Space station material information management system and method based on intelligent voice interaction

Abstract

The present invention discloses a space station material information management system and method based on intelligent voice interaction, which belongs to the field of space station material management technology, including: an on-orbit computing node, which includes a text extraction module, a semantic understanding and logic processing judgment module, and an on-orbit business management system; also includes a ground computing node, which includes a ground business management system and a semantic understanding and data judgment module connected in sequence; the ground business management system is used to receive incomprehensible text data sent by the on-orbit business management system, and then perform semantic analysis on the text data based on the ground intelligent solution service terminal through the semantic understanding and data judgment module. When the on-orbit semantics cannot be processed, the on-orbit business management system transmits the data to the ground intelligent solution service terminal for processing, thereby obtaining a highly accurate semantic understanding result, and then the ground updates the business data, thereby supporting the on-orbit semantic understanding and improving the accuracy of on-orbit material management.

Description

Technical Field

[0001] The present invention relates to the technical field of space station material management, and in particular to a space station material information management system and method based on intelligent voice interaction. Background Art

[0002] During the user's stay in orbit, they need to manually scan the materials to complete the operation of the material information management system. There are a large number of material handling and retrieval activities on orbit. When users manage materials on a daily basis, they need to use materials, software, and mobile phones in turn to achieve material management. Due to the influence of the weightless environment, the hands can no longer operate the mobile phone to scan the material label to identify the material information, operate the software to achieve material information management, and when the material is not identified by the material information management system without a code, the information management of this material cannot be performed. This leads to a large workload and low efficiency for the material information management activities on orbit. The identification means cannot cover all material objects and thus cannot accurately manage the status of materials.

[0003] Voice interaction is widely used in smart home, smart driving, telephone customer service and other fields. For voice interaction material management, it is generally applied to ground environment. For example, the invention patent application with publication number CN110597975A and patent name of coal mine material management dialogue method, system and all-in-one machine based on Self-Attention mechanism, and the invention patent application with publication number CN113762888A and patent name of a warehouse management intelligent robot have proposed relevant methods, devices, equipment and the like for voice interaction material management. The above-mentioned prior arts are all for voice interaction material management on the ground side, while the voice interaction of space station material management has its particularity in network environment, application level and management object, and is still different from the ground voice interaction system, which is specifically manifested in:

[0004] 1) The space station network is the internal network of the flight control system and cannot be connected to the Internet. It is in an offline state and cannot use the big data support of the Internet;

[0005] 2) The space station has limited on-orbit resources and cannot use large intelligent solution servers with high power and large heat dissipation;

[0006] 3) Space station material management concentrates on management activities such as inquiry, retrieval, installation, disposal, feedback, and transportation. On the basis of identification, corresponding management semantic understanding and process judgment are also required, and multiple rounds of interaction are carried out based on the results of the judgment, and finally the material information data is transmitted to the management system to realize the recording of management data.

[0007] In summary, the existing voice interaction material management solutions suitable for ground environments cannot be applied to space station environments. On this basis, how to improve the accuracy of space station material information management based on voice interaction is a technical problem that needs to be solved urgently. Summary of the invention

[0008] The purpose of the present invention is to overcome the problems of the prior art and provide a space station material information management system and method based on intelligent voice interaction.

[0009] The purpose of the present invention is to achieve the following technical solutions: a space station material information management system based on intelligent voice interaction, the system includes an on-orbit end and a ground end; the on-orbit end includes a mobile terminal and an on-orbit computing node, the mobile terminal includes a voice wake-up and audio acquisition module, a voice broadcast module and a front-end interactive interface, the on-orbit computing node includes a text extraction module, a semantic understanding and logic processing judgment module and an on-orbit business management system connected in sequence, the semantic understanding and logic processing judgment module, the on-orbit business management system are connected with the voice broadcast module, and the on-orbit business management system interacts with the front-end interactive interface for information;

[0010] The text extraction module is used to convert the voice command input by the user into text data; the semantic understanding and logic processing judgment module includes a semantic understanding unit and a logic processing judgment unit. The semantic understanding unit is used to perform word segmentation and semantic analysis on the text data to obtain semantic understanding result data; the logic processing judgment unit is used to analyze the semantic understanding result data, first identify the user's intention by matching keywords and word slots, and then match with the semantic dictionary database data. After the match is successful, the business management data and material dictionary data are obtained, and the business management data and material dictionary data are fed back to the on-orbit business management system. When the keyword and word slot match or the match with the semantic dictionary database data fails, the unrecognized result is fed back to the voice broadcast module or the voice broadcast module is used to ask questions to obtain complete information; the on-orbit business management system is used to perform material management according to the business management data and material dictionary data, and when the semantic understanding and logic processing judgment module still feeds back unrecognized results after multiple rounds of interaction, the text data is transmitted to the ground business management system;

[0011] The ground end includes a ground computing node, which includes a ground business management system and a semantic understanding and data judgment module connected in sequence; the ground business management system is connected to the on-orbit business management system, and is used to receive text data sent by the on-orbit business management system and transmit it to the semantic understanding and data judgment module. The semantic understanding and data judgment module performs semantic analysis on the text data based on the ground intelligent solution service terminal, and feeds back the semantic analysis results to the ground business management system. The ground business management system processes the material data according to the semantic analysis results and feeds back to the on-orbit business management system, thereby enabling the on-orbit business management system to update the material status and display it through the front-end interactive interface.

[0012] In one example, the on-orbit computing node also includes a noise processing module, which is connected to the input end of the text extraction module, and is used to divide the voice command data from high to low frequency, match the background noise of the space station to each divided voice frequency band, and then enhance the contrast between the human voice and the background noise, and perform human voice wave analysis, filtering and extraction to obtain voice data and transmit it to the text extraction module.

[0013] In one example, the text extraction module includes a special noun matching unit and / or a replacement unit; the special noun matching unit corrects the style extraction result according to the special nouns of the stored space station scene; the replacement unit is used to compulsorily replace non-special nouns with special nouns for space station material management.

[0014] In one example, the on-orbit business management system is further integrated with a material location information matching and accounting module for solving the multi-level location information of materials, the multi-level location information including multiple combinations of aircraft, cabin, area, quadrant, channel or multiple combinations of aircraft, cabin, area, quadrant, cell, the material location information matching and accounting module includes:

[0015] A simplified position solving unit, used for solving the simplified position information according to the mapping relationship between the stored simplified position information and the standard multi-level position information; and / or,

[0016] a disordered sequence position solving unit, used to extract position keywords from disordered sequence position information, and solve the disordered sequence position information by permuting and combining position keywords; and / or,

[0017] The unique single-level position solving unit is used to match the corresponding unique standard multi-level position information according to the provided unique single-level position.

[0018] In one example, the on-orbit end also includes an access control system connected to the on-orbit business management system, which stores multi-level location information of materials; the material location information matching and accounting module also includes a location information completion unit, which is used to complete the missing single-level location information according to the multi-level location information of materials stored in the access control system.

[0019] It should be further explained that the technical features corresponding to the above system examples can be combined or replaced with each other to form a new technical solution.

[0020] The present invention also includes a space station material information management method based on intelligent voice interaction, the method comprising the following steps:

[0021] Collect voice commands from space station users;

[0022] Convert voice commands into text data;

[0023] Perform word segmentation and semantic analysis on text data to obtain semantic understanding result data;

[0024] Analyze the semantic understanding result data, first identify the user's intention by matching keywords and word slots, then match with the semantic dictionary database data, obtain business management data and material dictionary data after successful matching, and feed the business management data and material dictionary data back to the on-orbit business management system. When the keyword and word slot match or the match with the semantic dictionary database data fails, feed back the unrecognized result and make a voice broadcast or ask questions through voice broadcast to obtain complete information;

[0025] Carry out material management based on business management data and material dictionary data, and transmit text data to the ground business management system when the semantic understanding and logic processing judgment module still returns no recognition results after multiple rounds of interaction;

[0026] The ground business management system receives the text data sent by the on-orbit business management system and transmits it to the semantic understanding and data judgment module;

[0027] The text data is semantically parsed based on the ground intelligent solution service terminal, and the semantic analysis results are fed back to the ground business management system. The ground business management system processes the material data according to the semantic analysis results and feeds back to the orbital business management system, which enables the on-orbit business management system to update the material status and display it through the front-end interactive interface.

[0028] In one example, before converting the voice command into text data, the method further includes:

[0029] Divide the voice command data according to frequency from high to low;

[0030] Match the background noise of the space station to each divided speech frequency band;

[0031] Enhance the contrast between human voice and background noise;

[0032] Perform human voice wave analysis, filtering and extraction to obtain voice data.

[0033] In one example, converting the voice command into text data further includes:

[0034] Correct the text extraction results according to the stored special nouns of the space station scene; and / or, forcibly replace non-special nouns with special nouns for space station material management.

[0035] In one example, after the unrecognized result is fed back, the multi-level location information calculation step of the material is also included:

[0036] The simplified position information is solved according to the mapping relationship between the stored simplified position information and the standard multi-level position information; the multi-level position information includes multiple combinations of aircraft, cabins, areas, quadrants, channels or multiple combinations of aircraft, cabins, areas, quadrants, cells; and / or,

[0037] Extracting position keywords from the disordered position information, and solving the disordered position information by permuting and combining the position keywords; and / or,

[0038] Match the corresponding unique standard multi-level location information according to the unique single-level location provided.

[0039] In one example, after the unrecognized result is fed back, the following further includes:

[0040] The missing single-level location information is completed based on the multi-level location information of materials stored in the access control system.

[0041] It should be further explained that the technical features corresponding to the various examples of the above method can be combined or replaced with each other to form a new technical solution.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. In one example, algorithm processing is performed according to the needs of space station material management services and scenarios. Due to the high network latency and limited resources on the orbit side, it is impossible to use a high-power large-scale intelligent solution server. When the on-orbit semantics of the present invention cannot be processed, the on-orbit business management system transmits the data to the ground natural semantic understanding engine (intelligent solution service terminal) for processing, thereby obtaining a highly accurate semantic understanding result, and then the business data is updated on the ground, thereby supporting the on-orbit semantic understanding and improving the accuracy of on-orbit material management. At the same time, the voice interaction mechanism and the semantic understanding mechanism adopt a universal design, which is completely separated from the on-orbit business management system for material management. After accurate voice recognition and conversion, semantic extraction of business keywords in different on-orbit business management systems, multiple rounds of interactive matching, data transmission, and semantic understanding of different businesses are performed through logical processing algorithms, thereby realizing functional reuse.

[0044] 2. In one example, the noise processing module can separate background noise and obtain pure human voice, providing a more accurate speech recognition basis for subsequent text extraction and semantic understanding.

[0045] 3. In one example, the text extraction results are corrected through the special noun matching unit, providing a more accurate speech recognition basis for subsequent semantic understanding; the non-special nouns are forcibly replaced with special nouns for space station material management through the replacement unit to maintain the consistency and professionalism of the text, thereby effectively identifying the special nouns for space station environmental material management and improving the accuracy and reliability of material management.

[0046] 4. In one example, by solving the multi-level location information of materials, the location information of materials at different levels such as aircraft, cabins, areas, quadrants, and channels can be solved to ensure the accuracy of the material location; at the same time, by simplifying the position solving unit, the disordered sequence position solving unit and the only single-level position solving unit, the module can effectively match and calculate the material location information, thereby improving the accuracy and efficiency of material management.

[0047] 5. In one example, the multi-level location information of materials stored in the access control system is used to complete the missing single-level location information, and standard multi-level location information can be obtained to ensure the integrity of the material location, thereby improving the accuracy and efficiency of material management. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The specific implementation methods of the present invention are further described in detail below in conjunction with the accompanying drawings. The accompanying drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The same reference numerals are used in these drawings to represent the same or similar parts. The schematic embodiments of the present application and their descriptions are used to explain the present application and do not constitute improper limitations on the present application.

[0049] Figure 1 A system framework diagram provided for an example of the present invention;

[0050] Figure 2 A system framework diagram provided for a preferred example of the present invention;

[0051] Figure 3 A flow chart of a method provided for an example of the present invention. DETAILED DESCRIPTION

[0052] The technical solution of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments 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 technicians in this field without creative work are within the scope of protection of the present invention.

[0053] In the description of the present invention, it should be noted that the directions or positional relationships indicated by "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc. are directions or positional relationships based on the drawings, which are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the use of ordinal numbers (e.g., "first and second", "first to fourth", etc.) is to distinguish objects, and is not limited to this order, and cannot be understood as indicating or implying relative importance.

[0054] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, "installation", "connection" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0055] In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0056] In one example, a space station material information management system based on intelligent voice interaction, such as Figure 1 As shown, the system includes an on-orbit end and a ground end, and the on-orbit end communicates with the ground end via the space-ground network. Among them, the on-orbit end includes a mobile terminal and an on-orbit computing node. The mobile terminal is integrated with a voice wake-up and audio acquisition module, a voice broadcast module and a front-end interactive interface. The on-orbit computing node can be an on-orbit computer, including a text extraction module, a semantic understanding and logic processing judgment module and an on-orbit business management system connected in sequence. The semantic understanding and logic processing judgment module, the on-orbit business management system are connected to the voice broadcast module, and the on-orbit business management system interacts with the front-end interactive interface. The ground end includes a ground computing node, which can be a ground computer, including a ground business management system and a semantic understanding and data judgment module connected in sequence. The on-orbit business management system communicates with the ground business management system via the space-ground network.

[0057] Optionally, the on-orbit computing node also includes an on-orbit database of the on-orbit business management system, and the on-orbit database stores material information, including material name information, material location information (including material arrangement and combination location information, i.e., the arrangement and combination of multi-level location information), and the mapping relationship between each material arrangement and combination location information and the standard location relationship. When the on-orbit business management system is not set in the on-orbit database, the material name information, material status information, material arrangement and combination location information, and the mapping relationship between each material arrangement and combination location information and the standard location relationship are stored through the storage module of the on-orbit business management system. The ground computing node also includes a ground database of the ground business management system, which is used to synchronize the data information stored in the on-orbit database. In order to reduce the amount of data transmission, only incremental data is synchronized, so that the audio and text of the user's current round of interaction can be transmitted to the ground for verification.

[0058] Optionally, the ground end also includes a user end, which includes a business software interface that interacts with the ground business management system, is used to manage material data, and feeds back material execution results to the user after this round of interaction is completed.

[0059] Furthermore, the voice wake-up and audio collection module, the voice broadcast module, the text extraction module, the semantic understanding and logic processing judgment module constitute the voice interaction system. Among them, the voice wake-up and audio collection module includes a voice wake-up unit and an audio collection unit. The voice wake-up unit is the basis of the voice interaction system, which is used to realize real-time voice monitoring of mobile devices such as mobile phones, thereby enabling intelligent voice interaction. When the voice wake-up unit is in a normal wake-up state, the status is displayed as "1", and then the audio collection unit collects audio data (voice instructions) and passes the audio data to the text extraction module in real time. When the voice wake-up unit is in an unwake-up state, the status is displayed as "0". At this time, the audio collection unit does not work and does not pass audio data to the text extraction module. In this example, the voice wake-up and audio collection module and the text extraction module transmit data through the communication network in the space station to transmit audio data in real time.

[0060] Furthermore, the speech synthesis broadcast module can be a speaker, etc., and the data format it receives is text. In the semantic understanding and logic processing judgment module, the on-orbit business management system execution and other links, a mechanism for outputting text data to the speech synthesis broadcast module will be set. The speech synthesis broadcast module will broadcast the received text in real time, and when the data is stacked, it will be broadcast in sequence. The semantic synthesis broadcast module is the bridge of the voice interaction system, which can broadcast the results of the on-orbit business management system execution to the user and provide correct feedback results for intelligent voice interaction.

[0061] Furthermore, the text extraction module is the key to the voice interaction system. It needs to accurately identify the user's voice commands and convert them into text data to output correct voice data for intelligent voice interaction. Specifically, the text extraction module can be a language model, such as a BERT model or a transformer model, which can accurately identify the entities of voice data (material names, material codes, business commands, location nouns, etc.), and arrange material-related nouns, attribute-related quantifiers, business command-related verbs, and location-related nouns into a keyword combination that the machine can understand, and output it to the semantic understanding and logic processing judgment module.

[0062] Furthermore, the semantic understanding and logic processing judgment module is the core of the voice interaction system, which has intent matching and extraction, semantic understanding and judgment, multi-round interaction and question-answering, and outputs material interface data to the on-orbit business management system to process the correct execution logic and information for intelligent voice interaction. The semantic understanding and logic processing judgment module includes a semantic understanding unit and a logic processing judgment unit. The semantic understanding unit is used to perform word segmentation and semantic analysis on text data to obtain semantic understanding result data. After the semantic understanding unit obtains the semantic understanding result data and extracts key information, it performs semantic data business matching, multi-round interaction logic management, and business interface data call processing on the semantic understanding result data, thereby achieving comprehensive and accurate processing of semantic content and on-orbit business management. Through internal processing in different situations, external data transmission, and multi-round user interaction, the management of material information data is finally realized. The logic processing judgment unit is used to analyze the semantic understanding result data, first identify the user's intention by matching keywords and word slots, and then match with the semantic dictionary database data. If the match is successful, the business management data and material dictionary data are obtained, and the business management data and material dictionary data are fed back to the on-orbit business management system. When the keyword and word slot match or the semantic dictionary database data fails to match, the unrecognized result is fed back to the voice broadcast module or the voice broadcast module is used to ask questions to obtain complete information. Preferably, for the space station material management scenario, key elements such as the dictionary library and the intention library are established, and the dictionary library and the intention library are integrated into the logic processing judgment unit. In the dictionary library, the material name, management business, material location, product code and number that the text extraction module and the on-orbit business management system need to execute need to be fully entered according to the space station rules, and similar vocabulary is configured for the dictionary content with synonyms to facilitate the matching of semantic understanding. It is necessary to establish an intent library for material management activities, and to establish descriptive speech patterns for certain material management activities in the corresponding intent library, and to establish the necessary word slots for identifying this need, so as to facilitate the system's intent recognition and matching. Both the word slots and extended questions need to be based on dictionaries and associated with dictionaries. Furthermore, corresponding management processes are set for different material management activities. The process needs to set the wake-up and initial speech at the beginning, and enter the extended questions and word slots through the intent matching process. The extracted word slot dictionary content is then passed to the business software through the set data interface for business processing of this material. At the same time, the content with follow-up questions, feedback, etc. is also passed to the speech synthesis broadcast module through the data interface for broadcast.

[0063] The text extraction module and the semantic understanding and logic processing judgment module are all configured on the on-orbit computing node and can directly interact with data. The semantic understanding and logic processing judgment module uses its own database to realize the configuration and call of the dictionary of material names and action names, the configuration and call of intention expansion questions and word slots, the configuration and call of the material management voice interaction process, and the configuration and call of the semantic understanding expansion interface and input and output data. The database is installed and run in an integrated manner when the voice modules are installed. Data revision and deletion can be realized only through front-end web operations, and the semantic database envelope is larger than the material database content, which is convenient for identifying content conversion.

[0064] Furthermore, the on-orbit business management system is used to manage materials based on business management data and material dictionary data, and transmits text data to the ground business management system when the semantic understanding and logic processing judgment module still feedbacks unrecognized results after multiple rounds of interaction. Specifically, after receiving the interface data output by the semantic understanding unit, the on-orbit business management system will perform corresponding business actions and records, and feedback the execution result data to the voice interaction system, thereby realizing material information management in the intelligent voice interaction mode. In particular, when the material location information described by the user is not standard, the on-orbit business management system will compare it with the material arrangement and combination in the database and then give the standard location name information.

[0065] Furthermore, the ground business management system is connected to the on-orbit business management system to receive text data sent by the on-orbit business management system and transmit it to the semantic understanding and data judgment module. The semantic understanding and data judgment module performs semantic analysis on the text data based on the ground intelligent solution service terminal, and feeds back the semantic analysis results to the ground business management system. The ground business management system processes the material data according to the semantic analysis results and feeds back to the on-orbit business management system, thereby enabling the on-orbit business management system to update the material status and display it through the front-end interactive interface.

[0066] It should be noted that due to the limitation of on-orbit resources, high-power intelligent solution servers cannot be used, and the space-ground network has high latency and the measurement and control area cannot communicate in real time. In order to improve the convenience of users using voice interaction functions on-orbit, the present invention proposes a space-ground integrated semantic understanding model composed of an on-orbit precise sentence processing service and a ground-based natural sentence processing service. The on-orbit precise sentence processing service can understand and extract keywords of normal sentences, such as: "Help me check the location of XX materials", "Help me check the details of XX materials", "XX materials have been taken", "XX materials have been taken", "Transfer XX materials to XX location", "Transfer XX materials at XX location to XX location", "Feedback record", "Discard XX to XX location" and other standard sentences. Through the semantic understanding and logic processing judgment module, it can be understood and parsed, and the keyword slot content can be extracted to the on-orbit business management system for business processing. However, if there are multiple materials and multiple locations in a sentence, such as: "XX materials, XX materials, XX materials from XX location have been taken", "XX materials from XX location are transferred to XX location, X pieces of XX materials from XX location are transferred to XX location", etc., as well as descriptions of batch material transfers, when the semantic understanding and logical processing judgment module alone cannot complete the task, the on-orbit business management system will transmit the recognized text to the high-power natural sentence ground intelligent solution service terminal set up on the ground. After the sentence is understood and output as a normal understanding sentence by the high-power natural sentence ground intelligent solution service terminal (loaded with a large language model, such as the GPT series model), the extracted single instruction keyword is output to the ground on-orbit business management system for business processing. After the ground on-orbit business management system processes the material data according to standard keywords, it will be synchronized to the on-orbit server through the ground-ground synchronization software. When the on-orbit server completes the data update, it will send the task execution data of the natural sentence to the mobile phone APP page, prompting the user that the business statement that was not understood on-orbit has been executed and the execution result is displayed (the sentence text content and the execution result are displayed at the same time on the page for user confirmation).

[0067] In one example, if Figure 2As shown, the on-orbit computing node also includes a noise processing module, which is connected to the input end of the text extraction module, and is used to divide the voice command data from high to low frequency, match the background noise of the space station to each voice frequency band, and suppress the background noise in combination with the characteristics of the specific distinction between human voice and background noise in sound waves, thereby enhancing the contrast between human voice and background noise, and then separating, filtering, and extracting human voice waves for the optimized audio data, thereby forming pure human voice audio for human voice wave analysis, filtering, and extraction, obtaining voice data, and finally transmitting the voice data to the text extraction module. The space station noise processing module is the bridge of the voice interaction system, which can eliminate and control the unique background noise of the space station, thereby outputting purer human voice data to improve the recognition accuracy and provide higher quality voice data for the text extraction module.

[0068] In one example, the text extraction module includes a special noun matching unit and / or a replacement unit, preferably including a noun matching unit and a replacement unit. The special noun matching unit corrects the style extraction result according to the special nouns of the stored space station scene, that is, the hot word configuration of the special nouns is required for the space station material management, such as "solar wing", "calcium removal", etc., to improve the recognition conversion accuracy. The replacement unit is used to compulsorily replace non-special nouns with special nouns for space station material management based on the same pronunciation or similar pronunciation. Specifically, a mandatory conversion replacement list is set for the special nouns for material management that cannot be accurately converted into sentence descriptions, such as Xianshan replaced by Quadrant, Tianhe replaced by Tianhe, Tianzhouba replaced by TZ-8, etc., to improve the consistency accuracy of matching nouns with space station special nouns and material management databases. At the same time, it is also necessary to train the language model, train the user's speech audio with the standard text, and improve the accuracy of user accent recognition.

[0069] In one example, the on-orbit business management system is also integrated with a material location information matching and accounting module, which is used to solve the multi-level location information of materials. The multi-level location information includes multiple combinations of aircraft, cabins, regions, quadrants, and channels or multiple combinations of aircraft, cabins, regions, quadrants, and cells, preferably five-level location information formed in the order of aircraft, cabins, regions, quadrants, and channels / cells. Specifically, the space station has the characteristics of multiple aircraft, multiple cabins, multiple regions, weightlessness and non-directionality. The precise management of space station materials has five-level definitions in aircraft, cabins, regions, quadrants, and channels / cells. The on-orbit business database uses the five-level location definition as the standard storage format for material information. The five-level location standard format brings certain difficulties to the user's voice interaction management of materials, that is, it requires the user to say all levels, each level needs to be said correctly, the order of each level needs to match, and each cabin has the same area and quadrant. In order to improve the efficiency and user experience of voice interaction management of material location information, a semantic understanding unit is introduced to solve the multi-level location information of materials. When the location description in the semantic understanding and logic processing judgment module is non-standard and cannot be understood, the accurate location information will be output after being solved by the material location information matching and accounting module of the on-orbit business management system, which is convenient for data updating and execution of the on-orbit business management system.

[0070] At this time, the material location information matching and accounting module includes a simplified location solution unit and / or a disordered order location solution unit and / or a unique single-level location solution unit, preferably including the above three units. Among them, the simplified location solution unit is used to solve the simplified location information according to the mapping relationship between the stored simplified location information and the standard multi-level location information, such as mapping the simplified location information XX-sleeping area 1 with the standard location information XX-working cabin-A area-IV quadrant-grid, and calculating and solving the missing non-critical compartment information and matching the corresponding location information. At the same time, in the algorithm, the description method can be simplified for unnecessary location levels, and the simplified description location information is mapped with the standard multi-level location information, such as: XX-A area-IV quadrant-grid omits the unnecessary "working cabin" segment, and XX-A area-IV quadrant uses an algorithm form in which the description definition without grid is defined as a channel in addition to omitting the unnecessary "working cabin". At this time, by querying the corresponding mapped standard multi-level location information, the location name simplified description statement can be realized. Position matching accounting and solving. The disordered position solving unit is used to extract the position keywords in the disordered position information, and solve the disordered position information by using the permutation and combination of position keywords. Specifically, when the position description order is disordered, such as the existence of many permutations and combinations such as work cabin-XX-A zone-IV quadrant-grid, work cabin-XX-IV quadrant-A zone-grid, A zone-grid-work cabin-XX-IV quadrant, etc., after the position content is output to the disordered position solving unit through voice, the disordered position solving unit adopts a mechanism of multiple permutations and combinations of positions corresponding to one accurate position to realize position verification and solution. The unique single-level position solving unit is used to match the corresponding unique standard multi-level position information according to the provided unique single-level position. Specifically, the space station has only one airlock, rear-end channel, node module and other material storage areas, and combined with the location characteristics of the small interior, no multiple area divisions, and only quadrants and channels, when the voice description says: "Put XX material in the airlock / node module", the on-orbit business management system can automatically match the location information to XX-airlock / node module through multiple locations corresponding to a unique location, and then bind it to the accounting and solution method of quadrant I by default, providing a standard material information processing format for the materials (after settlement and accounting: XX-airlock / node module-quadrant I), and the material location information can be updated without further interaction.

[0071] Furthermore, if the standard multi-level location information cannot be formed or the identification error occurs after the location is calculated through the material location information matching and calculation module, such as: "Area A-IV quadrant-grid, random sentence in the middle, work cabin-XX"; "Put XX material in the upper left corner of a cabinet in XX-A area / IV-III quadrant position / XX bag", the voice broadcast module starts the sentence follow-up question: Please say the location name again. After the user says the location again, the above-mentioned solution and accounting processing will be performed. If the location information still cannot be matched in the end, the user will be fed back that this business will be handed over to the ground business management system for processing. The ground will then analyze and understand it through the ground intelligent solution service terminal, and output the regular location name. After the material location information is processed by the ground on-orbit business management system, it will be synchronized into orbit to complete the update of the material location information.

[0072] In one example, if Figure 2 As shown, the on-orbit end also includes an access control system connected to the on-orbit business management system, and the access control system stores materials and multi-level location information of materials. The material location information matching and accounting module also includes a location information completion unit, which is used to complete the missing single-level location information according to the multi-level location information of materials stored in the access control system. Specifically, if there is no compartment-level location information in the location information, the material location information matching and accounting module calls the material flow compartment location information monitored by the access control system, and automatically completes the compartment information of the material location, so that the material location information forms a standard five-level location information format data. The standard location information is then processed together with keywords such as material name and business type, and the data update of the material location is finally completed.

[0073] The present invention also includes a space station material information management method based on intelligent voice interaction, which has the same inventive concept as the above-mentioned space station material information management system, such as Figure 3 As shown, the method comprises the following steps:

[0074] S1: Collect voice commands from space station users.

[0075] Specifically, the user wakes up the voice interaction system through the wake-up word. After the wake-up unit receives the set wake-up word, it will prompt "Please speak" as a feedback word. At this time, voice recording is turned on and the user's voice instructions are collected through the audio collection unit.

[0076] S2: Convert the voice command into text data.

[0077] The voice commands are transcribed in real time. Specifically, the voice data is read through the BERT model, transformer model and other large voice models for conversion and matching, so as to achieve the text content required by the user. If no voice command is received within 2 seconds after receiving the voice or the user says "finished" to end the voice input, the text converted from this voice will be automatically output to the semantic understanding unit.

[0078] S3: Perform word segmentation and semantic analysis on the text data to obtain semantic understanding result data.

[0079] The text data is segmented and identified, extracted, matched, processed, interacted, and called according to the judgment logic set by the semantic logic processing judgment algorithm. The analysis, judgment, questioning, feedback, prompting, processing of semantic data results, and business data interaction and management functions are realized for various voice material management conditions.

[0080] S4: Analyze the semantic understanding result data, first identify the user intention by matching keywords and word slots, and then match it with the semantic dictionary database data. If the match is successful, obtain the business management data and material dictionary data, and feed the business management data and material dictionary data back to the on-orbit business management system. When the keyword and word slot match or the match with the semantic dictionary database data fails, feedback the unrecognized result and make a voice broadcast or ask questions through voice broadcast to obtain complete information.

[0081] Specifically, the keyword results of semantic extraction (semantic understanding result data) are matched with the set management intent extension questions and word slots. If the recognized word segmentation text can match the word slot content of the intent extension question, the semantic dictionary database data matching is entered according to the judgment process; if the recognized word segmentation text cannot match the word slot content of the intent extension question, the feedback "The intent was not matched and has been handed over to the ground for processing and updating" is fed back and announced by the speech synthesis module, and the spoken sentence is introduced into the next logical judgment. According to the extension question and word slot, after the intent is matched, but the keyword slot information of the extension question is missing, the voice announcement module is used to ask "Please tell me a certain feature", and the spoken sentence is introduced into the next logical judgment. If it is judged to be an incomprehensible non-standard sentence, it will be processed by the ground end. The text data will be sent to the ground business management system via the on-orbit business management system, and then transmitted to the ground intelligent solution service terminal for analysis, extraction, and sorting, and then a standard and understandable material business sentence will be output. After the ground business management system updates the business data, the processing result will be uploaded to the on-orbit business management system for updating, and the mobile phone will be prompted that the material business of this sentence has been processed and updated.

[0082] S5: Perform material management based on business management data and material dictionary data, and transmit text data to the ground business management system when the semantic understanding and logic processing judgment module still returns unrecognized results after multiple rounds of interaction.

[0083] Specifically, after the on-orbit business management system obtains the business management data and material dictionary data, it matches it with its own database data and makes changes according to the management business. If it is executed correctly, the feedback "a certain business of a certain material has been executed" will be fed back and announced by the speech synthesis module, ending this round of dialogue. If the business cannot be executed due to the status of the material, the feedback "the material is in a certain state and a certain business cannot be performed" will be fed back to end this round of dialogue. If this material in the business database has been downloaded and does not exist, the feedback "the material has been downloaded and a certain business cannot be performed" will be fed back to end this round of dialogue. If there are multiple materials that cannot be matched to a specific material, the feedback "there are multiple materials, please select through the mobile phone" will be fed back, and all the relevant materials found will be pushed to the front-end interactive interface for the user to select and execute. When there are multiple auxiliary equipment objects, the feedback "please say device X or device Y" will be given, and the semantic dictionary database data matching link will be entered into the follow-up link again.

[0084] Furthermore, during the operation of the voice interaction system, the on-orbit business management system continuously inputs and outputs data with the semantic understanding and logic processing judgment module to ensure the accurate recording and change of material information. After the data is changed, it is recorded and unified in the database of the ground-to-earth material information management system. The on-orbit business management system will also transmit the text and audio of this conversation to the ground for ground personnel to verify the information.

[0085] S6: The ground business management system receives the text data sent by the on-orbit business management system and transmits it to the semantic understanding and data judgment module. It performs semantic analysis on the text data based on the ground intelligent solution service terminal and feeds back the semantic analysis results to the ground business management system. The ground business management system processes the material data according to the semantic analysis results and feeds back to the on-orbit business management system, thereby enabling the on-orbit business management system to update the material status and display it through the front-end interactive interface.

[0086] Specifically, the on-orbit business management system will transmit the recognized text to the high-power natural sentence ground intelligent solution service terminal set up on the ground. After the high-power natural sentence ground intelligent solution service terminal understands the sentence and outputs it as a normally understood sentence, the extracted single instruction keyword will be output to the ground on-orbit business management system for business processing. After the ground on-orbit business management system processes the material data according to the standard keywords, it will synchronize it to the on-orbit server through the ground-to-space synchronization software. When the on-orbit server completes the data update, it will send the task execution data of the natural sentence to the mobile phone APP page, prompting the user that the business statement that was not understood on-orbit has been executed and the execution result is displayed (the sentence text content and the execution result are displayed at the same time on the page for user confirmation).

[0087] In one example, when the voice command is converted into text data inaccurately, resulting in semantic misunderstanding, the user is given an execution reminder and is prompted with instructions on how to proceed with the next action:

[0088] 1) Error in text data extraction and no such material exists in the semantic understanding database. In this case, there are two reasons: the material information is not configured in the database of the on-orbit business management system (not entered into the material information management system, but there is a physical object) and there is no such material in fact (no such information and no physical object). However, at this time, the space station material information management cannot confirm the reason. It will only determine that the entity word is not matched, which will trigger a round of interaction: XX material does not exist, please say it again. If the entity word in the semantic understanding is still not matched again, the final feedback will be given: XX material does not exist, and your management operation has been passed to the ground for review. The ground will check the voice data transmitted down and reply to the user after verification: the material status has been updated (it is assumed that what the user said in the voice must actually exist in orbit).

[0089] 2) The text data extraction is wrong and the material exists in the semantic understanding database. Since the text extraction module itself has a certain recognition accuracy, the text extraction module cannot confirm that the output text data is completely correct. It will only compare it with the configuration words in the semantic understanding entity slot database. If the comparison is successful, it is considered that the text entity is extracted correctly. If the comparison is wrong, it is considered that there is no extraction abnormality in the text entity. In this case, the keyword combination is output to the on-orbit business management system, which is processed by the on-orbit business management system. There are two branch situations in the processing of the on-orbit business management system: logical errors (the material is no longer there or the status cannot be executed) and logical correctness (normal business processing can be performed). When the logic is wrong, the on-orbit business management system will feedback to the speech synthesis: because XX material is already in XX status, the command cannot be executed, please confirm whether it needs to be processed on the ground. The ground receives the synchronous data that needs to be processed, reviews the data, and gives the user feedback on the data update. When the logic is correct, the on-orbit business management system will process it normally and feedback to the speech synthesis: XX material information has been taken / transferred, etc. At this time, the voice and material change status synchronized to the ground will be checked by the ground personnel to correct the material information data.

[0090] 3) The text data is extracted correctly, but there is no such material in the semantic understanding database. This situation is the same as 1). There are two reasons: the material information is not configured in the database (not entered into the material information management system, but there is a physical object) and there is no such material in fact (no such information, no physical object). However, it is unclear what the reason is, and it will only be determined that the entity word is not matched. The same management strategy as 1) will be adopted.

[0091] 4) The text data is extracted correctly but the material information in the business database cannot be changed. In this case, for example, the consumed material entity will be identified as taken and output to the on-track business management system. The on-track business management system will process the logic error and adopt the same management strategy as the logic error in 2).

[0092] 5) Multiple corrections lead to abnormal understanding. In the existing feedback, there are corrections to the material names, codes and locations in the user's voice description. At present, the semantic understanding and logic processing judgment module on the track side does not have big data and AI solving capabilities, and cannot support judgment. It is necessary to configure the correction description to achieve this business processing as much as possible. When it is indeed impossible to extract effective and understandable keyword combinations normally, a new round of interaction will be triggered: You said you don’t understand, please say it again. At the same time, when training users, it is also proposed to use sentences that can be understood by the semantic understanding and logic processing judgment module as much as possible when describing the voice, so as to improve the user experience and information execution accuracy.

[0093] 6) Abnormal logical understanding. The combination of the previous and the next understanding is misplaced, and the intent and word slot cannot be matched. The material management business content to be carried out needs to be said again.

[0094] 7) The understanding of two materials and two types of management in one sentence is abnormal. In the existing feedback, there are sentences in the user's voice description that perform two different management actions on materials. The semantic understanding will extract keywords and output them to the on-track business management system. The on-track business management system can perform segmentation processing, but in order to ensure that complex descriptions are not executed abnormally, it is recommended that when training users, the language model used by the on-track business management system should be used to identify the sentences in the voice description to improve the user experience and information execution accuracy. In different states, the semantic understanding and logic processing judgment module outputs different business data values ​​to the on-track business management system, which facilitates the on-track business management system to distinguish between the output data and status, and at the same time sets the execution program of the corresponding value in the on-track business management system.

[0095] In one example, before converting the voice command into text data, a background noise processing step is also included:

[0096] S121: dividing the voice command data according to frequency from high to low;

[0097] S122: Matching the background noise of the space station to each divided speech frequency band;

[0098] S123: suppressing background noise and enhancing the contrast between human voice and background noise;

[0099] S124: Perform human voice wave analysis, filtering and extraction, output purer human voice data to improve recognition accuracy, and provide higher quality voice data for the text extraction module.

[0100] In one example, converting the voice command into text data further includes:

[0101] S21: Modify the text extraction results according to the stored special nouns of the space station scene, that is, configure hot words of special nouns according to the needs of space station material management, such as "solar wing", "calcium removal", etc., to improve the recognition conversion accuracy.

[0102] S22: Compulsory replacement of non-special nouns with special nouns for space station material management. Specifically, a mandatory conversion replacement list is set for the special nouns for material management that cannot be accurately converted into sentence descriptions, such as replacing Xianshan with Quadrant, Tianhe with Tianhe, and Tianzhouba with TZ-8, etc., to improve the consistency accuracy of nouns matching space station special nouns and material management databases.

[0103] In this example, steps S21 and S22 may be performed simultaneously, or only step S21 or S22 may be performed.

[0104] In one example, after matching with the semantic dictionary database data, if a matching dictionary is obtained, the business management data and material dictionary data will be passed to the on-track business management system. If the voice dictionary database data is not matched, the feedback "No corresponding material data found, please say the material name again" will be reported by the speech synthesis module, and the spoken sentence will be introduced into the next semantic dictionary database data match. When describing the location of the transfer business by voice, if the location is described standardly and recognized correctly, the location information will be directly passed to the next process. If the material location information description is inaccurate, it will be automatically passed to the on-track business management system. The material location information matching and accounting module of the on-track business management system will perform multi-level location information solution, and output the correct location sorting to the voice. After the semantic verification and confirmation of the location dictionary, all business data will be output to the on-track business management system to update the material location information.

[0105] Specifically, after the unrecognized result is fed back, that is, when the location description in the semantic understanding and logic processing judgment module is not standard and cannot be understood, the multi-level location information solution step of the material is introduced:

[0106] S451: Calculate the simplified location information according to the mapping relationship between the stored simplified location information and the standard multi-level location information. For example, the simplified location information XX-sleeping area 1 is mapped with the standard location information XX-working cabin-A zone-IV quadrant-grid, and the missing non-critical cabin information and the corresponding location information are calculated and solved. At the same time, the description method can be simplified for unnecessary location levels in the algorithm, and the simplified description location information is mapped with the standard multi-level location information, such as: XX-A zone-IV quadrant-grid omits the unnecessary "working cabin" segment, and XX-A zone-IV quadrant, in addition to omitting the unnecessary "working cabin", also adopts the algorithm form of a channel without a grid description definition. At this time, by querying the corresponding mapped standard multi-level location information, the location name simplified description statement can be calculated and solved.

[0107] S452: extracting the position keywords in the disordered position information, and using the permutation and combination position keywords to solve the disordered position information. When the position description order is disordered, such as the presence of many permutations and combinations such as work cabin-XX-A zone-IV quadrant-grid, work cabin-XX-IV quadrant-A zone-grid, A zone-grid-work cabin-XX-IV quadrant, etc., after the position content is output to the disordered position solving unit through voice, the disordered position solving unit uses a mechanism that multiple permutation and combination positions correspond to one accurate position to realize position verification and solving.

[0108] S453: Match the unique standard multi-level location information according to the unique single-level location provided. Specifically, for the space station with only one material storage area such as the airlock, rear channel, node module, etc., combined with the location characteristics of the small interior, no multiple area divisions, and only quadrants and channels, when the voice description is: "Put XX material in the airlock / node module", the on-orbit business management system can automatically match the location information to XX-airlock / node module through multiple locations corresponding to a unique location, and then bind it to the calculation and solution method of quadrant I by default, providing a standard material information processing format for the materials (after settlement and accounting: XX-airlock / node module-quadrant I), and the material location information can be updated without interaction again.

[0109] In this example, steps S451, S452 and S453 may be executed simultaneously, or only step S451, S452 or S453 may be executed.

[0110] In one example, after the unrecognized result is fed back, that is, when the location description in the semantic understanding and logic processing judgment module is not standard and cannot be understood, it also includes:

[0111] The missing single-level location information is completed based on the multi-level location information of materials stored in the access control system.

[0112] Specifically, if the position information is missing and the standard position information cannot be formed after multi-level position information solution, the material compartment position data recorded by the access control system is retrieved and combined with the voice position to form the standard position information for the voice. After semantic verification and confirmation of the position dictionary, it is output to the on-orbit business management system to update the material location information.

[0113] The above specific implementation methods are detailed descriptions of the present invention. It cannot be determined that the specific implementation methods of the present invention are limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions and substitutions can be made without departing from the concept of the present invention, which should be regarded as belonging to the protection scope of the present invention.

Claims

1. A space station material information management system based on intelligent voice interaction, characterized in that: It includes an on-orbit end and a ground end; the on-orbit end includes a mobile terminal and an on-orbit computing node, the mobile terminal includes a voice wake-up and audio acquisition module, a voice broadcast module and a front-end interactive interface, the on-orbit computing node includes a text extraction module, a semantic understanding and logic processing judgment module and an on-orbit business management system connected in sequence, the semantic understanding and logic processing judgment module, the on-orbit business management system are connected with the voice broadcast module, and the on-orbit business management system interacts with the front-end interactive interface for information; The text extraction module is used to convert the voice command input by the user into text data; the semantic understanding and logic processing judgment module includes a semantic understanding unit and a logic processing judgment unit. The semantic understanding unit is used to perform word segmentation and semantic analysis on the text data to obtain semantic understanding result data; the logic processing judgment unit is used to analyze the semantic understanding result data, first identify the user's intention by matching keywords and word slots, and then match with the semantic dictionary database data. After the match is successful, the business management data and material dictionary data are obtained, and the business management data and material dictionary data are fed back to the on-orbit business management system. When the keyword and word slot match or the match with the semantic dictionary database data fails, the unrecognized result is fed back to the voice broadcast module or the voice broadcast module is used to ask questions to obtain complete information; the on-orbit business management system is used to perform material management according to the business management data and material dictionary data, and when the semantic understanding and logic processing judgment module still feeds back unrecognized results after multiple rounds of interaction, the text data is transmitted to the ground business management system; The ground end includes a ground computing node, which includes a ground business management system and a semantic understanding and data judgment module connected in sequence; the ground business management system is connected to the on-orbit business management system, and is used to receive text data sent by the on-orbit business management system and transmit it to the semantic understanding and data judgment module. The semantic understanding and data judgment module performs semantic analysis on the text data based on the ground intelligent solution service terminal, and feeds back the semantic analysis results to the ground business management system. The ground business management system processes the material data according to the semantic analysis results and feeds back to the on-orbit business management system, thereby enabling the on-orbit business management system to update the material status and display it through the front-end interactive interface.

2. The space station material information management system based on intelligent voice interaction according to claim 1 is characterized in that: The on-orbit computing node also includes a noise processing module, which is connected to the input end of the text extraction module, and is used to divide the voice command data from high to low frequency, match the background noise of the space station to each divided voice frequency band, and then enhance the contrast between the human voice and the background noise, and perform human voice wave analysis, filtering and extraction to obtain voice data and transmit it to the text extraction module.

3. The space station material information management system based on intelligent voice interaction according to claim 1 is characterized in that: The text extraction module includes a special noun matching unit and / or a replacement unit; the special noun matching unit corrects the style extraction result according to the special nouns of the stored space station scene; the replacement unit is used to compulsorily replace non-special nouns with special nouns for space station material management.

4. The space station material information management system based on intelligent voice interaction according to claim 1 is characterized in that: The on-orbit business management system is also integrated with a material location information matching and accounting module, which is used to solve the multi-level location information of materials. The multi-level location information includes multiple combinations of aircraft, cabins, areas, quadrants, and channels or multiple combinations of aircraft, cabins, areas, quadrants, and cells. The material location information matching and accounting module includes: A simplified position solving unit, used for solving the simplified position information according to the mapping relationship between the stored simplified position information and the standard multi-level position information; and / or, a disordered sequence position solving unit, used to extract position keywords from disordered sequence position information, and solve the disordered sequence position information by permuting and combining position keywords; and / or, The unique single-level position solving unit is used to match the corresponding unique standard multi-level position information according to the provided unique single-level position.

5. The space station material information management system based on intelligent voice interaction according to claim 1 is characterized in that: The on-orbit end also includes an access control system connected to the on-orbit business management system, which stores multi-level location information of materials; the material location information matching and accounting module also includes a location information completion unit, which is used to complete the missing single-level location information according to the multi-level location information of materials stored in the access control system.

6. A space station material information management method based on intelligent voice interaction, characterized in that: The method comprises the following steps: Collect voice commands from space station users; Convert voice commands into text data; Perform word segmentation and semantic analysis on text data to obtain semantic understanding result data; Analyze the semantic understanding result data, first identify the user's intention by matching keywords and word slots, then match with the semantic dictionary database data, obtain business management data and material dictionary data after successful matching, and feed the business management data and material dictionary data back to the on-orbit business management system. When the keyword and word slot match or the match with the semantic dictionary database data fails, feed back the unrecognized result and make a voice broadcast or ask questions through voice broadcast to obtain complete information; Carry out material management based on business management data and material dictionary data, and transmit text data to the ground business management system when the semantic understanding and logic processing judgment module still returns no recognition results after multiple rounds of interaction; The ground business management system receives the text data sent by the on-orbit business management system and transmits it to the semantic understanding and data judgment module; The text data is semantically parsed based on the ground intelligent solution service terminal, and the semantic analysis results are fed back to the ground business management system. The ground business management system processes the material data according to the semantic analysis results and feeds back to the orbital business management system, which enables the on-orbit business management system to update the material status and display it through the front-end interactive interface.

7. The space station material information management method based on intelligent voice interaction according to claim 6 is characterized in that: Before converting voice commands into text data, it also includes: Divide the voice command data according to frequency from high to low; Match the background noise of the space station to each divided speech frequency band; Enhance the contrast between human voice and background noise; Perform human voice wave analysis, filtering and extraction to obtain voice data.

8. The space station material information management method based on intelligent voice interaction according to claim 6 is characterized in that: Convert voice commands into text data, including: Correct the text extraction results according to the stored special nouns of the space station scene; and / or, forcibly replace non-special nouns with special nouns for space station material management.

9. The space station material information management method based on intelligent voice interaction according to claim 6 is characterized in that: After the unrecognized result is fed back, the multi-level location information calculation steps of the materials are also included: The simplified position information is solved according to the mapping relationship between the stored simplified position information and the standard multi-level position information; the multi-level position information includes multiple combinations of aircraft, cabins, areas, quadrants, channels or multiple combinations of aircraft, cabins, areas, quadrants, cells; and / or, Extracting position keywords from the disordered position information, and solving the disordered position information by permuting and combining the position keywords; and / or, Match the corresponding unique standard multi-level location information according to the unique single-level location provided.

10. The space station material information management method based on intelligent voice interaction according to claim 6 is characterized in that: After the unrecognized result is reported, it also includes: The missing single-level location information is completed based on the multi-level location information of materials stored in the access control system.

Images