Robot man-machine interaction supervision system based on artificial intelligence

Abstract

The invention discloses a robot man-machine interaction supervision system based on artificial intelligence, and belongs to the technical field of robot supervision. The method comprises the following steps of: integrating an active inquiry condition of a user and an active inquiry processing result to obtain an inventory-insufficient article sorting table and an inexistent article sorting table corresponding to different articles; according to the sorting table of the goods with insufficient stock and the sorting table of the goods without existence, the goods demand conditions of different users and the stockout conditions and the replenishment conditions of different goods on shelf can be more accurately and efficiently obtained; the different abnormal feedback types are classified and marked by integrating and calculating the different abnormal feedback types appearing every day and the corresponding total appearing times, so that managers and operation and maintenance personnel can timely and efficiently obtain influences generated by the different abnormal feedback types; the method and the device are used for solving the technical problems of single monitoring data mining integration dimension and poor monitoring data utilization effect of robot man-machine interaction in the existing scheme.

Description

Technical Field

[0001] The present invention relates to the technical field of robot supervision, and in particular to a robot human-machine interaction supervision system based on artificial intelligence. Background Art

[0002] Service consumption robots refer to intelligent robots that can provide specific services and can provide users with various services, such as catering, cleaning, security, shopping, medical consultation, etc.; some common service consumption robots include restaurant service robots, hotel service robots, cleaning service robots, security service robots, shopping service robots, medical service robots, etc.

[0003] When implementing existing robot-human interaction supervision programs, most of them remain in the aspects of interactive question-and-answer and data recording. They do not expand the mining of recorded data from different dimensions to implement deeper data analysis and utilization, and do not dynamically manage the subsequent optimization of robot-human interaction based on the mined recorded data. This results in a single dimension of mining and integration of monitoring data for robot-human interaction and poor utilization of monitoring data. Summary of the Invention

[0004] The purpose of the present invention is to provide an artificial intelligence-based robot human-machine interaction supervision system to solve the technical problems of the existing solutions in which the monitoring data mining and integration of robot human-machine interaction are single in dimension and the monitoring data utilization effect is poor.

[0005] The purpose of the present invention can be achieved through the following technical solutions:

[0006] A robot human-machine interaction supervision system based on artificial intelligence, comprising:

[0007] The human-machine interaction information monitoring and analysis module is used to monitor and analyze the human-machine interaction service data of different service robots every day, and obtain individual human-machine interaction supervision data including query analysis data and interaction process monitoring data;

[0008] The human-machine interaction information anomaly assessment module is used to process and integrate all query analysis data uploaded to the cloud platform daily to obtain the corresponding query analysis integration table and push prompts in real time. It also performs data preprocessing and analysis on all interaction monitoring data to obtain different abnormal feedback states of the robot and dynamically optimize prompts for abnormalities in robot interaction. It includes:

[0009] When processing and integrating all query analysis data, all first match failure tags and second match failure tags are traversed and counted for all query analysis data. All corresponding insufficient inventory items and non-existent items are obtained based on all first match failure tags and second match identification tags, and are sorted and combined to obtain an insufficient inventory item sorting table and a non-existent item sorting table. The insufficient inventory item sorting table and the non-existent item sorting table are processed and combined to obtain a query analysis integration table.

[0010] During the real-time data preprocessing and analysis of all interactive monitoring data, all interactive monitoring data are traversed to count all abnormal feedback types and the total number of feedbacks corresponding to different abnormal feedback types from all robots, and the abnormal feedback impact coefficients corresponding to different abnormal feedback types are obtained by processing and calculating; the abnormal feedback impact coefficients are classified and marked to obtain a number of high-impact abnormal feedback and low-impact abnormal feedback;

[0011] All sorted high-impact exception feedback and low-impact exception feedback are synchronously pushed to management personnel and operation and maintenance personnel corresponding to robot interaction to determine the interaction optimization types and interaction optimization plans corresponding to different exception feedbacks, and the optimization of robot interaction is dynamically implemented according to the determined interaction optimization types and interaction optimization plans.

[0012] Preferably, the movement state of the shopping cart equipped with the robot is monitored, and a monitoring start instruction is generated when the shopping cart is pushed from the parking area and enters the target area. The start time of the shopping cart is recorded according to the monitoring start instruction, the real-time movement of the shopping cart is located, and the service data of the robot during the movement of the shopping cart is monitored and counted.

[0013] When a user asks a question to the robot, the name of the item in the query statement is obtained through the voice recognition algorithm and the keyword recognition algorithm, and the item name is traversed and matched in the item database to obtain query analysis data consisting of a successful match label, a first match failure label, or a second match failure label.

[0014] Preferably, if the corresponding item exists in the item database and the remaining quantity is not zero, a successful match tag is generated and the items of different brands corresponding to the item name and the corresponding placement location are pushed to the display screen on the shopping cart. Upon confirmation by the user, the distance between the shopping cart and the placement location of the corresponding item and the navigation route are displayed in real time.

[0015] If the corresponding item exists in the item database and the remaining quantity is zero, a first match failure tag is generated and a prompt indicating insufficient item inventory is generated;

[0016] If the corresponding item does not exist in the item database, a second matching failure tag is generated and a prompt indicating that the item does not exist is generated.

[0017] Preferably, when monitoring, recording and analyzing the process of user inquiry interaction, the total number of user inquiries, as well as all abnormal feedback types and the total number of feedback corresponding to different abnormal feedback types recorded by the robot are counted, and the total number of user inquiries, all abnormal feedback types and the total number of feedback corresponding to different abnormal feedback types are arranged and combined to obtain interaction process monitoring data;

[0018] When the user's shopping cart enters the checkout area, a monitoring end instruction is generated. The start checkout time of the shopping cart is recorded according to the monitoring end instruction, and the shopping cart number and the corresponding query analysis data and interaction process monitoring data are combined to obtain individual human-computer interaction supervision data and uploaded to the cloud platform in real time.

[0019] Preferably, identical items appearing in the insufficient inventory item sorting table and the non-existent item sorting table are deduplicated and associated with the total number of occurrences. The deduplicated insufficient inventory item sorting table and the non-existent item sorting table constitute an inquiry analysis integration table and push real-time prompts to management personnel to arrange item replenishment in a timely manner and the overall inquiry status of non-existent items.

[0020] Preferably, when all abnormal feedback types that occur are digitally processed, different abnormal feedback types are numbered and marked as i, i = 1, 2, 3, ..., n; n is a positive integer; and the total number of occurrences of different abnormal feedback types is marked as Ni;

[0021] According to the time sequence of occurrence, different abnormal feedback types are sequentially matched with the abnormal feedback type weight table pre-stored in the database to obtain the corresponding feedback type weight Qi.

[0022] Preferably, the values ​​of the marked data are extracted and the formula Calculate and obtain the abnormal feedback impact coefficient YFy corresponding to different abnormal feedback types.

[0023] Preferably, all abnormal feedback impact coefficients and corresponding abnormal feedback types are arranged in descending order according to the values ​​of the abnormal feedback impact coefficients, all abnormal feedback types that are not less than the abnormal feedback impact threshold are marked as high-impact abnormal feedback, and all abnormal feedback types that are less than the abnormal feedback impact threshold are marked as low-impact abnormal feedback.

[0024] Preferably, the average of all abnormal feedback impact coefficients is calculated and marked as the abnormal feedback impact threshold.

[0025] Compared with the existing solutions, the present invention achieves the following beneficial effects:

[0026] The present invention obtains a sorting table of insufficient inventory items and a sorting table of non-existent items corresponding to different items through integration of the user's active inquiries and the results of active inquiry processing. Based on the sorting table of insufficient inventory items and the sorting table of non-existent items, the item demand of different users and the out-of-stock and replenishment status of different shelf goods can be obtained more accurately and efficiently, thereby improving the diversity and reliability of robot interactive data mining.

[0027] The present invention integrates and calculates the different abnormal feedback types that appear every day and the corresponding total number of occurrences to obtain the corresponding abnormal feedback impact coefficient, and classifies and labels the different abnormal feedback types according to the abnormal feedback impact coefficient, so that management personnel and operation and maintenance personnel can obtain the impact of different abnormal feedback types in a timely and efficient manner, so as to actively optimize and maintain the interactions of different robots, thereby improving the diversity and reliability of the expansion and utilization of robot interaction abnormal data monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The present invention will be further described below with reference to the accompanying drawings.

[0029] Figure 1 This is a module block diagram of an artificial intelligence-based robot human-machine interaction supervision system of the present invention. DETAILED DESCRIPTION

[0030] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary operation and maintenance personnel in this field without making any creative efforts are within the scope of protection of the present invention.

[0031] like Figure 1 As shown, the present invention is an artificial intelligence-based robot human-machine interaction supervision system, which includes a human-machine interaction information monitoring and analysis module, a human-machine interaction information anomaly assessment module, a cloud platform and a database;

[0032] The human-machine interaction information monitoring and analysis module is used to monitor and analyze the human-machine interaction service data of different service robots every day to obtain individual human-machine interaction supervision data; it includes:

[0033] The movement status of a shopping cart equipped with a robot is monitored. When the shopping cart is pushed from the parking area and enters the target area, a monitoring start instruction is generated. The determination of whether the shopping cart has been pushed from the parking area and entered the target area can be achieved based on existing Internet of Things technologies, such as RFID (radio frequency identification) technology or positioning technologies such as GPS, Wi-Fi, and Bluetooth. The start time of use of the shopping cart is recorded according to the monitoring start instruction, the real-time movement of the shopping cart is located, and the service data of the robot during the movement of the shopping cart is monitored and counted.

[0034] The service robot in the embodiment of the present invention is a shopping robot in a shopping supermarket, which can be installed on a shopping cart so that users can query information, receive shopping guidance, and receive discount information in a timely and efficient manner.

[0035] When the user asks the robot a question, the robot uses the voice recognition algorithm and keyword recognition algorithm to obtain the item name in the query sentence, and then traverses and matches the item name in the item database;

[0036] Among them, the speech recognition algorithm and the keyword recognition algorithm are both existing conventional technical solutions, and the specific steps are not repeated here;

[0037] For example, when a user asks, "Where is the soy sauce?", the recognition algorithm obtains the item name: soy sauce.

[0038] If the corresponding item exists in the item database and the remaining quantity is not zero, a successful match tag is generated and the item of different brands corresponding to the item name and the corresponding placement location are pushed to the display screen on the shopping cart. The distance between the shopping cart and the placement location of the corresponding item and the navigation route are displayed in real time upon user confirmation; user confirmation can be achieved by clicking the "Confirm" button displayed on the display screen;

[0039] The item database is constructed by the names of all commodities in the corresponding shopping supermarket; the placement position is the number of the shelf;

[0040] If the corresponding item exists in the item database and the remaining quantity is zero, a first match failure tag is generated and a prompt indicating insufficient item inventory is generated; insufficient item inventory can be updated based on the quantity of goods on the shelves and the real-time settlement quantity of the same goods;

[0041] If the corresponding item does not exist in the item database, a second matching failure tag is generated and a prompt indicating that the item does not exist is generated;

[0042] The match success tag, the first match failure tag, or the second match failure tag constitutes query analysis data;

[0043] In the embodiment of the present invention, by implementing data recording and statistics on users' active inquiries, the results of each user's inquiry can be obtained, and reliable individual inquiry data support can be provided for subsequent item replenishment and new product analysis.

[0044] When monitoring, recording, and analyzing the process of user inquiry interactions, the total number of user inquiries, as well as all abnormal feedback types and the total number of feedback corresponding to different abnormal feedback types recorded by the robot are counted. The total number of user inquiries, all abnormal feedback types, and the total number of feedback corresponding to different abnormal feedback types are arranged and combined to obtain interaction process monitoring data.

[0045] Among them, all abnormal feedback types reported by the robot are monitored and counted by the robot itself, including but not limited to speech recognition failure types, keyword recognition failure types, network failure types, display failure types, etc.

[0046] When the user's shopping cart enters the checkout area, a monitoring end instruction is generated. The start checkout time of the shopping cart is recorded according to the monitoring end instruction. The shopping cart number and the corresponding query analysis data and interaction process monitoring data are combined to obtain individual human-computer interaction supervision data and uploaded to the cloud platform in real time;

[0047] In an embodiment of the present invention, by monitoring, recording and analyzing the process of user inquiry interaction, it is possible to actively record the abnormal interaction data each time it occurs, and at the same time provide reliable local interaction data support for targeted optimization of different aspects of abnormalities that may occur in subsequent robot interactions.

[0048] The human-machine interaction information anomaly assessment module is used to process and integrate all query analysis data uploaded to the cloud platform daily to obtain the corresponding query analysis integration table and push prompts in real time. It also performs data preprocessing and analysis on all interaction monitoring data to obtain different abnormal feedback states of the robot and dynamically optimize prompts for abnormalities in robot interaction. It includes:

[0049] When processing and integrating all query analysis data, all first match failure tags and second match failure tags are traversed and counted for all query analysis data. All corresponding insufficient inventory items and non-existent items are obtained based on all first match failure tags and second match identification tags, and are sorted and combined to obtain an insufficient inventory item sorting table and a non-existent item sorting table.

[0050] Duplicate items appearing in the insufficient inventory item ranking table and the non-existing item ranking table are removed and the total number of occurrences is correlated. The deduplicated insufficient inventory item ranking table and the non-existing item ranking table form an inquiry analysis integration table, which provides real-time notifications to managers to promptly arrange item replenishment and the overall inquiry status of non-existing items.

[0051] In an embodiment of the present invention, by integrating the user's active inquiries and the results of the active inquiry processing and obtaining the insufficient inventory item ranking table and the non-existent item ranking table corresponding to different items, the insufficient inventory item ranking table and the non-existent item ranking table can be used to more accurately and efficiently obtain the item demand situation of different users and the out-of-stock and replenishment situation of different shelf goods, thereby improving the diversity and reliability of robot interactive data mining.

[0052] During the real-time data preprocessing and analysis of all interactive monitoring data, all interactive monitoring data are traversed and counted to obtain all abnormal feedback types and the total number of feedbacks corresponding to different abnormal feedback types from all robots;

[0053] When all abnormal feedback types are digitally processed, different abnormal feedback types are numbered and marked as i, i = 1, 2, 3, ..., n; n is a positive integer; and the total number of occurrences of different abnormal feedback types is marked as Ni;

[0054] According to the time sequence of occurrence, different abnormal feedback types are sequentially matched with the abnormal feedback type weight table pre-stored in the database to obtain the corresponding feedback type weight Qi;

[0055] Among them, the abnormal feedback type weight table is pre-set with several abnormal feedback types and corresponding feedback type weights. The specific values ​​of the feedback type weights corresponding to different abnormal feedback types are determined by the robot developer based on development work experience;

[0056] Extract the values ​​of the marked data and use the formula Calculate and obtain the abnormal feedback influence coefficient YFy corresponding to different abnormal feedback types;

[0057] According to the values ​​of the abnormal feedback impact coefficients, all abnormal feedback impact coefficients and corresponding abnormal feedback types are sorted in descending order. At the same time, the average of all abnormal feedback impact coefficients is calculated and marked as the abnormal feedback impact threshold. All abnormal feedback types with a value not less than the abnormal feedback impact threshold are marked as high-impact abnormal feedback, and all abnormal feedback types with a value less than the abnormal feedback impact threshold are marked as low-impact abnormal feedback.

[0058] All sorted high-impact exception feedback and low-impact exception feedback are synchronously pushed to the management personnel and the operation and maintenance personnel corresponding to the robot interaction to determine the interaction optimization types and interaction optimization plans corresponding to different exception feedbacks, and the optimization of the robot interaction is dynamically implemented according to the determined interaction optimization types and interaction optimization plans; among them, the interaction optimization types and interaction optimization plans are determined by the operation and maintenance personnel according to the actual application scenarios and actual application requirements.

[0059] In the implementation of the present invention, the different abnormal feedback types that occur daily and the corresponding total number of occurrences are integrated and calculated to obtain the corresponding abnormal feedback impact coefficient, and the different abnormal feedback types are classified and marked according to the abnormal feedback impact coefficient, so that management personnel and operation and maintenance personnel can obtain the impact of different abnormal feedback types in a timely and efficient manner, so that they can actively optimize and maintain and update the interactions of different robots, thereby improving the diversity and reliability of the expansion and utilization of robot interaction abnormal data monitoring.

[0060] In addition, the formulas involved in the above are all calculated by removing dimensions and taking their numerical values. They are a formula that is closest to the actual situation obtained by collecting a large amount of data and simulating it through simulation software.

[0061] In the several embodiments provided by the present invention, it should be understood that the disclosed system can be implemented in other ways. For example, the embodiments of the invention described above are merely illustrative. For example, the division of modules is only a logical function division, and other division methods may be used in actual implementation.

[0062] Modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical modules, and may be located in one place or distributed across multiple network modules. Some or all of these modules may be selected to achieve the objectives of this embodiment based on actual needs.

[0063] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing module, each module may exist physically separately, or two or more modules may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or hardware plus software functional modules.

[0064] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the basic characteristics of the present invention.

[0065] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, ordinary operation and maintenance personnel in this field should understand that the technical solutions of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A robot human-machine interaction supervision system based on artificial intelligence, characterized in that: It includes a human-machine interaction information monitoring and analysis module, which is used to monitor and analyze the human-machine interaction service data of different service robots every day, and obtain individual human-machine interaction supervision data including query analysis data and interaction process monitoring data; The human-machine interaction information anomaly assessment module is used to process and integrate all query analysis data uploaded to the cloud platform daily to obtain the corresponding query analysis integration table and push prompts in real time. It also performs data preprocessing and analysis on all interaction monitoring data to obtain different abnormal feedback states of the robot and dynamically optimize prompts for abnormalities in robot interaction. It includes: When processing and integrating all query analysis data, all first match failure tags and second match failure tags are traversed and counted for all query analysis data. All corresponding insufficient inventory items and non-existent items are obtained based on all first match failure tags and second match identification tags, and are sorted and combined to obtain an insufficient inventory item sorting table and a non-existent item sorting table. The insufficient inventory item sorting table and the non-existent item sorting table are processed and combined to obtain a query analysis integration table. During the real-time data preprocessing and analysis of all interactive monitoring data, all interactive monitoring data are traversed to count all abnormal feedback types and the total number of feedbacks corresponding to different abnormal feedback types from all robots, and the abnormal feedback impact coefficients corresponding to different abnormal feedback types are obtained by processing and calculating; the abnormal feedback impact coefficients are classified and marked to obtain a number of high-impact abnormal feedback and low-impact abnormal feedback; All sorted high-impact exception feedback and low-impact exception feedback are synchronously pushed to management personnel and operation and maintenance personnel corresponding to robot interaction to determine the interaction optimization types and interaction optimization plans corresponding to different exception feedbacks, and the optimization of robot interaction is dynamically implemented according to the determined interaction optimization types and interaction optimization plans.

2. The artificial intelligence-based robot human-machine interaction supervision system according to claim 1, characterized in that: Monitor the movement of a shopping cart equipped with a robot. When the shopping cart is pushed from the parking area and enters the target area, a monitoring start instruction is generated. The start time of the shopping cart is recorded according to the monitoring start instruction. The real-time movement of the shopping cart is also located. The service data of the robot during the movement of the shopping cart is monitored and compiled. When a user asks a question to the robot, the name of the item in the query statement is obtained through the voice recognition algorithm and the keyword recognition algorithm, and the item name is traversed and matched in the item database to obtain query analysis data consisting of a successful match label, a first match failure label, or a second match failure label.

3. The artificial intelligence-based robot human-machine interaction supervision system according to claim 2, characterized in that: If the corresponding item exists in the item database and the remaining quantity is not zero, a successful match tag is generated and the items of different brands corresponding to the item name and the corresponding placement location are pushed to the display screen on the shopping cart. After the user confirms, the distance between the shopping cart and the placement location of the corresponding item and the navigation route are displayed in real time; If the corresponding item exists in the item database and the remaining quantity is zero, a first match failure tag is generated and a prompt indicating insufficient item inventory is generated; If the corresponding item does not exist in the item database, a second matching failure tag is generated and a prompt indicating that the item does not exist is generated.

4. The artificial intelligence-based robot human-machine interaction supervision system according to claim 3 is characterized in that: When monitoring, recording, and analyzing the process of user inquiry interactions, the total number of user inquiries, as well as all abnormal feedback types and the total number of feedback corresponding to different abnormal feedback types recorded by the robot are counted. The total number of user inquiries, all abnormal feedback types, and the total number of feedback corresponding to different abnormal feedback types are arranged and combined to obtain interaction process monitoring data. When the user's shopping cart enters the checkout area, a monitoring end instruction is generated. The start checkout time of the shopping cart is recorded according to the monitoring end instruction, and the shopping cart number and the corresponding query analysis data and interaction process monitoring data are combined to obtain individual human-computer interaction supervision data and uploaded to the cloud platform in real time.

5. The artificial intelligence-based robot human-machine interaction supervision system according to claim 1, characterized in that: The same items that appear in the insufficient inventory item sorting table and the non-existent item sorting table are deduplicated and the total number of occurrences are associated. The deduplicated insufficient inventory item sorting table and the non-existent item sorting table constitute an inquiry analysis integration table and push real-time prompts to management personnel to arrange item replenishment in a timely manner and the overall inquiry status of non-existent items.

6. The artificial intelligence-based robot human-machine interaction supervision system according to claim 5, characterized in that: When all abnormal feedback types are digitally processed, different abnormal feedback types are numbered and marked as i, i = 1, 2, 3, ..., n; n is a positive integer; and the total number of occurrences of different abnormal feedback types is marked as Ni; According to the time sequence of occurrence, different abnormal feedback types are sequentially matched with the abnormal feedback type weight table pre-stored in the database to obtain the corresponding feedback type weight Qi.

7. The artificial intelligence-based robot human-machine interaction supervision system according to claim 6, characterized in that: Extract the values ​​of the marked data and use the formula Calculate and obtain the abnormal feedback impact coefficient YFy corresponding to different abnormal feedback types.

8. The artificial intelligence-based robot human-machine interaction monitoring system according to claim 7, characterized in that: All abnormal feedback impact coefficients and corresponding abnormal feedback types are sorted in descending order according to the values ​​of the abnormal feedback impact coefficients. All abnormal feedback types whose values ​​are not less than the abnormal feedback impact threshold are marked as high-impact abnormal feedback, and all abnormal feedback types whose values ​​are less than the abnormal feedback impact threshold are marked as low-impact abnormal feedback.

9. The artificial intelligence-based robot human-machine interaction supervision system according to claim 8, characterized in that: The mean of all abnormal feedback impact coefficients is calculated and marked as the abnormal feedback impact threshold.

Images