Intelligent elevator maintenance system

Abstract

The elevator maintenance intelligent system comprises a first wearable device and a second wearable device, and the first wearable device is provided with a multi-mode communication module, an AR interaction module and an audio collection and output module. The second wearable device is provided with a biological environment monitoring module, a positioning verification module, a risk monitoring module and a self-adaptive collaborative verification module; the multimode communication module is used for performing bidirectional data transmission on the AR interaction module, the audio acquisition and output module, the biological environment monitoring module, the positioning verification module, the risk monitoring module and the self-adaptive collaborative verification module with an external cloud platform or an Internet of Things system; according to the system, by integrating the multi-mode communication module, the AR interaction module, the biological environment monitoring module, the high-precision positioning module, the risk monitoring module and the like, comprehensive intelligent management of elevator maintenance operation is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent maintenance of special equipment, and particularly to an intelligent elevator maintenance system. Background Art

[0002] At present, there are still significant deficiencies in the safety monitoring and personnel management technologies for the operation environments of special equipment (such as hoistways, underground facilities, etc.), which are mainly reflected in the following aspects:

[0003] 1. Isolated health monitoring data: Existing equipment mainly independently monitors the vital signs of operators (such as heart rate, blood oxygen, etc.), and fails to conduct linkage analysis and risk early warning with environmental parameters (such as toxic gas concentration, temperature and humidity, oxygen content, etc.), resulting in a lag in the identification of potential safety hazards and the inability to achieve comprehensive risk assessment.

[0004] 2. Insufficient positioning accuracy: Traditional GPS modules have serious signal attenuation in enclosed or complex spaces (such as hoistways, underground areas), and the positioning error exceeds 5 meters, making it difficult to meet the requirements of high-precision position verification, and unable to effectively monitor whether operators enter compliant areas or dangerous restricted areas, increasing the risks of illegal operations and accidents.

[0005] 3. Lack of safety authentication mechanism for interaction devices: The operation of special equipment involves professional operations, and the use of relevant intelligent equipment by unauthorized personnel needs to be strictly restricted. However, existing equipment generally lacks multi-modal biometric authentication (such as iris, voiceprint, face recognition, etc.), and cannot ensure that operators hold certificates to work, presenting potential safety hazards of equipment abuse or misoperation.

[0006] The above technical defects limit the safety control capabilities in special operation scenarios, and there is an urgent need for a comprehensive solution integrating environmental perception, high-precision positioning, and intelligent authentication to improve operation safety and management efficiency. Summary of the Invention

[0007] The technical problem solved by the present invention is to provide an intelligent elevator maintenance system that reduces the fault diagnosis time and improves the interception rate of illegal operations.

[0008] The technical solution adopted by the present invention to solve its technical problems is: An intelligent elevator maintenance system, including a first wearable device and a second wearable device. The first wearable device is provided with a multi-mode communication module, an AR interaction module, and an audio collection and output module. The second wearable device is provided with a biological environment monitoring module, a positioning verification module, a risk monitoring module, and an adaptive collaborative verification module;

[0009] The multi-mode communication module is used for two-way data transmission between the AR interaction module, the audio collection and output module, the biological environment monitoring module, the positioning verification module, the risk monitoring module, and the adaptive collaborative verification module and an external cloud platform or Internet of Things system;

[0010] The AR interaction module is used for virtual-real linkage with remote experts;

[0011] The audio acquisition and output module is used to communicate with the intelligent system;

[0012] The biological environment monitoring module is used to detect the physical condition of the wearer and the surrounding environment conditions;

[0013] The positioning and verification module is used to verify that the operation location is within the set location range;

[0014] The risk monitoring module is used to detect major risks, major quality problems and serious non-standard behaviors during the repair process;

[0015] The adaptive collaborative verification module is used to trigger the on-site workflow through matching, comparison and analysis of on-site operation data and device background data, and at the same time conduct quality inspection on the results.

[0016] Furthermore: The multi-mode communication module includes a 5G module and a WIFI module, and also includes a switching module for network selection according to signal strength.

[0017] Furthermore: The AR interaction module includes AR glasses, a multi-modal biometric recognition unit, a permission control unit, a login record unit, a gesture interaction unit, and a remote collaboration unit;

[0018] The AR glasses are used to present the guidance of remote experts and the on-site operation pictures, enabling the wearer to intuitively carry out maintenance operations;

[0019] The multi-modal biometric recognition unit is used to identify the identity of the wearer, so that only authorized personnel can use the system;

[0020] The permission control unit is used to allocate corresponding operation permissions according to the identity of the wearer;

[0021] The login record unit is used to record the login time and location of the wearer;

[0022] The gesture interaction unit is used to identify the gesture commands of the wearer and realize natural interaction with the system;

[0023] The remote collaboration unit is used to communicate with remote experts for remote experts to guide on-site operations.

[0024] Furthermore: The audio acquisition and output module includes bone conduction headphones and a voiceprint feature extraction unit, an environmental noise reduction unit and a two-way voice interaction unit integrated in the bone conduction headphones;

[0025] The voiceprint feature extraction unit is used to collect user voice in real time, extract voiceprint features, and match with the pre-stored voiceprint library of authorized personnel to complete identity authentication;

[0026] The environmental noise reduction unit is used to eliminate high-frequency mechanical noise and low-frequency vibration noise in the industrial environment, and retain user voice and key safety warning sounds;

[0027] The two-way voice interaction unit is used to enable authorized users to send voice commands to the intelligent device through bone conduction headphones and conduct two-way interaction with the expert system in the external cloud platform or Internet of Things system.

[0028] Furthermore: The biological environment monitoring module includes a micro non-invasive sensor for detecting the wearer's blood pressure, blood oxygen, heart rate, and temperature indicators, and also includes an environmental sensor for monitoring the temperature, humidity, air pressure, VOC, and brightness indicators of the working environment.

[0029] Furthermore: The positioning verification module includes a dual-frequency positioning unit, a barometric pressure sensor, a data processing unit, and a warning unit;

[0030] The dual-frequency positioning unit is used to receive satellite signals in the GPS L5 frequency band and the Beidou B2b frequency band, and achieve horizontal positioning through joint solution;

[0031] The barometric pressure sensor is used to measure barometric pressure data and convert it into altitude information;

[0032] The data processing unit is used to combine the horizontally collected coordinates and altitude information into three-dimensional operation position data, and compare it with the standard position parameters in the pre-stored elevator guide BIM model;

[0033] The warning unit is used to trigger an abnormal warning when the deviation between the operation position and the standard position of the BIM model exceeds a preset threshold.

[0034] Furthermore: The risk monitoring module includes a three-dimensional compliance detection unit and an operation compliance detection unit;

[0035] The three-dimensional compliance detection unit is used to compare the actual image data of the elevator site collected with a preset three-dimensional specification model;

[0036] The operation compliance detection unit is used to identify non-compliant behaviors in the operation by real-time analyzing the operation video data of the operator, and provide operation guidance or correction prompts in real time.

[0037] Furthermore: The three-dimensional compliance detection unit adopts a point cloud registration technology based on the ICP algorithm to compare the actual image data of the elevator site collected with a preset three-dimensional specification model, detect whether the installation or operation status of the elevator meets the specifications, and generate correction prompts for non-compliant situations;

[0038] The operation compliance detection unit is built based on the SlowFast dual-channel network, and a quantized version is deployed on edge devices. It analyzes the operator's operation video data in real time, associates the operation coding tree and action specification scoring system constructed by the elevator maintenance SOP, identifies non-compliant behaviors in the operation, and provides real-time operation guidance or correction prompts.

[0039] Furthermore, the adaptive collaborative verification module includes a data matching and comparison unit, a multimodal cross-validation unit, an exception handling unit and a result quality inspection unit;

[0040] The data matching and comparison unit is used to match, compare and analyze the field operation data with the equipment background data to trigger the field workflow;

[0041] The multimodal cross-validation unit is used to perform multimodal cross-validation of dynamic weight allocation using a Bayesian network based on multi-source data provided by the biological environment monitoring module, the AR interaction module, the audio acquisition and output module, the positioning verification module, and the risk monitoring module;

[0042] The exception handling unit is used to trigger an alternative verification method when a single algorithm verification fails, including switching to other biometric verification or retrieving historical operation data for auxiliary verification;

[0043] The result quality inspection unit is used to perform quality inspection on the output results of the collaborative verification algorithm.

[0044] Furthermore, the dynamic weight allocation of the multimodal cross-validation module is based on a Bayesian network, and the weight adjustment formula is:

[0045]

[0046] Where Hi is the i-th verification algorithm and E is the environmental credibility.

[0047] The beneficial effects of the present invention are:

[0048] 1. This system realizes comprehensive intelligent management of elevator maintenance operations by integrating multi-mode communication, AR interaction, biological environment monitoring, high-precision positioning and risk monitoring modules.

[0049] 2. The multi-mode communication module ensures stable data transmission between the system and the external cloud platform or IoT system, providing a reliable basis for remote monitoring and guidance.

[0050] 3. The AR interaction module uses AR glasses and other devices to enable the wearer to intuitively receive guidance from remote experts and to achieve virtual and real linkage with the on-site work scenes, greatly improving work efficiency and accuracy.

[0051] 4. The biological environment monitoring module not only monitors the wearer's physical condition in real time, but also comprehensively monitors environmental parameters, effectively preventing potential safety hazards caused by environmental factors. The positioning verification module uses technologies such as dual-frequency positioning and air pressure sensing to achieve high-precision three-dimensional positioning, ensuring that the operators are always working within the compliant area and reducing the risk of non-compliant operations.

[0052] 5. The risk monitoring module conducts real-time analysis on the actual image data of the elevator site and the operation video data of the operators through three-dimensional compliance detection and operation compliance detection, promptly discovers and corrects non-compliant behaviors, and effectively guarantees the operation safety and quality.

[0053] 6. The adaptive collaborative verification module further improves the security and reliability of the system through multi-modal cross-verification and exception handling mechanisms. Even when a single algorithm verification fails, it can quickly switch to other verification methods to ensure that the system is always in the best working state. Description of the Drawings

[0054] Figure 1 It is a schematic structural diagram of the elevator maintenance intelligent system according to the embodiment of the present application.

[0055] Figure 2 It is a flowchart of the voice recognition algorithm in the elevator maintenance intelligent system according to the embodiment of the present application.

[0056] Figure 3 It is an example diagram of one embodiment of the multi-modal cross-verification module in the elevator maintenance intelligent system according to the embodiment of the present application.

[0057] Figure 4 It is a schematic flowchart of the elevator maintenance intelligent system according to the embodiment of the present application.

[0058] In the figure, the labels are: the first wearable device 1, the multi-mode communication module 11, the AR interaction module 12, the audio acquisition and output module 13, the second wearable device 2, the biological environment monitoring module 21, the positioning verification module 22, the risk monitoring module 23, and the adaptive collaborative verification module 24. Detailed Embodiments

[0059] In order to make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of the specific embodiments of the present invention will be given with reference to the accompanying drawings.

[0060] Such as Figure 1As shown in the figure, an embodiment of the present application discloses an intelligent elevator maintenance system, including a first wearable device 1 and a second wearable device 2. A multi-mode communication module 11, an AR interaction module 12, and an audio acquisition and output module 13 are provided on the first wearable device 1. A biological environment monitoring module 21, a positioning verification module 22, a risk monitoring module 23, and an adaptive collaborative verification module 24 are provided on the second wearable device;

[0061] The multi-mode communication module 11 is used to perform two-way data transmission between the AR interaction module 12, the audio acquisition and output module 13, the biological environment monitoring module 21, the positioning verification module 22, the risk monitoring module 23, and the adaptive collaborative verification module 24 and an external cloud platform or an Internet of Things system;

[0062] The AR interaction module 12 is used to perform virtual-real linkage with remote experts;

[0063] The audio acquisition and output module 13 is used to communicate with the intelligent system;

[0064] The biological environment monitoring module 21 is used to detect the physical condition of the wearer and the surrounding environment condition;

[0065] The positioning verification module 22 is used to verify that the operation position is within the set position range;

[0066] The risk monitoring module 23 is used to detect major risks, major quality problems, and serious non-standard behaviors during the repair process;

[0067] The adaptive collaborative verification module 24 is used to trigger the on-site workflow by matching, comparing, and analyzing on-site operation data with device background data, and at the same time perform quality inspection on the results.

[0068] It should be explained that the above first wearable device 1 can be a head-mounted device, such as a safety helmet, a helmet, etc. The second wearable device 2 can be a wearable device, such as a safety vest or a safety suit, etc.

[0069] Specifically, such as Figure 4As shown in the figure, when the operator is performing on-site operations, the positioning verification module 22 confirms whether the operator has reached the designated elevator position. If the position deviation exceeds the set threshold, the system automatically triggers an alarm and locks the function of the AR interaction module 12. The biological environment monitoring module 21 continuously collects physiological data such as the operator's heart rate and blood oxygen, as well as environmental temperature, humidity, and harmful gas concentration. When the data is abnormal, a voice prompt is issued through the audio collection output module 13 and synchronized to the cloud management platform. During the operation, the operator can retrieve the three-dimensional elevator structure diagram through the AR interaction module 12, and remote experts can guide complex operations through the virtual-real overlay annotation function. The risk monitoring module 23 detects major risks, major quality problems, and serious non-standard behaviors during the maintenance process. The adaptive collaborative verification module 24 automatically triggers corresponding work processes based on the matching results of on-site operation data and background data, and conducts quality inspections on the operation results to ensure that each step of the operation complies with the specifications. In addition, the system also has a high degree of flexibility and can dynamically adjust the working parameters of each module according to the actual situation of the operation environment to meet different maintenance needs.

[0070] This system realizes the comprehensive intelligent management of elevator maintenance operations by integrating modules such as multi-mode communication, AR interaction, biological environment monitoring, high-precision positioning, and risk monitoring.

[0071] In this embodiment, the multi-mode communication module 11 includes a 5G module and a WIFI module, and also includes a switching module for network selection according to signal strength.

[0072] Specifically, the 5G module and the WIFI module can automatically switch according to the current environmental signal strength to ensure the stability and efficiency of data transmission. In areas with good signals, the system preferentially selects the WIFI module for data transmission to reduce operating costs; while in areas with weak signals, it automatically switches to the 5G module to ensure the real-time and reliability of data transmission. This intelligent switching mechanism enables maintenance personnel to obtain stable data communication support in different environments.

[0073] In this embodiment, the AR interaction module 12 includes AR glasses, a multi-modal biometric recognition unit, a permission control unit, a login record unit, a gesture interaction unit, and a remote collaboration unit;

[0074] The AR glasses are used to present the guidance of remote experts and the on-site operation screen, enabling the wearer to intuitively perform maintenance operations;

[0075] The multi-modal biometric recognition unit is used to identify the identity of the wearer, so that only authorized personnel can use the system;

[0076] The permission control unit is used to allocate corresponding operation permissions according to the identity of the wearer;

[0077] The login record unit is used to record the login time and location of the wearer;

[0078] The gesture interaction unit is used to recognize the gesture commands of the wearer and achieve natural interaction with the system;

[0079] The remote collaboration unit is used to communicate with remote experts for the remote experts to guide on-site operations.

[0080] Specifically, the AR glasses are monocular or binocular AR glasses, integrated with non-contact multi-modal biometric recognition. Only authorized personnel with the required certificates can activate the equipment and receive maintenance tasks and guidance from remote experts through the AR glasses. The permission control unit strictly restricts the operation permissions of wearers with different identities to ensure the safety and standardization of maintenance operations. The login record unit details the time and location of each login, providing strong support for subsequent traceability and management. The gesture interaction unit enables the wearer to interact with the system through simple gesture commands, greatly improving the operation convenience. The remote collaboration unit supports real-time communication between the wearer and remote experts, enabling experts to remotely guide on-site operations and solving the pain point of difficult handling of complex on-site problems.

[0081] In this structure, the AR interaction module 12 enables the wearer to intuitively receive guidance from remote experts through devices such as AR glasses and perform virtual-real linkage with the on-site operation screen, greatly improving the operation efficiency and accuracy.

[0082] In this embodiment, the audio acquisition and output module 13 includes a bone conduction headset and a voiceprint feature extraction unit, an environmental noise reduction unit, and a two-way voice interaction unit integrated in the bone conduction headset;

[0083] The voiceprint feature extraction unit is used to collect user voice in real time and extract voiceprint features, and match them with the pre-stored voiceprint library of authorized personnel to complete identity authentication;

[0084] The environmental noise reduction unit is used to eliminate high-frequency mechanical noise and low-frequency vibration noise in the industrial environment and retain user voice and key safety warning sounds;

[0085] The two-way voice interaction unit is used for authorized users to send voice commands to intelligent devices through the bone conduction headset and perform two-way interaction with the expert system in the external cloud platform or Internet of Things system.

[0086] Specifically, the bone conduction earphone includes a voiceprint feature extraction array and an environmental noise reduction algorithm. Authorized personnel can communicate with intelligent equipment, that is, connect to the expert system in the background through the multimode communication module 11. When employees have problems on site, they can directly ask questions and receive feedback and guidance through the bone conduction earphone, forming an interaction to remind on-site workers of safety and quality risks, etc. Moreover, the sound pickup device of the bone conduction earphone can filter out noisy environmental sounds without affecting on-site workers' ability to hear on-site sounds, ensuring on-site safety.

[0087] Among them, the speech recognition technology uses an algorithm based on Mel-scale Frequency Cepstral Coefficients (MFCC for short). This algorithm uses a 26-dimensional Mel filter bank and combines first-order and second-order difference techniques. In addition, it also uses an x-vector mapping based on the ResNet-34 architecture for identity verification. The specific algorithm process is as follows Figure 2 shown.

[0088] Among them, the environmental noise reduction algorithm combines a Gammatone filter bank with a deep noise reduction network. The deep noise reduction network uses a dual-path bidirectional long short-term memory network (BiLSTM, 128 units per path) and a time-domain mask (time threshold of 5 milliseconds), and is expected to increase the signal-to-noise ratio to ≥23 dB.

[0089] In this embodiment, the biological environment monitoring module 21 includes a micro non-invasive sensor for detecting the blood pressure, blood oxygen, heart rate, and temperature indicators of the wearer, and also includes an environmental sensor for monitoring the temperature, humidity, air pressure, VOC, and brightness indicators of the working environment.

[0090] Specifically, the sensors for detecting blood pressure, blood oxygen, heart rate, and temperature can be patch-type sensors integrated in the wearable device, which can monitor the physiological parameters of the wearer in real time. Once an abnormality is detected, an alarm will be immediately issued through the audio acquisition output module 13 and synchronized to the cloud management platform through the multimode communication module 11 for timely rescue measures. The

[0091] The environmental sensor can monitor key indicators such as temperature and humidity, air pressure, harmful gas concentration (such as VOC), and brightness in the working environment in real time. These data are crucial for evaluating the safety and comfort of the working environment. When the environmental parameters exceed the preset safety range, the system will immediately trigger a warning and issue instructions for emergency evacuation or taking corresponding protective measures to the operators through the AR interaction module 12 or the audio acquisition output module 13, thus effectively avoiding potential safety risks.

[0092] In this embodiment, the positioning and verification module 22 includes a dual-frequency positioning unit, a barometric pressure sensor, a data processing unit, and a warning unit;

[0093] The dual-frequency positioning unit is used to receive satellite signals in the GPS L5 frequency band and the Beidou B2b frequency band, and achieve horizontal positioning through joint solution;

[0094] The barometric pressure sensor is used to measure barometric pressure data and convert it into altitude information;

[0095] The data processing unit is used to combine the horizontally acquired coordinates and altitude information in real time into three-dimensional operation position data, and compare it with the standard position parameters in the pre-stored elevator guide rail BIM model;

[0096] The warning unit is used to trigger an abnormal alarm when the deviation between the operation position and the standard position of the BIM model exceeds a preset threshold.

[0097] Specifically, by receiving satellite signals from the GPS L5 frequency band and the Beidou B2b frequency band, the dual-frequency positioning unit can accurately determine the horizontal position of the operator. Combining the altitude information provided by the barometric pressure sensor, the data processing unit can calculate the three-dimensional operation position data of the operator in real time. This data is then compared with the standard position parameters in the elevator guide rail BIM model pre-stored in the system to ensure that the operator is always in a compliant operation area. Once the deviation between the operation position and the standard position in the BIM model exceeds the preset safety threshold, the warning unit will immediately trigger an abnormal alarm, issue a warning to the operator through the AR interaction module 12 or the audio acquisition output module 13, and even lock the function of the AR interaction module 12 if necessary to prevent non-compliant operations.

[0098] This high-precision three-dimensional positioning technology not only improves the safety of operations, but also provides valuable data support for subsequent operation analysis and optimization.

[0099] In this embodiment, the risk monitoring module 23 includes a three-dimensional compliance detection unit and an operation compliance detection unit;

[0100] The three-dimensional compliance detection unit is used to compare the actual image data of the elevator site collected with a preset three-dimensional specification model;

[0101] The operation compliance detection unit is used to identify non-compliant behaviors in operations by analyzing the operation video data of the operator in real time, and provide operation guidelines or correction prompts in real time.

[0102] Specifically, the three-dimensional compliance detection unit uses point cloud registration technology based on the ICP algorithm to compare the actual image data of the elevator site collected with the preset three-dimensional specification model, detect whether the installation or operation status of the elevator meets the specifications, and generate a correction prompt for non-compliance situations;

[0103] The operation compliance detection unit is based on the SlowFast dual-path network, deployed in a quantized version on edge devices. By analyzing the operation video data of the operator in real time and associating with the operation coding tree and action specification scoring system constructed by the elevator maintenance SOP, it identifies non-compliant behaviors during the operation and provides operation guidance or correction prompts in real time, thus effectively ensuring the safety and quality of the operation.

[0104] In this embodiment, the adaptive collaborative verification module 24 includes a data matching and comparison unit, a multi-modal cross-verification unit, an exception handling unit, and a result quality inspection unit;

[0105] The data matching and comparison unit is used to match, compare, and analyze the on-site operation data with the device background data to trigger the on-site workflow;

[0106] The multi-modal cross-verification unit is used to perform multi-modal cross-verification with dynamic weight allocation based on the Bayesian network for the multi-source data provided by the biological environment monitoring module 21, the AR interaction module 12, the audio acquisition and output module 13, the positioning verification module 22, and the risk monitoring module 23;

[0107] The exception handling unit is used to trigger an alternative verification method when a single algorithm verification fails, including switching to other biometric verifications or retrieving historical operation data for auxiliary verification;

[0108] The result quality inspection unit is used to perform quality inspection on the output results of the collaborative verification algorithm.

[0109] Specifically, the dynamic weight allocation of the multi-modal cross-verification module is based on the Bayesian network, and its weight adjustment formula is:

[0110]

[0111] Where Hi is the i-th verification algorithm and E is the environmental credibility.

[0112] The following is one specific embodiment. Specifically, as shown in Figure 3, for example: If a person's pharyngitis may cause the voiceprint verification to fail and the AR glasses cannot be unlocked, the adaptive collaborative algorithm will trigger other verifications at this time, such as face biometric verification, or retrieve past operation data for the person to perform problem verification, avoiding misjudgment caused by a single algorithm. The specific steps are as follows:

[0113] S1: Voiceprint verification fails. Detect whether the ambient noise is greater than 65 db? If yes, go to step S2; otherwise, go to step S3;

[0114] S2: Start iris recognition. If the passing rate is greater than 92%, the AR glasses are temporarily authorized; otherwise, the AR glasses are locked;

[0115] S3: Retrieve 3 historical questions. If the correct rate is greater than 80%, the AR glasses are temporarily authorized; otherwise, the AR glasses are locked.

[0116] The adaptive collaborative verification module 24 in this system further improves the security and reliability of the system through a multi-modal cross-verification and exception handling mechanism. Even when a single algorithm verification fails, it can quickly switch to other verification methods to ensure that the system is always in the best working state.

[0117] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. An intelligent elevator maintenance system, characterized in that: It includes a first wearable device (1) and a second wearable device (2). A multimode communication module (11), an AR interaction module (12), and an audio collection and output module (13) are provided on the first wearable device (1). A biological environment monitoring module (21), a positioning verification module (22), a risk monitoring module (23), and an adaptive collaborative verification module (24) are provided on the second wearable device; The multimode communication module (11) is used for two-way data transmission between the AR interaction module (12), the audio collection and output module (13), the biological environment monitoring module (21), the positioning verification module (22), the risk monitoring module (23), the adaptive collaborative verification module (24) and an external cloud platform or an Internet of Things system; The AR interaction module (12) is used for virtual-real linkage with remote experts; The audio collection and output module (13) is used for communicating with an intelligent system; The biological environment monitoring module (21) is used for detecting the physical condition of the wearer and the surrounding environment condition; The positioning verification module (22) is used for verifying that the operation location is within the set location range; The risk monitoring module (23) is used for implementing the detection of major risks, major quality problems and serious non-standard behaviors during the repair process; The adaptive collaborative verification module (24) is used for triggering the on-site workflow through the matching, comparison and analysis of on-site operation data and device background data, and at the same time performing quality inspection on the results.

2. The elevator maintenance intelligent system according to claim 1, characterized in that: The multimode communication module (11) includes a 5G module and a WIFI module, and also includes a switching module for network selection according to the signal strength.

3. The elevator maintenance intelligent system according to claim 1, characterized in that: The AR interaction module (12) includes AR glasses, a multimodal biometric recognition unit, a permission control unit, a login record unit, a gesture interaction unit, and a remote collaboration unit; The AR glasses are used for presenting the guidance of remote experts and the on-site operation picture, so that the wearer can intuitively perform maintenance operations; The multimodal biometric recognition unit is used for identifying the identity of the wearer, so that only authorized personnel can use the system; The permission control unit is used for allocating corresponding operation permissions according to the identity of the wearer; The login record unit is used for recording the login time and location of the wearer; The gesture interaction unit is used for identifying the gesture commands of the wearer to achieve natural interaction with the system; The remote collaboration unit is used for communicating with remote experts for remote experts to guide on-site operations.

4. The elevator maintenance intelligent system according to claim 1, characterized in that: The audio collection and output module (13) includes bone conduction headphones and a voiceprint feature extraction unit, an environmental noise reduction unit, and a two-way voice interaction unit integrated in the bone conduction headphones; The voiceprint feature extraction unit is used for collecting user voices in real time and extracting voiceprint features, and matching with a pre-stored voiceprint library of authorized personnel to complete identity authentication; The environmental noise reduction unit is used for removing high-frequency mechanical noises and low-frequency vibration noises in the industrial environment, and retaining user voices and key safety warning sounds; The two-way voice interaction unit is used for authorized users to send voice commands to intelligent devices through bone conduction headphones and perform two-way interaction with an expert system in an external cloud platform or an Internet of Things system.

5. The elevator maintenance intelligent system according to claim 1, characterized in that: The biological environment monitoring module (21) includes a micro non-invasive sensor for detecting the wearer's blood pressure, blood oxygen, heart rate, and temperature indicators, and also includes an environmental sensor for monitoring the temperature, humidity, air pressure, VOC, and brightness indicators of the working environment.

6. The elevator maintenance intelligent system according to claim 1, characterized in that: The positioning and verification module (22) includes a dual-frequency positioning unit, a barometric pressure sensor, a data processing unit, and a warning unit; The dual-frequency positioning unit is used to receive satellite signals in the GPS L5 frequency band and the Beidou B2b frequency band, and realizes horizontal positioning through joint solution; The barometric pressure sensor is used to measure barometric pressure data and convert it into altitude information; The data processing unit is used to combine the horizontally acquired coordinates and altitude information into three-dimensional operation position data, and compare it with the standard position parameters in the pre-stored BIM model of the elevator guide rail; The warning unit is used to trigger an abnormal alarm when the deviation between the operation position and the standard position of the BIM model exceeds a preset threshold.

7. The elevator maintenance intelligent system according to claim 1, characterized in that: The risk monitoring module (23) includes a three-dimensional compliance detection unit and an operation compliance detection unit; The three-dimensional compliance detection unit is used to compare the actually acquired on-site image data of the elevator with a preset three-dimensional specification model; The operation compliance detection unit is used to identify non-compliant behaviors during the operation by real-time analyzing the operation video data of the operator, and provide operation guidance or correction prompts in real time.

8. The elevator maintenance intelligent system according to claim 7, wherein: The three-dimensional compliance detection unit uses the point cloud registration technology based on the ICP algorithm to compare the actually acquired on-site image data of the elevator with a preset three-dimensional specification model, detect whether the installation or operation status of the elevator meets the specifications, and generate correction prompts for non-compliant situations; The operation compliance detection unit is constructed based on the SlowFast dual-path network, deployed in a quantized version on edge devices. By real-time analyzing the operation video data of the operator, associating with the operation coding tree and action specification scoring system constructed by the elevator maintenance SOP, it identifies non-compliant behaviors during the operation and provides operation guidance or correction prompts in real time.

9. The elevator maintenance intelligent system according to claim 1, wherein: The adaptive collaborative verification module (24) includes a data matching and comparison unit, a multi-modal cross-verification unit, an exception handling unit, and a result quality inspection unit; The data matching and comparison unit is used to match, compare, and analyze the on-site operation data with the device background data to trigger the on-site workflow; The multi-modal cross-verification unit is used to perform multi-modal cross-verification with dynamic weight allocation based on the Bayesian network using the multi-source data provided by the biological environment monitoring module (21), the AR interaction module (12), the audio acquisition and output module (13), the positioning and verification module (22), and the risk monitoring module (23); The exception handling unit is used to trigger an alternative verification method when a single algorithm verification fails, including switching to other biometric verifications or retrieving historical operation data for auxiliary verification; The result quality inspection unit is used to perform quality inspection on the output results of the collaborative verification algorithm.

10. The elevator maintenance intelligent system according to claim 9, characterized in that: The dynamic weight allocation of the multi-modal cross-verification module is based on the Bayesian network, and its weight adjustment formula is: Where Hi is the i-th verification algorithm and E is the environmental credibility.

Images