Voice control task processing method based on artificial intelligence

Abstract

An embodiment of the present invention relates to a voice control task processing method based on artificial intelligence, including: receiving broadcast signals of multiple beacons and recording the signal strength of each broadcast signal; determining the current position of a cleaning terminal according to the signal strength; processing a first voice message to determine the task category; when it is the first category, sending query information to a server according to the current position; when there is a task to be completed, receiving the task information to be completed sent by the server; generating a first task broadcast information according to the task name, task time and task requirements; broadcasting the task according to the first task broadcast information; receiving a second voice message; the second voice message is a response to any task in the first task broadcast information; processing the second voice message to determine the intention information; and executing the corresponding task according to the intention information.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a voice control task processing method based on artificial intelligence. Background Art

[0002] With the continuous development of science and technology, people's demand for time management and task management is increasing. Traditional watches are mainly used to display time, with relatively single functions, and can no longer meet the diverse needs of modern people.

[0003] In today's fast-paced life, people often need a convenient and efficient way to manage their tasks and schedules. The emergence of voice control technology provides a new way to solve this problem. Through voice control, users can operate the device more conveniently without manual operation, which improves the efficiency of use.

[0004] At the same time, the task punch card management system can help people better plan their time and tasks and improve work and study efficiency. Combining voice control technology with the task punch card management system can provide users with more intelligent and convenient services.

[0005] At present, although there are some task management software and smart watches, products that perfectly integrate voice control with task punch-in management systems are still relatively rare. Therefore, there is an urgent need for a voice control task processing method based on artificial intelligence. Summary of the invention

[0006] The purpose of the present invention is to provide a voice control task processing method based on artificial intelligence to solve the problems existing in the prior art.

[0007] To achieve the above object, the present invention provides a first aspect of a method for processing a voice control task based on artificial intelligence, the method comprising:

[0008] receiving broadcast signals of a plurality of beacons and recording the signal strength of each broadcast signal; the broadcast signal includes a broadcast signal ID;

[0009] Determining the current location of the cleaning terminal according to the signal strength;

[0010] Receiving a first voice message; the first voice message includes an area name;

[0011] Processing the first voice information to determine a task category; the task category includes a first category, a second category and a third category;

[0012] When it is the first category, query information is sent to the server according to the current location; the query information is used to instruct the server to query whether there is a task to be completed in the area where the current location is located;

[0013] When there are tasks to be completed, receiving task information to be completed sent by the server; the task information to be completed includes task name, task time and task requirements;

[0014] Generate first task broadcast information according to the task name, task time and task requirements;

[0015] Performing task broadcasting according to the first task broadcasting information;

[0016] receiving a second voice message; the second voice message being a response to any task in the first task broadcast message;

[0017] Processing the second voice information to determine intention information;

[0018] According to the intention information, the corresponding task is executed.

[0019] In a possible implementation, determining the current position of the cleaning terminal according to the signal strength specifically includes:

[0020] According to the positions of the beacons corresponding to the multiple signal strengths, the current position of the cleaning terminal is determined by a positioning algorithm.

[0021] In a possible implementation manner, the processing the second voice information to determine the intention information specifically includes:

[0022] Converting the second voice information to obtain corresponding text;

[0023] Performing part-of-speech tagging on the text;

[0024] Perform syntactic analysis on the text after part-of-speech tagging to generate an embedded representation of the sentence;

[0025] Match the embedded representation with a predefined instruction set and calculate the similarity based on cosine similarity or dot product;

[0026] Based on the similarity, determine the user intent.

[0027] In a possible implementation manner, the processing the second voice information to determine the intention information specifically includes:

[0028] When the second voice information is an ambiguous voice command, acquiring a plurality of clean areas in the voice information before the second voice information;

[0029] Using the multiple clean areas and the current position as inputs of a Bayesian network to calculate a posterior probability that each clean area is referred to;

[0030] Select the clean area with the highest posterior probability as the inferred reference area;

[0031] Generate second task broadcast information according to the cleaning area; the second task broadcast information is used to inquire whether to use the cleaning area as a reference area;

[0032] receiving a third voice message for the second task broadcast information; the third voice message is used to indicate whether to use the cleaning area as a reference area;

[0033] If the cleaning area is taken as the reference area, the task of the reference area is executed.

[0034] In a possible implementation, the method further includes:

[0035] When the first voice information is of the second category, determining the query content according to the first voice information; the query content includes a cleaning standard query, a remaining task query, and a future task query;

[0036] When the task category is the second category, if the query content is a cleaning standard query or a future task query, generating third broadcast information, and broadcasting the third broadcast information;

[0037] When the query content is a query of remaining tasks, the number of completed tasks corresponding to the area name is queried; the number of completed tasks includes the completion steps of a task or the number of completed tasks among multiple tasks;

[0038] Compare the number of completed tasks with a preset total number to determine the completion ratio; the preset total number includes the total steps of a task or the number of multiple tasks;

[0039] According to the completion ratio, fourth task broadcast information is generated; the fourth task broadcast information is used to indicate the progress of the task.

[0040] In a possible implementation, the method further includes:

[0041] When the first voice information is of the third category, the first voice information is sent to a server to feedback an abnormal situation through the first voice information.

[0042] In a possible implementation, the method further includes:

[0043] Receive the wake-up word;

[0044] According to the wake-up word, start the intercom function;

[0045] Querying a preset cleaning terminal list, and determining a cleaning terminal in the cleaning terminal list that has an intercom function activated;

[0046] The fourth voice message is sent to the cleaning terminal with the intercom function activated.

[0047] In a possible implementation, the method further includes:

[0048] receiving a fifth voice message, where the fifth voice message is used to instruct to establish a group or join a group;

[0049] When the fifth voice message is for establishing a group, establishing and saving the group according to the fifth voice message;

[0050] When the fifth voice message is about joining a group, a joining request message is sent to the cleaning terminal corresponding to the name of the user to be joined in the fifth voice message, or a joining request message is sent to the server according to the name of the group to be joined in the fifth voice message, so as to join the corresponding group through the joining request message.

[0051] In a second aspect, the present invention provides a computer server, comprising: a memory, a processor and a transceiver;

[0052] The processor is used to couple with the memory, read and execute instructions in the memory, so as to implement the voice control task processing method based on artificial intelligence according to any one of the first aspects;

[0053] The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

[0054] In a third aspect, the present invention provides a chip system, comprising a processor, the processor being coupled to a memory, the memory storing program instructions, and when the program instructions stored in the memory are executed by the processor, the artificial intelligence-based voice control task processing method described in any one of the first aspects is implemented.

[0055] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, and the computer program is executed by a processor to implement the artificial intelligence-based voice control task processing method described in any one of the first aspects.

[0056] The artificial intelligence-based voice control task processing method provided by the present invention introduces voice control into the cleaning field. Cleaning personnel do not need to manually operate complex equipment interfaces. Cleaning terminals can complete task processing and clocking-in operations through simple voice commands, bringing new intelligent solutions to the cleaning industry and promoting technological progress and development in the industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 One of the flowcharts of the method for processing a voice control task based on artificial intelligence provided by an embodiment of the present invention;

[0058] Figure 2 For the present invention Figure 1 One of the specific flow charts of step 110;

[0059] Figure 3 For the present invention Figure 1 The specific flow chart 2 of step 110;

[0060] Figure 4 A second flowchart of a method for processing a voice control task based on artificial intelligence provided by an embodiment of the present invention;

[0061] Figure 5 A schematic diagram of the computer server structure provided in Embodiment 2 of the present invention;

[0062] Figure 6 A schematic diagram of the chip system structure provided in Embodiment 3 of the present invention;

[0063] Figure 7 This is a schematic diagram of the chip system structure provided in Example 4 of the present invention. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments.

[0066] Embodiment 1

[0067] Figure 1 The present invention provides a flow chart of a method for processing voice control tasks based on artificial intelligence. The execution subject of the method is a cleaning terminal. The cleaning terminal of the present invention does not have a positioning function. The cleaning terminal of the present invention can be a watch, bracelet, etc. with a processing function. Figure 1 As shown, this application includes the following steps:

[0068] Step 101, receiving broadcast signals of multiple beacons and recording the signal strength of each broadcast signal; the broadcast signal includes a broadcast signal ID;

[0069] Specifically, the current position of the cleaning terminal may be determined by a positioning algorithm according to the positions of the beacons corresponding to the multiple signal strengths.

[0070] Among them, the beacon can be other electronic devices with positioning function, including but not limited to Wi-Fi network positioning and electronic devices, Bluetooth network positioning and electronic devices, UWB positioning and electronic devices, ZigBee positioning and electronic devices, etc.

[0071] Step 102, determining the current location of the cleaning terminal according to the signal strength;

[0072] Specifically, beacons are deployed indoors or in local outdoor areas, and the broadcast signal strength of the beacons is identified by using smart watches worn by personnel. The location information of the cleaning terminal is then obtained through the positioning algorithm. The cleaning terminal and the cleaning staff ID are pre-bound, so the obtained location information of the cleaning terminal can be used as the location coordinates of the cleaning staff.

[0073] Step 103, receiving a first voice message; the first voice message includes an area name;

[0074] Step 104: Process the first voice information to determine the task category.

[0075] The cleaning terminal can receive the voice information of the cleaning staff in real time, and process the voice information to determine the task category. The task categories include the first category, the second category and the third category.

[0076] The received first voice message includes the area name. After semantic analysis and keyword matching, the first voice message is matched with a preset instruction set to determine which type of task category the first voice message corresponds to. The following is a detailed description of various cleaning tasks:

[0077] Specifically, the tasks in this application include but are not limited to assigned tasks, planned tasks, equipment maintenance tasks, inspection tasks, inspection reporting tasks, etc. When an emergency occurs in the venue (such as a guest visiting and a cleaner is urgently needed), the cleaning supervisor can view the location and status of the cleaner through the SAAS platform (hereinafter referred to as the sass platform) or the Smart Cleaning mobile application, and send the assigned task to one or more people based on merit.

[0078] The scheduled task is a fixed time and place in the daily work schedule that requires the cleaner to clean. The server can generate scheduled tasks on a daily, weekly, monthly or custom time period through the intelligent cleaning platform. The scheduled task can select a cleaner or a cleaning team to complete the task. The scheduled task will be sent to the cleaning terminal of the cleaner or the cleaning team before the task execution cycle starts. Among them, the server here runs an intelligent cleaning system that the cleaning manager can log in to.

[0079] Therefore, when the cleaning staff is in the cleaning area where the task is located, when the cleaning staff arrives at a specific cleaning area, the cleaning terminal can automatically remind them to punch in and associate the punch-in information with the task in that area. This ensures that each area is cleaned and recorded in a timely manner. For large cleaning places, such as shopping malls and office buildings, the positioning function can be used to monitor the work trajectory of cleaning staff in real time to ensure that no area is missed and improve the cleaning quality and efficiency. The voice-controlled punch-in method greatly saves the time of cleaning staff and improves the convenience and efficiency of punching in. They no longer need to look for punch-in equipment or perform tedious manual operations, and can focus more on cleaning work.

[0080] When a device fails or requires maintenance, the device will automatically issue a warning and generate a task to be sent to the wearable device of the cleaner responsible for the scene or area.

[0081] Inspection tasks are patrol inspections to facilitate cleaning staff or cleaning supervisors to check whether there are any uninspected scenes. This type of task can be set through Zhiqingjie, including the inspection route and inspection points, inspection execution time range, inspection execution personnel or execution team, etc. When the inspection execution start time arrives, the server will automatically generate the inspection task and send it to the wearable device of the execution personnel or execution team.

[0082] When a patrol task is executed, if a scene that needs cleaning or maintenance is found, you can fill in the patrol task report. The report content includes the report scene, report description, photos, voice and other information. After the report task is filled in and saved, the server will automatically dispatch the task to the wearable device of the cleaner responsible for the scene.

[0083] After the server dispatches the task, the task will be pushed to the cleaning terminals around the task location, and the cleaning staff can complete the operation by voice "accept the order" or "do not accept the order". At the same time, the current time point will be recorded in the database. This time point will serve as the data support for the cleaner to respond to the task from the task dispatch. When the cleaner says not to accept the order, the task will be left in the order pool list for other cleaners to view. When the cleaner says to accept the order, the task will be pushed to the cleaner's received order list, and the task status will be changed to received order.

[0084] The task categories include a first category, a second category and a third category; the first category is a task operation category, the second category is a task information query category, and the third category is an abnormal feedback category.

[0085] The following is a specific description of the preset instruction set.

[0086] This application is aimed at cleaning scenarios and has a dedicated instruction set. For the first category, i.e., the task operation category, the following instructions can be provided:

[0087] “Confirm to grab the order of [specific task name] task”: used to grab the order of the task when assigning tasks to multiple people.

[0088] “Start cleaning [specific area name]”: used to start the cleaning task of the specified area.

[0089] “Pause cleaning [area name]”: Pauses the current cleaning task in the area. This can be used to temporarily respond to emergencies, such as when avoiding people or waiting for tool replenishment.

[0090] “Resume cleaning [area name]”: Resume cleaning of the area after pausing the task.

[0091] “Finish cleaning [area name] and clock in”: After finishing cleaning the designated area, directly clock in to mark the task as completed.

[0092] For the second category, i.e. task information query, the following instructions may be used:

[0093] "Query [area name] cleaning standards": Get the cleaning standard requirements for a specific area so that cleaning staff can clearly understand the work details.

[0094] "View remaining task list": Use voice to query the list of cleaning tasks that have not yet been completed and understand the subsequent work arrangements.

[0095] "Query tasks that will start in the next ten minutes": used to query scheduled tasks or tasks triggered by traffic that are about to be triggered, that is, to query the task list within ten minutes when the task starts or the estimated traffic volume reaches the threshold. This avoids multiple trips by cleaning staff.

[0096] For the third category, namely the exception feedback category, the following instructions can be provided:

[0097] "[Area Name] Cleaning Tool Damaged": Report to the server if a tool is damaged during the cleaning process in a certain area. The server can automatically record and notify relevant personnel to arrange tool replacement or repair.

[0098] "[Area name] found stains that were difficult to remove": When encountering special stains that are difficult to handle, the cleaning staff can provide feedback through this command, and the server will provide corresponding cleaning suggestions or arrange for professional assistance.

[0099] "[Area name] found that the paper output device was damaged by human intervention": When it is found that the paper output device was damaged by human intervention, this command will be used to report to the administrator. The administrator will notify the public security department based on the situation on the scene and protect the scene.

[0100] Step 105: When it is the first category, sending query information to the server according to the current location; the query information is used to instruct the server to query whether there is a task to be completed in the area where the current location is located;

[0101] When the cleaner enters the scene, the cleaning terminal automatically sends the current location to the server, and the server queries whether there are any tasks to be completed in the current scene. The tasks to be completed can be tasks pre-assigned to the cleaner, or tasks obtained by the cleaner through voice order grabbing on the cleaning terminal.

[0102] Step 106, when there are tasks to be completed, receiving the task information to be completed sent by the server;

[0103] Specifically, the cleaning terminal receives the task information to be completed, which may be a list, and each task information to be completed may include a task name, a task time, and a task requirement. The task time may include at least one of the task start time, the task end time, and the task duration. The task requirement may be some additional requirements for completing the task, such as requiring the task to be completed within 1 hour.

[0104] Step 107, generating first task broadcast information according to the task name, task time and task requirements;

[0105] Specifically, the completed tasks are processed and converted into voice. The first task broadcast information may be voice generated according to the order of task name, task area, task time, etc. for the completed tasks, and the subsequent cleaning terminal may broadcast the first task broadcast information.

[0106] Step 108, performing task broadcasting according to the first task broadcasting information;

[0107] Specifically, when the task to be completed is an unstarted task, the task number and task information will be automatically played, including the task name, task location, task time, task requirements, etc. The cleaning staff can choose to start the task or listen to the task information again by the task name or task order.

[0108] Step 109, receiving a second voice message; the second voice message is a response to any task in the first task broadcast information;

[0109] Specifically, the second voice message may be a response message of the cleaner received by the cleaning terminal. The cleaner may choose to start a task or all tasks. When the cleaner says "start the third task", the server will modify the status of the third task to in progress. At the same time, the current time point is recorded, which will be used as the start time of the task.

[0110] When the task to be completed is an ongoing task, there are two scenarios for the ongoing task. One is that after starting the task, the task is completed at the task location until the task is completed. The other is that during the task, the cleaner leaves the site to get tools or do other tasks, and returns to continue the task after a period of time.

[0111] The first is that after the cleaner completes the task, he / she says to the watch "All tasks completed" or "The third task completed". The server will automatically change the task status to completed and record the current time point, which will be used as the time when the task is completed.

[0112] In the second scenario, when the cleaner leaves the current task scene, the departure time is automatically recorded and the task status is changed to paused.

[0113] When the cleaner returns to the task scene, the location information of the cleaning terminal will be automatically uploaded to the server. The server will query whether there are any tasks to be completed in the current scene.

[0114] If there is a task that is paused, the cleaner can say, "Continue the task in progress" or "Continue the third task", and the cleaner can continue to complete the tasks in this scene. When the task is completed, the first scene logic is executed.

[0115] Step 110, processing the second voice information to determine the intention information;

[0116] Specifically, this application includes the following two situations when determining intent information:

[0117] The first type is when the second voice message is a clear instruction, such as Figure 2 As shown, this application includes the following steps:

[0118] Step 201, converting the second voice information to obtain corresponding text;

[0119] Step 202, performing part-of-speech tagging on the text;

[0120] Specifically, the text obtained by speech recognition is input into the BERT model for lexical analysis. The model can automatically identify the words in the sentence and accurately mark the parts of speech, such as accurately distinguishing nouns such as area names and verbs such as words related to cleaning actions.

[0121] Step 203, performing syntactic analysis on the text after part-of-speech tagging to generate an embedded representation of the sentence;

[0122] Specifically, in syntactic analysis, the BERT model can accurately determine the structure and grammatical relationship of sentences by learning from large-scale text corpora. For example, for "start the lobby cleaning process", the model can clearly parse "start" as the core verb, "lobby" as the target object, and "cleaning process" as the action category.

[0123] Step 204, matching the embedded representation with the predefined instruction set, and obtaining similarity based on cosine similarity or dot product calculation;

[0124] Specifically, the semantic matching algorithm based on the attention mechanism is used to deeply match the semantic representation analyzed by the BERT model with the pre-defined voice command set, so as to accurately determine the intention of the voice command. For "starting the lobby cleaning process", after semantic matching, it can be determined that its intention is to officially start the cleaning task sequence of the lobby area.

[0125] Step 205: Determine the user intention based on the similarity.

[0126] Specifically, according to the similarity, it is determined that the one with the highest similarity is the user intention.

[0127] The second type is when the second voice information is an ambiguous voice command, such as Figure 3 As shown, this application includes the following steps:

[0128] Step 301, obtaining a plurality of clean areas in the voice information before the second voice information;

[0129] Specifically, when the cleaning staff's expression is semantically ambiguous, such as "let's do it here first", the cleaning terminal starts the semantic disambiguation algorithm.

[0130] A Bayesian network can be pre-built, and the network nodes include possible cleaning areas (such as lobby, toilet, corridor, etc.), areas recently mentioned by cleaning staff, and areas where cleaning staff are currently located. The conditional probability table of each node in the Bayesian network is determined based on historical data and prior knowledge. For example, it is known that in 6 of the past 10 tasks, the cleaning staff mentioned the toilet and then mentioned "here" to refer to the toilet. Then the conditional probability from "recently mentioned the toilet" to "here refers to the toilet" is 0.6.

[0131] When receiving ambiguous voice commands, the cleaning terminal collects current relevant evidence, such as the area information and current location information recently mentioned by the cleaning staff, and inputs this evidence into the Bayesian network.

[0132] Step 302, using the multiple clean areas and the current position as inputs of a Bayesian network to calculate the posterior probability that each clean area is referred to;

[0133] Specifically, the Bayesian inference algorithm is used to calculate the posterior probability of each clean area being referred to under given evidence. For example, the posterior probability of "here refers to the bathroom" is calculated to be 0.7, the posterior probability of "here refers to the lobby" is calculated to be 0.2, and so on.

[0134] Step 303, selecting the clean area with the highest posterior probability as the inferred reference area;

[0135] Specifically, the posterior probability that the bathroom is referred to here is greater than the posterior probability that the lobby is referred to here, so the bathroom is referred to here, and the bathroom is the reference area.

[0136] Step 304: generating second task broadcast information according to the cleaning area; the second task broadcast information is used to inquire whether to use the cleaning area as a reference area;

[0137] Specifically, the cleaning area with the highest posterior probability is selected as the inferred reference area of ​​"here" in the fuzzy instruction. A second task broadcast information is generated, and the second task broadcast information is played to voice-inquire whether the cleaning staff confirms, and wait for further voice confirmation or correction instructions.

[0138] Step 305, receiving a third voice message for the second task broadcast information; the third voice message is used to indicate whether to use the cleaning area as a reference area;

[0139] Specifically, if the third voice information is to confirm that the cleaning area is used as the reference area, the task of the cleaning area is executed.

[0140] Step 306: If the cleaning area is used as the reference area, the task of the reference area is executed.

[0141] Further, such as Figure 4 As shown, the application also includes the following steps:

[0142] Step 401, when the first voice information is of the second category, determining query content according to the first voice information; the query content includes cleaning standard query, remaining task query and future task query;

[0143] Step 402, when the task category is the second category, if the query content is a cleaning standard query or a future task query, generate third broadcast information, and broadcast the third broadcast information;

[0144] Step 403: when the query content is a query of remaining tasks, query the corresponding number of completed tasks according to the area name; the number of completed tasks includes the completion steps of a task or the number of completed tasks among multiple tasks;

[0145] Step 404, comparing the number of completed tasks with a preset total number to determine the completion ratio; the preset total number includes the total number of steps of a task or the number of multiple tasks;

[0146] Step 405: Generate fourth task broadcast information according to the completion ratio; the fourth task broadcast information is used to indicate the progress of the task.

[0147] Specifically, during the execution of a task, cleaning staff can use voice commands to check the progress of the task or obtain other relevant information at any time. For example, if you say "check the progress of lobby cleaning", the cleaning terminal will perform an intelligent ratio calculation based on the number of cleaning steps that have been successfully completed and the pre-set total number of steps, and then respond with a voice message "Lobby cleaning is 40% complete, please focus on cleaning the seats in the public rest area next."

[0148] Furthermore, the present application also includes: when the first voice information is of the third category, sending the first voice information to a server to feedback abnormal situations through the first voice information.

[0149] If an abnormal situation occurs, such as "scratches on the marble floor of the lobby", the cleaning terminal will immediately record the abnormal information in detail and push it to the terminal device of the manager in real time after receiving this instruction, and respond to the cleaning staff with a voice message "The abnormal situation of scratches on the lobby floor has been successfully reported, please pay attention to on-site protection and wait patiently for further instructions". After receiving the notification, the manager can make full use of the intercom function of the cleaning terminal to communicate directly with the cleaning staff in real time, deeply understand the specific details, and quickly arrange subsequent appropriate handling measures, such as emergency deployment of professional maintenance personnel.

[0150] Through the above AI model-based and deeply optimized voice control solution combined with specific algorithms, the deep integration and intelligent interaction of voice commands and cleaning tasks are realized. With the help of advanced AI voice recognition, semantic understanding technology and intelligent algorithm design, a highly intelligent, precise, convenient and efficient voice control function is achieved in the cleaning task scenario.

[0151] Furthermore, the present application also includes the following method:

[0152] Receive a wake-up word; activate the intercom function according to the wake-up word; query a preset cleaning terminal list and determine the cleaning terminal in the cleaning terminal list with the intercom function activated; send a fourth voice message to the cleaning terminal with the intercom function activated.

[0153] Specifically, users do not need to manually start recording voice. You can input the voice to the cleaning terminal: "Xiaoqu, Xiaoqu". After the cleaning terminal receives the wake-up word, it compares the wake-up word with the preset wake-up word. If they are consistent, the intercom function will be awakened. When the intercom function is turned on, the cleaning terminal will receive the user's voice and forward the voice to other cleaning terminals that have turned on the intercom function. When the cleaning terminal receives "over" or "end", the cleaning terminal will automatically turn off the intercom function.

[0154] Furthermore, the present application also includes the following method:

[0155] receiving a fifth voice message, where the fifth voice message is used to instruct to establish a group or join a group;

[0156] When the fifth voice message is for establishing a group, establishing and saving the group according to the fifth voice message;

[0157] When the fifth voice message is about joining a group, a joining request message is sent to the cleaning terminal corresponding to the name of the user to be joined in the fifth voice message, or a joining request message is sent to the server according to the name of the group to be joined in the fifth voice message, so as to join the corresponding group through the joining request message.

[0158] Specifically, the cleaning terminal can receive the user's voice "I need to set up a walkie-talkie group" to start a new group, and the cleaning terminal will automatically pop up the interface for creating a group. The cleaning terminal can enter, save and modify group information through interface operations. Similarly, a group can be built and saved by describing the detailed information of the group.

[0159] When the cleaning terminal receives the voice message "Drag Zhang San and Li Si into the group just created", the terminal will automatically drag Zhang San and Li Si into the group.

[0160] The cleaning terminal can also receive voice messages, such as "I want to join a group", to submit an application to join the group. In a group call, only group members can answer the call, and others cannot. Unlike traditional intercoms, there is no need to set up cumbersome channels. Just join the group and you can talk. When there is no group, the call will be blocked.

[0161] The artificial intelligence-based voice control task processing method provided by the present invention introduces voice control into the cleaning field. Cleaning personnel do not need to manually operate complex equipment interfaces. Cleaning terminals can complete task processing and clocking-in operations through simple voice commands, bringing new intelligent solutions to the cleaning industry and promoting technological progress and development in the industry.

[0162] Embodiment 2

[0163] Embodiment 2 of the present invention provides a computer server, such as Figure 5 As shown, it includes: a memory, a processor and a transceiver;

[0164] The processor is used to couple with the memory, read and execute instructions in the memory, so as to implement any one of the artificial intelligence-based voice control task processing methods provided in the first embodiment above;

[0165] The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

[0166] Embodiment 3

[0167] Embodiment 3 of the present invention provides a chip system, such as Figure 6 As shown, it includes a processor, a coupling between the processor and a memory, the memory stores program instructions, and when the program instructions stored in the memory are executed by the processor, any one of the artificial intelligence-based voice control task processing methods provided in Example 1 is implemented.

[0168] Embodiment 4

[0169] Embodiment 4 of the present invention provides a computer-readable storage medium, such as Figure 7 As shown, it includes a program or instruction. When the program or instruction is run on a computer, any one of the artificial intelligence-based voice control task processing methods provided in Example 1 is implemented.

[0170] The professionals should further realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to the function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0171] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0172] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A voice control task processing method based on artificial intelligence, characterized in that: The method comprises: receiving broadcast signals of a plurality of beacons and recording the signal strength of each broadcast signal; the broadcast signal includes a broadcast signal ID; Determining the current location of the cleaning terminal according to the signal strength; Receiving a first voice message; the first voice message includes an area name; Processing the first voice information to determine a task category; the task category includes a first category, a second category and a third category; When it is the first category, query information is sent to the server according to the current location; the query information is used to instruct the server to query whether there is a task to be completed in the area where the current location is located; When there are tasks to be completed, receiving task information to be completed sent by the server; the task information to be completed includes task name, task time and task requirements; Generate first task broadcast information according to the task name, task time and task requirements; Performing task broadcasting according to the first task broadcasting information; receiving a second voice message; the second voice message being a response to any task in the first task broadcast message; Processing the second voice information to determine intention information; According to the intention information, the corresponding task is executed.

2. The method according to claim 1, characterized in that Determining the current position of the cleaning terminal according to the signal strength specifically includes: According to the positions of the beacons corresponding to the multiple signal strengths, the current position of the cleaning terminal is determined by a positioning algorithm.

3. The method according to claim 1, characterized in that The processing of the second voice information to determine the intention information specifically includes: Converting the second voice information to obtain corresponding text; Performing part-of-speech tagging on the text; Perform syntactic analysis on the text after part-of-speech tagging to generate an embedded representation of the sentence; Match the embedded representation with a predefined instruction set and calculate the similarity based on cosine similarity or dot product; Based on the similarity, determine the user intent.

4. The method according to claim 1, characterized in that: The processing of the second voice information to determine the intention information specifically includes: When the second voice information is an ambiguous voice command, acquiring a plurality of clean areas in the voice information before the second voice information; Using the multiple clean areas and the current position as inputs of a Bayesian network to calculate a posterior probability that each clean area is referred to; Select the clean area with the highest posterior probability as the inferred reference area; Generate second task broadcast information according to the cleaning area; the second task broadcast information is used to inquire whether to use the cleaning area as a reference area; receiving a third voice message for the second task broadcast information; the third voice message is used to indicate whether to use the cleaning area as a reference area; If the cleaning area is taken as the reference area, the task of the reference area is executed.

5. The method according to claim 1, characterized in that The method further comprises: When the first voice information is of the second category, determining the query content according to the first voice information; the query content includes a cleaning standard query, a remaining task query, and a future task query; When the task category is the second category, if the query content is a cleaning standard query or a future task query, generating third broadcast information, and broadcasting the third broadcast information; When the query content is a query of remaining tasks, the number of completed tasks corresponding to the area name is queried; the number of completed tasks includes the completion steps of a task or the number of completed tasks among multiple tasks; Compare the number of completed tasks with a preset total number to determine the completion ratio; the preset total number includes the total steps of a task or the number of multiple tasks; According to the completion ratio, fourth task broadcast information is generated; the fourth task broadcast information is used to indicate the progress of the task.

6. The method according to claim 1, characterized in that: The method further comprises: When the first voice information is of the third category, the first voice information is sent to a server to feedback an abnormal situation through the first voice information.

7. The method according to claim 1, characterized in that The method further comprises: Receive the wake-up word; According to the wake-up word, start the intercom function; Querying a preset cleaning terminal list, and determining a cleaning terminal in the cleaning terminal list that has an intercom function activated; The fourth voice message is sent to the cleaning terminal with the intercom function activated.

8. The method according to claim 1, characterized in that The method further comprises: receiving a fifth voice message, where the fifth voice message is used to instruct to establish a group or join a group; When the fifth voice message is for establishing a group, establishing and saving the group according to the fifth voice message; When the fifth voice message is about joining a group, a joining request message is sent to the cleaning terminal corresponding to the name of the user to be joined in the fifth voice message, or a joining request message is sent to the server according to the name of the group to be joined in the fifth voice message, so as to join the corresponding group through the joining request message.

9. A chip system, characterized in that: It includes a processor, the processor is coupled to a memory, the memory stores program instructions, and when the program instructions stored in the memory are executed by the processor, the artificial intelligence-based voice control task processing method described in any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is executed by the processor to implement the artificial intelligence-based voice control task processing method described in any one of claims 1 to 8.

Images