, guidance services, business consultation services, and business processing services. In response to the human-computer interaction services, the guidance services, the business consultation services, or the business processing services, the robot initiates a robot scheduling request to the motion control module or the robot management platform.

[0006] According to a self-service robot system provided by the present invention, determining the scheduling task based on the robot scheduling request and the status and movement information of all the robots includes: Based on the robot scheduling request, a preset identifier is determined, and an initial screening is performed based on the preset identifier to obtain candidate robots that meet the robot scheduling request. The target robot is determined based on the type of the robot scheduling request and the status and movement information of the candidate robots; Based on the target robot and the destination and target business service corresponding to the robot scheduling request, the scheduling task is determined.

[0007] According to a self-service robot system provided by the present invention, the initial screening based on the preset identifier to obtain candidate robots that meet the robot scheduling request includes: If the business location is in business hours based on the preset identifier and the target business service is bound to the preset business point, then the robots in the robot list that meet the robot scheduling request are selected as the candidate robots based on the robot's battery level.

[0008] According to a self-service robot system provided by the present invention, the status information is used to characterize whether the robot's working status is normal, the types of services supported by the robot, and whether it is in a service processing state; the movement information is used to characterize the robot's movement status; and the step of determining a target robot based on the type of the robot scheduling request and the status and movement information of the candidate robots includes: When the robot scheduling request is of the general scheduling type, the candidate robot that is in normal working state, not in business processing state and not in outbound movement is selected as the target robot based on the status information and movement information of the candidate robot. If the robot scheduling request is of the type of business scheduling, based on the status information and movement information of the candidate robots, the candidate robots that are in normal working condition, support the business corresponding to the robot scheduling request, and are not in the business processing state or in outbound movement are selected as the target robots.

[0009] According to the present invention, a self-service robot system includes a perception module for acquiring customer gestures and / or facial expressions and environmental information in the application scenario. The interaction module integrates one or more of a speech recognition system, a natural language processing system, and a speech synthesis system. The interaction module is used to identify customer intentions based on customer voice information and environmental information in the application scenario, and to generate a robot scheduling request based on the customer intentions, including: The interaction module is used to identify the customer intent corresponding to the voice information based on the voice recognition system and the natural language processing system, and generate a robot scheduling request based on the customer intent if the customer intent is determined to be accurate based on the environmental information.

[0010] According to a self-service robot system provided by the present invention, the interaction module is further configured to determine the priority of the customer's intention based on the environmental information, and encapsulate the priority and the environmental information in the robot scheduling request. The robot management platform is further configured to adjust the scheduling strategy based on the priority and the environmental information in the robot scheduling request.

[0011] According to a self-service robot system provided by the present invention, the motion control module of the target robot is used to drive the target robot to the destination in response to the scheduling task, including: The motion control module of the target robot is used to respond to the scheduling task, determine the scheduling path based on the destination and the preset path planning algorithm, and optimize the scheduling path based on the preset path optimization algorithm during the process of driving the target robot to the destination based on the scheduling path until the destination is reached.

[0012] According to a self-service robot system provided by the present invention, the motion control module of the target robot is used to optimize the scheduling path based on a preset path optimization algorithm during the process of driving the target robot to the destination based on the scheduling path until the destination is reached, including: Acquire environmental point cloud information during robot movement, and based on the environmental point cloud information, determine whether the robot is in a degraded or non-degraded scene, and whether there are new static obstacles or dynamic targets on or around the scheduling path. When the robot is determined to be in a degraded scenario, the robot is controlled to move along the scheduled path. If a dynamic target is detected, the movement is paused and the robot continues to move after the dynamic target moves away from the scheduled path. When the robot is determined to be in a non-degenerate scenario, the robot is controlled to perform rolling optimization in the local time domain with the scheduling path as a reference. The optimization is achieved by solving for the minimum control parameters of the cost function.

[0013] The present invention also provides a self-service robot system control method for controlling the self-service robot system described in any of the above claims, the method comprising: The control interaction module or the customer-facing business system initiates a robot scheduling request; The robot management platform responds to the robot scheduling request. The robot management platform is used to determine the scheduling task based on the robot scheduling request and the status and movement information of all the robots. The scheduling task includes the target robot, the destination, and the target business service, and transmits the scheduling task to the target robot. The target robot is controlled to respond to the scheduling task. The motion control module of the target robot is used to drive the target robot to the destination so that the target robot can perform the target business service.

[0014] The self-service robot system and its control method provided by this invention include at least one robot and a robot management platform. The robot includes a motion control module, a business module, and an interaction module. The business module integrates all business services under a preset application scenario. The interaction module identifies the customer's intent based on the customer's voice information, or based on the customer's voice information and environmental information under the application scenario, and generates a robot scheduling request based on the customer's intent. The robot management platform responds to the robot scheduling request initiated by the business module or the customer-facing business system, and determines the scheduling task based on the robot scheduling request, as well as the status and movement information of all robots. The scheduling task represents the target robot, the destination, and the target business service. The robot management platform transmits the scheduling task to the target robot. The motion control module of the target robot responds to the scheduling task and drives the target robot to the destination so that the target robot can perform the target business service. The robot in this invention integrates all business services under a preset application scenario, and is uniformly scheduled by the robot management platform in response to robot scheduling requests. This eliminates the need to configure multiple robots or separate scheduling processes for each type of robot, avoiding increased deployment costs and resource redundancy. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is one of the structural block diagrams of the self-service robot system provided by the present invention.

[0017] Figure 2 This is the second structural block diagram of the self-service robot system provided by the present invention.

[0018] Figure 3 This is a schematic diagram of the locations of financial outlets provided by the present invention.

[0019] Figure 4 This is a schematic diagram of the robot scheduling process in one embodiment of the present invention.

[0020] Figure 5 This is a schematic diagram of the robot management process after the scheduling task is completed in one embodiment of the present invention.

[0021] Figure 6 This is a flowchart illustrating the self-service robot system control method provided by the present invention.

[0022] Figure label: 100 - Robot; 110 - Motion control module; 120 - Business module; 130 - Interaction module; 140 - Communication module; 200 - Robot management platform. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0024] Figure 1 This is one of the structural block diagrams of the self-service robot system provided by the present invention. Figure 2 This is the second structural block diagram of the self-service robot system provided by the present invention, as shown below. Figure 1 and Figure 2 As shown, the present invention provides a self-service robot system, including at least one robot 100 and a robot management platform 200. The robot 100 includes a motion control module 110, a business module 120 and an interaction module 130.

[0025] The business module 120 integrates all business services under preset application scenarios.

[0026] The interaction module 130 is used to identify customer intent based on customer voice information, or based on customer voice information and environmental information in the application scenario, and generate a robot scheduling request based on the customer intent.

[0027] The robot management platform 200 is used to respond to a robot 100 scheduling request initiated by the interaction module 130 or the customer-facing business system, and determine a scheduling task based on the robot 100 scheduling request and the status and movement information of all the robots 100. The scheduling task includes a target robot, a destination, and a target business service. The robot management platform 200 transmits the scheduling task to the target robot.

[0028] The motion control module 110 of the target robot is used to drive the target robot to the destination in response to the scheduling task, so that the target robot can perform the target business service.

[0029] The number of robots 100 is determined based on the application scenario, which includes industries such as finance, healthcare, and transportation, such as financial branches, airport halls, train stations, subway stations, shopping malls, hospitals, and other public and service locations. The robot management platform 200 includes a cloud platform or server. The confidentiality level of the robot management platform 200 is determined based on the confidentiality level of the application scenario. The robot management platform 200 performs unified scheduling and management of all or some of the robots 100 within the application scenario; this invention does not impose specific limitations on this.

[0030] The motion control module 110, business module 120, and interaction module 130 are all functional modules integrated into the robot 100. The motion control module 110, business module 120, interaction module 130, and robot management platform 200 can communicate with each other. When the robot 100 can provide services to customers, it can directly send a scheduling command to its own motion control module 110. The motion control module 110 responds to the scheduling command and drives the robot 100 to the location corresponding to the scheduling command. When the business module 120 cannot meet all customer needs (e.g., there are many customers and they require different services), the interaction module 130 sends a robot 100 scheduling request to the robot management platform 200. The robot 100 scheduling request includes customer needs and the destination corresponding to the customer needs. Customer needs include business needs and non-business needs. Business needs indicate that the customer needs non-general business services under the preset application scenario. Optionally, the robot 100 scheduling request also includes the customer's current location.

[0031] Taking a financial branch as an example, business services include interaction, greeting, business consultation, business processing, number taking, and guidance. Interaction, greeting, business consultation, number taking, and guidance are all general business services, while business processing is a non-general business service.

[0032] Customer-facing business systems are used to represent business systems deployed on non-robot 100 devices that provide business services in application scenarios, such as tablets, counter equipment, etc. When customers use customer-facing business systems, if they have any questions or need auxiliary resources for business processing, they can initiate robot 100 scheduling requests through the customer-facing business systems. Customer-facing business systems can also proactively initiate robot 100 scheduling requests when customers are processing business.

[0033] The robot management platform 200 receives scheduling requests from robot 100, parses the requests to obtain customer needs and their corresponding destinations, and determines the target service based on these needs. The robot management platform 200 obtains the status and movement information of all robots 100, or retrieves this information from a list of robots 100, which includes robot identifiers and periodically uploaded status and movement information. Based on this information, the robot management platform 200 selects target robots capable of fulfilling the scheduling requests, generates scheduling tasks based on the target robot, destination, and target service, and transmits these tasks to the target robots.

[0034] After receiving the scheduling task, the target robot parses the task to obtain the destination and target business service, and transmits the destination to the motion control module 110. The motion control module 110 drives the target robot to the destination. After arriving at the destination, the motion control module 110 sends arrival information, and the robot controls the business module to enter the target business and provide the target business service to the customer.

[0035] The robot management platform 200 can be deployed in a single application location within an application scenario, or it can share services and manage multiple application locations. Taking financial branches as an example, the robot management platform 200 can be deployed independently in a single branch or manage multiple branches. Status information reflects the business capabilities and status of robot 100. Robot 100's business capabilities refer to the services it currently supports; business status indicates whether robot 100 is currently processing a service (including: processing service, greeting, idle, charging). Mobility information reflects robot 100's mobility capabilities and mobility status. Robot 100's mobility capabilities indicate whether it can move when scheduled; mobility status indicates whether robot 100 is moving or stationary.

[0036] Before use, Robot 100 needs to scan the application site map, and the destination is one of the points set by the staff on the application site map.

[0037] Taking a financial branch as an example, before using Robot 100, it needs to scan the branch map. Robot 100's automatic movement range is within the map's area. After the map scan is complete, branch staff will select suitable locations on the map based on the branch layout, setting each location as a different service point in a many-to-many manner. Service point types include: business points, greeting points, customer rest points, charging points, and idle waiting points. Business points are locations for handling banking services, such as account opening and transactions; greeting points are locations that typically greet customers at the branch entrance; customer rest points are locations in the customer rest area, serving as the destination for Robot 100 when guiding customers to the rest area; charging points are locations where Robot 100 charges; and idle waiting points are locations where Robot 100 is available for summoning or direct use by customers. Service point types can be added, removed, or modified according to the actual needs of the branch.

[0038] Each service at the branch has a unique service code, and each service that can be processed on Robot 100 must be linked to a service point on the map, such as... Figure 3 As shown, location ① can be set as a "Account Opening" or "Deposit Certificate Opening" service point, location ② as a "Financial Transaction" or "Card Opening" service point, location ③ as a "Password Modification" service point, and location ④ as an idle waiting point. That is, one location can be set as a service point for multiple services, and multiple service points can be set up for the same service.

[0039] Optionally, the robot 100 further includes an interaction module 130, which integrates one or more of a speech recognition system, a natural language processing system, and a speech synthesis system. In a preferred embodiment, a voice assistant is provided for the interaction module to achieve intelligent voice control.

[0040] Correspondingly, after the target robot's motion control module responds to the scheduling task and drives the target robot to the destination, the motion control system will inform the robot assistant that it has arrived. The robot assistant will then inform the voice assistant that it has arrived, and the voice assistant will then notify the business system to enter the corresponding business process.

[0041] Optionally, the robot 100 further includes a communication module 140, which is used to realize internal communication within the robot 100 and communication between the robot 100 and the robot management platform 200. In a preferred embodiment, a robot assistant is provided for the communication module, which is used to control the communication module to realize internal communication within the robot 100 and communication between the robot 100 and the robot management platform 200.

[0042] Specifically, robot 100 includes a main control processor, a motion control processor, and an AI accelerator processor. Business module 120 is mounted on the main control processor, interaction module 130 is mounted on the AI ​​accelerator processor, motion control module 110 is mounted on the motion control processor, and communication module 140 may include CAN bus, EtherCAT bus, PCIE interface, 4G / 5G / 6G communication module 140, WiFi communication module 140, serial communication module, etc. The main control processor is connected to the AI ​​accelerator processor through PCIE interface, connected to the motion control processor through CAN bus and / or EtherCAT bus, and connected to robot management platform 200 through 4G / 5G / 6G communication module 140 and / or WiFi communication module 140. The internal communication module, interaction module, and motion control module of the robot can communicate internally through the serial communication module.

[0043] The interaction module 130 also integrates a central control system for controlling communication; a speech recognition system for converting customer speech into text based on Automatic Speech Recognition (ASR) technology; a natural language processing system for performing semantic analysis on the text recognized by the speech recognition system based on Natural Language Processing (NLP) technology to understand the meaning expressed by the customer, identify the customer's intent, provide the required service (i.e., the target business service) based on the customer's intent, and output a response based on the target business service; and a speech synthesis system for converting the response output by the natural language processing system into a speech response based on Text-To-Speech (TTS) technology.

[0044] By integrating the interaction module 130 locally into the robot 100, the robot 100 is equipped with local voice interaction and intent recognition functions, enabling it to receive and process customer voice and autonomously generate robot scheduling requests. Preferably, when generating robot scheduling requests, the robot scheduling strategy can be made more intelligent by comprehensively considering the user's intent priority and real-time environmental information and introducing multi-dimensional optimization factors.

[0045] It is understood that the robot 100 in this invention integrates all business services under the preset application scenario, and is uniformly scheduled by the robot management platform 200 in response to the scheduling request of the robot 100. There is no need to configure multiple robots 100, and no need to configure exclusive scheduling processes for each type of robot 100, thus avoiding the problems of increased deployment costs and resource redundancy.

[0046] As an example, the business services include human-computer interaction services, guidance services, business consultation services, and business processing services. In response to the human-computer interaction services, the guidance services, the business consultation services, or the business processing services, the robot initiates a robot 100 scheduling request to the motion control module 110 or the robot management platform 200.

[0047] Human-computer interaction services are implemented through touchscreen and / or voice interaction. Touchscreen interaction involves receiving customer commands and providing feedback via the touchscreen. Voice interaction can be applied to any of the following services: guidance, business consultation, and business processing. During the process of providing services to customers, Robot 100 can interact with customers entirely through voice to complete the transaction.

[0048] Taking a financial branch scenario as an example, Robot 100 can provide customers with voice-activated number retrieval services during the guidance, business consultation, and business processing processes. During guidance, customers can directly say "I need a number" to Robot 100, or say they want to conduct XX business, such as "I want to apply for a card." Robot 100 checks if it supports this business; if so, it directly proceeds to the business; otherwise, it provides the number retrieval service. During business consultation, after a customer inquires about a certain business, Robot 100 will ask if the customer needs a number. If the customer replies yes, the number retrieval service will be provided. During business processing, when a customer needs to retrieve a number to conduct other business after completing the current business on Robot 100, they can say "I need a number" or "I want to conduct XX business" on the menu page, and Robot 100 will provide the number retrieval service.

[0049] The guidance service includes greeting and location assistance. Taking a financial branch as an example, when Robot 100 is positioned at the branch entrance, it automatically wakes up its screen upon detecting a customer via a human body sensor and camera. It then proactively greets the customer with a voice message, inquires about their needs, and provides services such as business consultation and ticket dispensing. Robot 100 can also provide directions for customers unfamiliar with the branch. This function can be triggered directly by the customer's verbal request, by operating the Robot 100's screen, or by other electronic devices or systems. If Robot 100 is not in a business-related state after being woken up, the customer can say "Take me to XX" or "I want to go to XX" to the Robot 100. For example, if the customer says "Take me to the rest area," Robot 100 will lead the customer to the rest area; if the customer says "I want to go to the self-service area," Robot 100 will lead the customer to the self-service area. Customers can also choose to directly operate the Robot 100's screen, select the desired location, and issue a guidance command. Furthermore, it supports triggering through customer-facing business systems, such as issuing guidance commands via a PAD.

[0050] Business consultation service refers to Robot 100 answering customers' business-related questions. Taking a financial branch scenario as an example, after Robot 100 is activated, customers can consult with it in a conversational manner about business-related issues, such as how to open an account or change a password. During the business consultation process, when Robot 100 recognizes a business transaction, it can also proactively inquire about the customer's need for a number and guide them to take a number and wait in line.

[0051] When providing business processing services, Robot 100 is in a business processing state, supporting two business processing modes: Robot 100 working with customer-facing business systems and Robot 100 operating independently. The Robot 100 working with customer-facing business systems can be understood as the "Robot 100+" mode, which breaks down traditional business processes. Some processes are completed on Robot 100, while others are completed on other devices. For example, opening a deposit certificate usually requires completing a lot of forms. After arriving at the branch, the customer can directly use a PAD to access the deposit certificate opening service, complete the forms, and then click the "Certificate Issuance" button on the page to summon Robot 100 to process the subsequent deposit certificate issuance. After summoning Robot 100, it will proactively go to the customer's branch to provide the service. Another example is for more complex counter services, where more complex processes are completed at the counter, while simpler processes (such as medium retrieval) are completed on Robot 100, improving the efficiency of counter processing. Robot 100 operating independently means that all business processes are completed on Robot 100.

[0052] Based on the above explanation of each service, it can be understood that the robot, in response to the human-computer interaction service, the guidance service, the business consultation service, or the business processing service, initiates a robot 100 scheduling request to the motion control module 110 or the robot management platform 200. Specifically, when a customer makes a request to the robot 100, the robot 100 determines which of the following services the customer's request belongs to: the human-computer interaction service, the guidance service, the business consultation service, or the business processing service; whether it needs assistance from other robots 100; and whether it needs to move. If movement is required and no assistance from other robots 100 is needed, then the robot 100 initiates a robot 100 scheduling request to the motion control module 110; if assistance from other robots 100 is required, then the robot 100 scheduling request is initiated to the robot management platform 200. For example, if multiple customers simultaneously require guidance services but have different destinations, they need to initiate a robot 100 scheduling request to the robot management platform 200.

[0053] It is understood that this invention integrates services such as voice interaction, business consultation, business processing, voice number retrieval, greeting, and location guidance, solving the problem of single and redundant deployment of Robot 100. By cooperating with the customer-facing business system, Robot 100 improves the customer experience through business processing methods that combine business segmentation and multiple channels. Robot 100 proactively provides services to customers, greatly facilitating customers with mobility impairments and elderly customers.

[0054] As one embodiment, determining the scheduling task based on the robot 100 scheduling request, and the status and movement information of all the robots 100, includes: Based on the robot 100 scheduling request, a preset identifier is determined, and an initial screening is performed based on the preset identifier to obtain candidate robots 100 that meet the robot 100 scheduling request. The target robot is determined based on the type of the robot 100 scheduling request and the status and movement information of the candidate robot 100; Based on the target robot and the destination and target business service corresponding to the scheduling request of the robot 100, the scheduling task is determined.

[0055] The preset identifier serves as a unique identifier for the application location within the application scenario, used to determine whether the application location is in normal working condition. The robot 100 dispatch request carries the preset identifier, or the parameters carried in the robot 100 dispatch request have a unique mapping relationship with the preset identifier. Therefore, the robot management platform 200 can determine the preset identifier through the robot 100 dispatch request. The robot management platform 200 can also determine whether the application location is in normal working condition based on the preset identifier. If so, it filters out all robots 100 within the application location that are fault-free or have a battery level higher than a preset value, and uses these filtered robots 100 as candidate robots 100, completing the initial screening.

[0056] The types of robot 100 scheduling requests include general scheduling and business scheduling. General scheduling means that the robot 100 is summoned directly through a separate general summoning function without being summoned in the business process, and the destination point is selected. It does not carry business parameters. Business scheduling means that the robot 100 is summoned in the business process according to the current business needs during the business process. The summoning information carries business parameters.

[0057] Determining a target robot based on the type of the robot 100 scheduling request and the status and movement information of the candidate robots 100 includes performing a secondary screening based on the type of the robot 100 scheduling request and the status and movement information of the candidate robots 100 to select the target robot from the candidate robots 100.

[0058] Optionally, the initial screening based on the preset identifier to obtain candidate robots 100 that meet the scheduling request of robot 100 includes: If the business location is in business hours based on the preset identifier and the target business service is bound to the preset business point, the robot 100 that meets the robot 100 scheduling request in the robot 100 list is selected as the candidate robot 100 based on the robot 100's battery level.

[0059] If the business location does not exist, is not in business hours, or the target business service is not bound to a preset business point, the process ends.

[0060] Optionally, the status information is used to characterize whether the working status of the robot 100 is normal, the types of services supported by the robot 100, and whether it is in a service processing state; the movement information is used to characterize the movement status of the robot 100; and the step of determining the target robot based on the type of the robot 100 scheduling request and the status and movement information of the candidate robots 100 includes: When the type of the robot 100 scheduling request is general scheduling, based on the status information and movement information of the candidate robot 100, the candidate robot 100 that is in normal working state, not in business processing state and not in motion is selected as the target robot. If the type of the robot 100 scheduling request is a business scheduling request, based on the status information and movement information of the candidate robot 100, the candidate robot 100 that is in normal working condition, supports the business corresponding to the robot 100 scheduling request, and is not in a business processing state and is not in motion is selected as the target robot.

[0061] like Figure 4 As shown, taking a financial branch as an example, the robot 100 scheduling process includes two rounds of screening. The first round initially screens out multiple robots 100 that may meet the summoning requirements, and the second round screens out one robot 100 that meets the summoning requirements from the robots 100 screened in the first round.

[0062] The first round of screening includes: determining whether the requested branch exists based on the organization number; if yes, continue; otherwise, end and cannot be summoned; determining whether the branch is currently in business hours; if yes, continue; otherwise, end and cannot be summoned; determining whether the service initiating the summoning has been bound to a service point; if yes, continue; otherwise, end and cannot be summoned; and selecting all robots 100 in the robot 100 list whose battery capacity meets the service requirements as candidate robots 100.

[0063] The second round of screening includes: determining whether the hardware status and map information of the current robot 100 are normal; if so, this requirement is met; determining whether the current call type is a general call; if so, checking whether the current robot 100 is in operation (the robot 100 is in operation when it is handling banking services such as taking a number, opening a card, or buying wealth management products); otherwise, this requirement is met; otherwise, based on the requested service, determining whether the current robot 100's business environment supports handling the service; if so, this requirement is met. For card issuance, it needs to be determined whether robot 100 has a card; for bill issuance, it needs to be determined whether robot 100 has a bill, etc. It checks whether robot 100 is currently engaged in business (robot 100 is in a business state when it is handling banking services such as ticket issuance, card issuance, or wealth management purchases); otherwise, this requirement is met. It checks whether robot 100 has been scheduled for other business operations; otherwise, this requirement is met. It checks whether robot 100 is currently moving; if so, a termination command is issued. If termination is successful, this requirement is met, and the process proceeds to the next step; if termination fails, the current screening ends, robot 100 is removed, and the process moves to the next robot 100. Otherwise, this requirement is met, and the process proceeds to the next step. When a robot 100 meets all the above conditions, it will be locked and summoned to the business location to handle the business.

[0064] like Figure 5 As shown, after robot 100 completes its task, robot management platform 200 comprehensively schedules robot 100 to return based on information such as robot 100's battery level and location. The task completion scheduling logic includes: determining whether it is working time; if it is not working time, it is directly dispatched back to the charging point; determining robot 100's battery level; if robot 100's battery level is lower than the set service battery level and robot 100 is not charging, it is dispatched back to the charging point; and determining whether robot 100 is at the welcoming point; if not, robot 100 is dispatched back to the welcoming point.

[0065] It is understood that the present invention can improve screening efficiency by determining the target robot 100 through two screenings. By selecting the target robot based on the type of scheduling request of robot 100 and the status and movement information of candidate robots 100, the accuracy of screening and determining the target robot is high, which shortens the customer waiting time and improves the customer experience.

[0066] As one embodiment, the robot includes a perception module for acquiring customer gestures and / or facial expressions, as well as environmental information in the application scenario. The interaction module integrates one or more of a speech recognition system, a natural language processing system, and a speech synthesis system. The interaction module is used to identify customer intentions based on customer voice information and environmental information in the application scenario, and to generate a robot scheduling request based on the customer intentions, including: The interaction module is used to identify the customer intent corresponding to the voice information based on the voice recognition system and the natural language processing system, and generate a robot scheduling request based on the customer intent if the customer intent is determined to be accurate based on the environmental information.

[0067] Optionally, the perception module includes a camera and sensors. The camera is used to acquire the customer's gestures and / or facial expressions, and the sensors are used to acquire environmental information in the application scenario, such as weather, temperature, and humidity. The camera can also be used to acquire environmental images of the application scenario, enabling the robot to understand the environment in the application scenario from multiple dimensions.

[0068] In one embodiment, determining the accuracy of the customer's intent based on the environmental information can be achieved by judging the consistency between the environmental information and the customer's intent. For example, if the interaction module recognizes the customer's voice as "I need hot water," and the sensor detects a low temperature in the application scenario (e.g., below zero), then the environmental information is determined to be consistent with the customer's intent, i.e., the customer's intent is accurate. Preferably, a consistency database of customer intent and environmental information can be pre-built. After extracting the customer intent and environmental information, a search and comparison with the consistency database is performed to determine the consistency between the customer intent and the environmental information.

[0069] By incorporating a perception module, the robot can recognize customers' gestures and facial expressions, and combine this with current time, location, and even weather information to comprehensively assess the urgency and true needs of the customer's intentions. For example, when a customer says "I need help" and points to spilled liquid on the ground, the robot can recognize the keywords through voice recognition, confirm the scene through a camera, and automatically generate a cleaning task scheduling request, rather than a general help scheduling request. Another example is providing towels to customers entering the application scenario on rainy days.

[0070] In another embodiment, the confidence levels of the speech recognition results, the customer's gestures and / or facial expressions, and environmental information can be weighted and fused to calculate a comprehensive confidence level to determine the accuracy of the customer's intent. If the comprehensive confidence level is greater than or equal to a confidence threshold, the customer's intent is deemed accurate. The weights for the speech recognition results, customer gestures and / or facial expressions, and environmental information can be 0.6, 0.3, and 0.1, respectively, and the confidence threshold can be 0.85.

[0071] For example, the interaction module associates the voice recognition result (such as 'I need hot water') with the 'customer rubbing hands' captured by the camera and the 'ambient temperature 18℃' captured by the sensor. Through a weighted fusion algorithm (such as voice intent weight 0.6, behavioral feature weight 0.3, and environmental data weight 0.1), when the overall confidence level is ≥85%, the customer's intent is determined to be accurate.

[0072] Furthermore, the weights and confidence thresholds can be adjusted based on environmental information. For example, in noisy environments (sensor-detected noise ≥60 dB), the speech recognition weight is reduced to 0.4, the behavioral feature weight is increased to 0.4, and the environmental data weight is increased to 0.2, while the confidence threshold is lowered to 80%. In quiet environments, the initial weights and thresholds are maintained to improve the adaptability to different scenarios.

[0073] Furthermore, when the overall confidence level is low, the robot can proactively initiate a secondary interaction to confirm the customer's intent. In addition, to avoid unnecessarily consuming system resources, after initiating the interaction, it is determined whether the termination condition for the secondary interaction has been met. If so, the secondary interaction is stopped.

[0074] For scenarios with low confidence (e.g., overall confidence < 85%), the interaction module automatically initiates a second inquiry, with the content optimized based on environmental information. For example, if the voice recognition result is "I need help," but no clear behavior or environmental anomaly is detected, the robot proactively asks, "Do you need items (such as hot water or tissues), or are you feeling unwell and need assistance?" The termination condition for the second interaction can be: "If the customer does not respond twice consecutively, or explicitly refuses further interaction, the scheduling request generation is terminated, and the interaction log is recorded for backend analysis."

[0075] It is understood that by introducing a perception module, this invention enables multimodal interaction between the robot and the customer. It can combine the customer's gestures and / or facial expressions with environmental information in the application scenario to judge the identified customer's intentions and reduce the error rate of robot scheduling requests.

[0076] As an example, the interaction module is further configured to determine the priority of the customer's intent based on the environmental information, and encapsulate the priority and the environmental information in the robot scheduling request. The robot management platform is further configured to adjust the scheduling strategy based on the priority and the environmental information in the robot scheduling request.

[0077] Optionally, the present invention pre-establishes initial priorities for various customer intentions. The interaction module can adjust the initial priority of the customer intention based on environmental information to obtain the priority of the customer intention. Specifically, the interaction module can determine whether an abnormal event with a higher priority than the customer intention has occurred based on environmental information. If so, the initial priority of the customer intention is reduced.

[0078] Environmental information can include customer physiological status (such as posture and heart rate), sudden scene signals (such as fire alarms and equipment malfunctions), and public safety signals (such as emergency broadcasts), collected through devices such as cameras, infrared sensors, and scene-linked sensor networks mounted on the robot. Regarding customer physiological status, abnormal events refer to events that reach abnormal physiological indicators, such as abnormal customer posture, which means the customer remains still for ≥5 seconds after falling or the body tilt angle is >45°. Preferably, after detecting an abnormal customer posture, the robot performs secondary confirmation through voice inquiry (such as "Do you need help?") and infrared sensor detection of vital signs (such as breathing fluctuations), avoiding misjudging normal behaviors such as bending over to pick up items or lying down to rest as falls.

[0079] The priorities of the above environmental information and customer intent can be divided into emergency life safety level (such as falls, sudden illness), public emergency level (such as fire, earthquake early warning), routine service level (such as water supply, food delivery), low priority (such as consulting non-emergency information), etc. Each level corresponds to a fixed priority value (such as levels 1-4, the smaller the value, the higher the priority). The initial priority is preset according to the business type (such as level 3 for routine service level).

[0080] The interaction module adjusts the initial priority based on environmental information, including but not limited to increasing the priority of customer intents corresponding to emergency events and decreasing the priority of regular service intents interrupted by emergency events. When an emergency life-safety level event is detected, the priority value is directly adjusted to level 1; when a public emergency level event is detected, it is adjusted to level 2, and all level 3-4 tasks are suspended, reducing the priority of customer intents. After determining the priority, the interaction module needs to send a task priority change notification to the business module. The business module suspends the resource occupation of currently low-priority services (such as stopping path planning and releasing occupied hardware interfaces) to reserve resources for high-priority tasks. It should be noted that the priority adjustment has an expiration period. The priority adjustment for public emergency level events is valid until the event is resolved (such as when a fire alarm is deactivated), while the priority for emergency life-safety level events continues until staff confirm the customer's safety, to avoid long-term priority occupation that could prevent regular services from being unrecoverable.

[0081] The scheduling strategy adjustment scheme of the robot management platform includes: after receiving a scheduling request with priority, if the current target robot is executing a low-priority task, immediately send a task interruption command and re-plan the execution path of the high-priority task; if there are no idle robots, prioritize interrupting the robot performing the low-priority task, or schedule the nearest robot to handle the high-priority task.

[0082] For example, when receiving a customer's voice command "Please fetch me water," if the system detects an abnormal pattern of the customer suddenly falling and remaining motionless, it identifies the "fall" as a potentially life-threatening emergency. This emergency has a much higher priority than regular service tasks, and the "water fetching" task is temporarily suspended. Suspending the water fetching task requires recording the current progress (e.g., planned route, not yet started). After the emergency is resolved, the robot management platform decides whether to resume execution or notify the customer to re-initiate the request based on the remaining task queue and the robot's current status (e.g., whether it is idle, battery level).

[0083] It is understood that this invention determines the priority of customer intent through environmental information and encapsulates the priority and environmental information in the robot scheduling request. The robot management platform is also used to adjust the scheduling strategy based on the priority and environmental information in the robot scheduling request. By comprehensively considering the user intent priority and real-time environmental information, multi-dimensional optimization factors are introduced. Through dynamic adjustment of priority and linkage with environmental information, priority response to urgent tasks is achieved, safety risks are reduced, and the flexibility and reliability of services are improved.

[0084] As one embodiment, the motion control module 110 of the target robot, in response to the scheduling task, drives the target robot to the destination, including: The motion control module 110 of the target robot is used to respond to the scheduling task, determine the scheduling path based on the destination and the preset path planning algorithm, and optimize the scheduling path based on the preset path optimization algorithm during the process of driving the target robot to the destination based on the scheduling path until the destination is reached.

[0085] Optionally, determining the scheduling path based on the destination and a preset path planning algorithm includes: The system invokes a pre-built Venn diagram topology network and calculates the optimal global path from the target robot's current position to its destination based on a search algorithm. This optimal global path is then used as the scheduling path. The goal of the search algorithm is to maximize the distance to the nearest obstacle, ensuring that the optimal global path has a central characteristic and maximizes the safe passage space.

[0086] A Voronoi diagram is a spatial partitioning method based on distance metrics. It divides a plane into a series of convex polygonal cells, each cell associated with a given station (a point in a set of stations) such that the distance from any point within a cell to that station is less than the distance to all other stations, thus forming a neighborhood centered on the station. The edges of the Voronoi diagram (i.e., cell boundaries) constitute a continuous topological structure, where the adjacency relationships between cells define an undirected graph network. Each cell serves as a topological unit, and their shared boundaries represent the proximity between stations. It is widely used in problems such as proximity queries, path planning, and resource allocation.

[0087] Specifically, the Venograph topology network is constructed based on the following steps: construct an environmental grid map of the application site, and use the outlines of static obstacles (walls, counters, furniture, etc.) in the environmental grid map as generation points to construct the Venograph topology network. The Venograph topology network consists of several ridges (a set of points that are equidistant from the nearest obstacles on both sides) and vertices.

[0088] Optionally, the path optimization algorithm includes an obstacle perception algorithm and a relocation algorithm. Correspondingly, the motion control module of the target robot is used to optimize the scheduled path based on a preset path optimization algorithm during the process of driving the target robot to the destination based on the scheduled path, until the destination is reached, including: Based on multiple sensors including LiDAR, depth cameras, Time-of-Flight (TOF) laser arrays, and ultrasound, the system acquires environmental point cloud information during robot movement. It then determines whether the robot is in a degraded scene (such as a long, narrow, enclosed area like a corridor) or a non-degraded scene (such as a wide, open area like a hall). Simultaneously, it identifies any new static obstacles or dynamic targets on or around the robot's path. Specifically, LiDAR and depth cameras are used for detecting distant targets, depth cameras and TOF laser arrays are used for detecting transparent, low-lying obstacles at mid-to-near distances, and ultrasound is used for detecting nearby obstacles. When the robot is determined to be in a degraded scenario, the robot is controlled to move along the scheduled path. If a dynamic target is detected, the movement is paused and the robot continues to move after the dynamic target moves away from the scheduled path. When the robot is determined to be in a non-degenerate scenario, the robot is controlled to perform rolling optimization in the local time domain with the scheduling path as a reference. The optimization is achieved by solving for the minimum control parameters (robot speed and angular velocity) of the cost function.

[0089] The cost function expression is as follows: C total =ω path *C path +ωo *C obs +ω d *C dyn +ω s *C smooth ; Among them, C total C represents the total cost. path ω represents the path-following deviation cost, penalizing the deviation of the trajectory from the global reference path (or local target point). path C represents the path following deviation cost weight. obs This represents the cost of distance to static obstacles, penalizing trajectory points that are too close to all static obstacles, based on real-time perception updates, ω. o C represents the cost weight for static obstacle distances. dyn Represents the collision prediction cost for dynamic obstacles. Based on the predicted trajectory of the dynamic obstacle, the probability of the robot's future trajectory colliding with it is calculated and determined. ω d C represents the weight of the dynamic obstacle prediction collision cost. smooth Represents the cost of motion smoothness, penalizing drastic changes in velocity and acceleration, ω s This represents the cost weight for motion smoothness.

[0090] It is important to note that the robot automatically calibrates the extrinsic parameters of the dual LiDAR during startup and movement, and performs real-time compensation for point cloud distortion caused by the robot's own movement to ensure the accuracy of environmental point cloud information. In addition, a jump-point downsampling strategy is adopted when processing environmental point cloud information to reduce the computing power required for point cloud processing while ensuring that features are not lost. Furthermore, a dual ground point cloud filtering algorithm based on ground segmentation and normal vectors is used to filter the downsampled environmental point cloud information, remove ground point clouds, and prevent ground textures, reflections, or slopes from being misidentified as obstacles.

[0091] It is understood that the present invention generates scheduling paths based on scheduling tasks, realizes on-demand planning triggered by events, greatly reduces problems such as robot swaying, and optimizes the scheduling paths based on preset path optimization algorithms to improve the scheduling safety of robots.

[0092] The self-service robot system control method provided by the present invention is described below. The self-service robot system control method described below can be referred to in correspondence with the self-service robot system described above.

[0093] like Figure 6As shown, the present invention also provides a self-service robot system control method for controlling the self-service robot system described in any of the above claims. The self-service robot system includes at least one robot and a robot management platform. The robot includes a motion control module, a business module, and an interaction module. The business module integrates all business services under a preset application scenario. The method includes: Step S100: The control interaction module or the customer-facing business system initiates a robot scheduling request; Step S200: Control the robot management platform to respond to the robot scheduling request. The robot management platform is used to determine the scheduling task based on the robot scheduling request and the status information and movement information of all the robots. The scheduling task includes the target robot, the destination and the target business service, and transmits the scheduling task to the target robot. Step S300: Control the target robot to respond to the scheduling task. The motion control module of the target robot is used to drive the target robot to the destination so that the target robot can perform the target business service.

[0094] The embodiments of the self-service robot system control method provided by the present invention can refer to the above-described self-service robot system, and its technical effects are also corresponding to those of the self-service robot system, so they will not be repeated here.

[0095] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A self-service robot system, characterized in that, It includes at least one robot and a robot management platform, wherein the robot includes a motion control module, a business module and an interaction module; The business module integrates all business services under the preset application scenarios; The interaction module is used to identify customer intent based on customer voice information, or based on customer voice information and environmental information in the application scenario, and generate a robot scheduling request based on the customer intent. The robot management platform is used to respond to robot scheduling requests initiated by the interaction module or the customer-facing business system, and determine scheduling tasks based on the robot scheduling requests and the status and movement information of all the robots. The scheduling task includes a target robot, a destination, and a target business service. The robot management platform transmits the scheduling task to the target robot. The motion control module of the target robot is used to drive the target robot to the destination in response to the scheduling task, so that the target robot can perform the target business service.

2. The self-service robot system according to claim 1, characterized in that, The business services include human-computer interaction services, guidance services, business consultation services, and business processing services. In response to the human-computer interaction services, the guidance services, the business consultation services, or the business processing services, the robot initiates a robot scheduling request to the motion control module or the robot management platform.

3. The self-service robot system according to claim 1, characterized in that, The step of determining the scheduling task based on the robot scheduling request, as well as the status and movement information of all the robots, includes: Based on the robot scheduling request, a preset identifier is determined, and an initial screening is performed based on the preset identifier to obtain candidate robots that meet the robot scheduling request. The target robot is determined based on the type of the robot scheduling request and the status and movement information of the candidate robots; Based on the target robot and the destination and target business service corresponding to the robot scheduling request, the scheduling task is determined.

4. The self-service robot system according to claim 3, characterized in that, The initial screening based on the preset identifier to obtain candidate robots that meet the robot scheduling request includes: If the business location is in business hours based on the preset identifier and the target business service is bound to the preset business point, then the robots in the robot list that meet the robot scheduling request are selected as the candidate robots based on the robot's battery level.

5. The self-service robot system according to claim 3, characterized in that, The status information is used to characterize whether the robot's working status is normal, the types of services the robot supports, and whether it is in a service processing state. The movement information is used to characterize the robot's movement status. Determining the target robot based on the type of the robot scheduling request and the status and movement information of the candidate robots includes: When the robot scheduling request is of the general scheduling type, the candidate robot that is in normal working state, not in business processing state and not in outbound movement is selected as the target robot based on the status information and movement information of the candidate robot. If the robot scheduling request is of the type of business scheduling, based on the status information and movement information of the candidate robots, the candidate robots that are in normal working condition, support the business corresponding to the robot scheduling request, and are not in the business processing state or in outbound movement are selected as the target robots.

6. The self-service robot system according to any one of claims 1 to 5, characterized in that, The robot includes a perception module for acquiring customer gestures and / or facial expressions, as well as environmental information in the application scenario. The interaction module integrates one or more of a speech recognition system, a natural language processing system, and a speech synthesis system. The interaction module is used to identify customer intentions based on their voice information and environmental information in the application scenario, and to generate a robot scheduling request based on the customer intentions, including: The interaction module is used to identify the customer intent corresponding to the voice information based on the voice recognition system and the natural language processing system, and generate a robot scheduling request based on the customer intent if the customer intent is determined to be accurate based on the environmental information.

7. The self-service robot system according to any one of claims 6, characterized in that, The interaction module is further configured to determine the priority of the customer's intent based on the environmental information, and encapsulate the priority and the environmental information in the robot scheduling request. The robot management platform is further configured to adjust the scheduling strategy based on the priority and the environmental information in the robot scheduling request.

8. The self-service robot system according to any one of claims 1 to 5, characterized in that, The motion control module of the target robot, used to drive the target robot to the destination in response to the scheduling task, includes: The motion control module of the target robot is used to respond to the scheduling task, determine the scheduling path based on the destination and the preset path planning algorithm, and optimize the scheduling path based on the preset path optimization algorithm during the process of driving the target robot to the destination based on the scheduling path until the destination is reached.

9. The self-service robot system according to claim 8, characterized in that, The motion control module of the target robot is used to optimize the scheduling path based on a preset path optimization algorithm during the process of driving the target robot to the destination based on the scheduling path, until the destination is reached, including: Acquire environmental point cloud information during robot movement, and based on the environmental point cloud information, determine whether the robot is in a degraded or non-degraded scene, and whether there are new static obstacles or dynamic targets on or around the scheduling path. When the robot is determined to be in a degraded scenario, the robot is controlled to move along the scheduled path. If a dynamic target is detected, the movement is paused and the robot continues to move after the dynamic target moves away from the scheduled path. When the robot is determined to be in a non-degenerate scenario, the robot is controlled to perform rolling optimization in the local time domain with the scheduling path as a reference. The optimization is achieved by solving for the minimum control parameters of the cost function.

10. A self-service robot system control method, used to control the self-service robot system according to any one of claims 1 to 9, characterized in that, The method includes: The control interaction module or the customer-facing business system initiates a robot scheduling request; The robot management platform responds to the robot scheduling request. The robot management platform is used to determine the scheduling task based on the robot scheduling request and the status and movement information of all the robots. The scheduling task includes the target robot, the destination, and the target business service, and transmits the scheduling task to the target robot. The target robot is controlled to respond to the scheduling task. The motion control module of the target robot is used to drive the target robot to the destination so that the target robot can perform the target business service.

Images