An intelligent hotel internet of things management method and system based on artificial intelligence

Abstract

The application provides an intelligent hotel Internet of Things management method and system based on artificial intelligence, and relates to the technical field of intelligent management. The method comprises the following steps: constructing a three-dimensional model of a hotel building, performing regional division and Internet of Things device deployment, and collecting N hotel regional operation data streams; deploying a multi-modal sensing network in the guest room department of a target intelligent hotel; building a distributed edge node set; correlating and mapping the N hotel regional operation data streams and multi-user multi-modal data streams to the distributed edge node set for abnormal feature recognition; constructing a hotel operation and maintenance strategy library, determining target hotel operation and maintenance strategy parameters, and performing early warning operation and maintenance management and control on the target intelligent hotel. The application solves the technical problem of low hotel operation and maintenance management efficiency caused by the difficulty in discovering device faults and energy waste in a timely manner due to reliance on manual inspection and regular inspection, and improves the operation and maintenance management efficiency of the intelligent hotel by identifying device abnormalities and user behavior abnormalities.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent management, and particularly relates to a smart hotel Internet of Things management method and system based on artificial intelligence. BACKGROUND

[0002] In the current hotel industry, equipment operation and maintenance management is an indispensable part of the operation process. Since the hotel involves a variety of facilities and equipment, and most of these devices rely on manual inspection and regular inspection to ensure their normal operation. Due to the reliance on manual inspection, it is difficult to discover equipment abnormalities and hidden dangers in time, resulting in some potential problems being ignored, and even causing serious equipment failure, affecting hotel operation. In addition, the current hotel operation and maintenance management fails to fully identify and analyze the behavior characteristics of users, and cannot effectively identify abnormal behavior of users during their stay, which may lead to waste of energy, such as air conditioners being turned on for a long time while users are not in the room, lighting systems remaining on when no one is present, etc., not only increasing the operating cost of the hotel, but also affecting the user's stay experience, thereby affecting the efficiency of hotel operation and maintenance management.

[0003] In summary, in the prior art, due to the reliance on manual inspection and regular inspection, it is difficult to discover equipment failure and energy waste in time, resulting in low efficiency of hotel operation and maintenance management. SUMMARY

[0004] The purpose of the present application is to provide a smart hotel Internet of Things management method and system based on artificial intelligence, to solve the technical problem of low efficiency of hotel operation and maintenance management due to the reliance on manual inspection and regular inspection in the prior art.

[0005] In view of the above problems, the present application provides a smart hotel Internet of Things management method and system based on artificial intelligence.

[0006] In a first aspect, the present application provides an intelligent hotel Internet of Things management method based on artificial intelligence. The method is implemented by an intelligent hotel Internet of Things management system based on artificial intelligence. The method includes: constructing a hotel building three-dimensional model, performing regional division and Internet of Things device deployment on the hotel building three-dimensional model to obtain N hotel Internet of Things regions, and collecting N hotel region operation data streams of the N hotel Internet of Things regions; deploying a multi-modal sensing network in a guest room department of a target intelligent hotel, collecting multi-user multi-modal data streams in real time through the multi-modal sensing network; performing load demand prediction and edge node configuration in the N hotel Internet of Things regions, and building a distributed edge node set; associating and mapping the N hotel region operation data streams and the multi-user multi-modal data streams to the distributed edge node set for abnormal feature identification to obtain a hotel abnormal feature set; constructing a hotel operation and maintenance strategy library, performing operation and maintenance strategy analysis on the hotel abnormal feature set using the hotel operation and maintenance strategy library, determining target hotel operation and maintenance strategy parameters, and performing early warning operation and maintenance control on the target intelligent hotel based on the target hotel operation and maintenance strategy parameters.

[0007] Optionally, a regional division grid is preset according to hotel management accuracy, a hotel building three-dimensional model is divided into floors and regionally equalized according to the regional division grid, a hotel floor equalization region set is obtained, each region information in the hotel floor equalization region set is functionally identified to obtain hotel floor region application functions, the hotel floor equalization region set is merged and connected according to the hotel floor region application functions, N hotel key regions are determined, monitoring demand analysis and deployment coverage analysis are performed on the N hotel key regions, N hotel region monitoring demand parameters and N hotel region device deployment parameters are obtained, Internet of Things devices are selected and deployed based on the N hotel region monitoring demand parameters and the N hotel region device deployment parameters, and the N hotel Internet of Things regions are obtained.

[0008] Optionally, a multi-level function tag system is constructed, the hotel floor region application functions are multi-level identified according to the multi-level function tag system, hotel region multi-level function classification tags are obtained, the hotel floor equalization region set is tag-mapped and valued according to the hotel region multi-level function classification tags, an equalization region hotel function tag set is obtained, a function tag overlap threshold is preset, the equalization region hotel function tag set is overlap-determined using the function tag overlap threshold, a consistent equalization region tag set is obtained, and the hotel floor equalization region set is merged and connected based on the consistent equalization region tag set, and the N hotel key regions are determined.

[0009] Optionally, historical data mining is performed based on the N hotel IoT areas to obtain N hotel area historical operation data sets; calculation task extraction is performed on the N hotel area historical operation data sets to obtain N hotel area calculation task sets and N hotel area calculation amount sets; demand quantization analysis is performed on the N hotel area calculation task sets to determine N hotel area task calculation demands, including calculation accuracy, calculation privacy level, and calculation rate; based on the N hotel area calculation amount sets and the N hotel area task calculation demands, load demand prediction and edge node configuration are performed on the N hotel IoT areas to obtain a distributed edge node set.

[0010] Optionally, a distributed edge topology network is built, and based on the distributed edge topology network, distributed network calculation total resources are determined; based on the N hotel area calculation amount sets and the N hotel area task calculation demands, cluster analysis is performed on the N hotel IoT areas to obtain a distributed edge calculation area set; the distributed network calculation total resources are used as constraint parameters to perform load demand prediction and calculation resource allocation on the distributed edge calculation area set to determine edge area calculation resource allocation information; based on the edge area calculation resource allocation information, edge node configuration is performed on the distributed edge calculation area set to obtain the distributed edge node set.

[0011] Optionally, based on hotel operation management standards, a hotel abnormal operation feature library is defined, including hotel device-level abnormalities and user behavior abnormalities; based on the hotel abnormal operation feature library, abnormal data association and abnormality identification training are performed to build a hotel abnormal feature identifier, and the hotel abnormal feature identifier is stored in the distributed edge node set; the N hotel area operation data streams and the multi-user multi-modal data streams are associated and mapped to the distributed edge node set to obtain an edge node associated data stream set; the hotel abnormal feature identifier is called by the distributed edge node set to perform abnormal feature identification on the edge node associated data stream set to obtain the hotel abnormal feature set.

[0012] Optionally, based on the hotel abnormal operation feature library, abnormal data association is performed to obtain a hotel abnormal operation feature data set; the hotel abnormal operation feature data set is evaluated and labeled to obtain a hotel abnormal operation type sample set and a hotel abnormal operation level sample set; based on the hotel abnormal operation type sample set and the hotel abnormal operation level sample set, identification training is combined to build the hotel abnormal feature identifier.

[0013] Optionally, the hotel operation strategy library is used to perform strategy matching analysis on the hotel abnormal feature set, to obtain a plurality of hotel operation strategy parameters; simulation and effect evaluation are performed on the plurality of hotel operation strategy parameters, to obtain a plurality of hotel operation parameter effects; the plurality of hotel operation strategy parameters are compared and optimized according to the plurality of hotel operation parameter effects, to determine the target hotel operation strategy parameter.

[0014] Optionally, a user behavior portrait is constructed according to user historical check-in data and user biological features; the target hotel operation strategy parameter is corrected based on the user behavior portrait, to determine a hotel operation personalized strategy parameter.

[0015] In a second aspect, the present application further provides a smart hotel Internet of Things management system based on artificial intelligence, which is used to execute the smart hotel Internet of Things management method based on artificial intelligence as described in the first aspect. The smart hotel Internet of Things management system based on artificial intelligence comprises: a region division module, which is used to construct a hotel building three-dimensional model, divide regions of the hotel building three-dimensional model, and deploy Internet of Things devices, to obtain N hotel Internet of Things regions, and collect N hotel region running data streams of the N hotel Internet of Things regions; a perception network layout module, which is used to layout a multi-modal perception network in a guest room department of a target smart hotel, and collect multi-user multi-modal data streams in real time through the multi-modal perception network; an edge node building module, which is used to perform load demand prediction and edge node configuration in the N hotel Internet of Things regions, and build a distributed edge node set; an abnormal feature identification module, which is used to associate and map the N hotel region running data streams and the multi-user multi-modal data streams to the distributed edge node set for abnormal feature identification, to obtain a hotel abnormal feature set; and an operation management module, which is used to construct a hotel operation strategy library, use the hotel operation strategy library to perform operation strategy analysis on the hotel abnormal feature set, determine a target hotel operation strategy parameter, and perform early warning operation management and control on the target smart hotel based on the target hotel operation strategy parameter.

[0016] The one or more technical solutions provided in the present application have at least the following beneficial effects:

[0017] By constructing a three-dimensional model of a hotel building, regional division and Internet of Things device deployment are performed on the three-dimensional model of the hotel building to obtain N hotel Internet of Things regions, and N hotel region operation data streams of the N hotel Internet of Things regions are collected; a multi-modal sensing network is laid out in a guest room department of a target smart hotel, and multi-user multi-modal data streams are collected in real time through the multi-modal sensing network; load demand prediction and edge node configuration are performed in the N hotel Internet of Things regions, and a distributed edge node set is built; the N hotel region operation data streams and the multi-user multi-modal data streams are associated and mapped to the distributed edge node set for abnormal feature identification to obtain a hotel abnormal feature set; a hotel operation and maintenance strategy library is constructed, the hotel operation and maintenance strategy library is used for operation and maintenance strategy analysis on the hotel abnormal feature set, target hotel operation and maintenance strategy parameters are determined, and the target smart hotel is prewarned, operated, managed and controlled based on the target hotel operation and maintenance strategy parameters. That is, by constructing a three-dimensional model of a hotel, performing regional division and Internet of Things device deployment, determining N key regions, laying out a multi-modal sensing network in a guest room, and building a distributed edge node set, the operation data streams of the N key regions of the hotel and the data streams collected by the multi-modal sensing network are associated and mapped to the distributed edge node set for abnormal feature identification, so as to determine a target hotel operation and maintenance strategy, and improve the efficiency of hotel operation and maintenance management.

[0018] The above description is only a summary of the technical solutions of the present application. In order to enable the technical means of the present application to be more clearly understood, and to be implemented according to the content of the description, and in order to enable the above and other purposes, characteristics and advantages of the present application to be more apparent and easy to understand, the following specific embodiments of the present application are described. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 A flowchart of a smart hotel Internet of Things management method based on artificial intelligence according to the present application;

[0020] Figure 2 A structural diagram of a smart hotel Internet of Things management system based on artificial intelligence according to the present application.

[0021] BRIEF DESCRIPTION OF DRAWINGS DETAILED DESCRIPTION

[0022] The application provides an intelligent hotel Internet of Things management method and system based on artificial intelligence, which solves the technical problem of low hotel operation and management efficiency caused by the difficulty in finding equipment failure and energy waste in time due to the dependence on manual inspection and timing inspection in the prior art. By constructing a three-dimensional model of the hotel, region division and Internet of Things device deployment are performed, N key areas are determined, a multi-modal sensing network is arranged in the guest room, a distributed edge node set is built, the operation data flow of the N key areas of the hotel and the data flow collected by the multi-modal sensing network are associated and mapped to the distributed edge node for abnormal feature recognition, so that the target hotel operation strategy is determined, and the efficiency of hotel operation and management is improved.

[0023] The technical solutions in the application will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. It should be understood that the application is not limited by the example embodiments described herein. Based on the embodiments of the application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the application. In addition, it should be noted that, for the convenience of description, only part of the application is shown in the drawings, not all.

[0024] Embodiment one, please refer to the attached Figure 1 The application provides an intelligent hotel Internet of Things management method based on artificial intelligence, wherein the intelligent hotel Internet of Things management method based on artificial intelligence is executed by an intelligent hotel Internet of Things management system based on artificial intelligence, and the intelligent hotel Internet of Things management method based on artificial intelligence specifically includes the following steps:

[0025] S100: Construct a three-dimensional model of a hotel building, divide the three-dimensional model of the hotel building into regions, and deploy Internet of Things devices to obtain N hotel Internet of Things regions, and collect N hotel region operation data flows of the N hotel Internet of Things regions.

[0026] Further, the S100 of the application includes:

[0027] According to the hotel management precision, a grid is preset for region division, the hotel building three-dimensional model is floor segmented and regionally divided according to the grid for region division, a hotel floor regionally divided region set is obtained; each region information in the hotel floor regionally divided region set is respectively functionally identified, and a hotel floor region application function is obtained; the hotel floor regionally divided region set is merged and connected according to the hotel floor region application function, and N hotel key regions are determined; monitoring demand analysis and deployment coverage analysis are performed on the N hotel key regions, and N hotel region monitoring demand parameters and N hotel region device deployment parameters are obtained; based on the N hotel region monitoring demand parameters and the N hotel region device deployment parameters, Internet of Things devices are selected and deployed, and the N hotel Internet of Things regions are obtained.

[0028] A multi-level function tag system is constructed, the hotel floor region application function is multi-level identified according to the multi-level function tag system, and a hotel region multi-level function classification tag is obtained; the hotel floor regionally divided region set is labeled and valued according to the hotel region multi-level function classification tag, and a regionally divided region hotel function tag set is obtained; a function tag overlap threshold is preset, the function tag overlap threshold is used to determine the overlap of the regionally divided region hotel function tag set, and a consistent regionally divided region tag set is obtained; based on the consistent regionally divided region tag set, the hotel floor regionally divided region set is merged and connected, and the N hotel key regions are determined.

[0029] Specifically, using existing building information modeling technology, according to the blueprint and design specifications of the hotel, a three-dimensional model of the hotel is constructed, including the external structure, internal floors and facility layout of the hotel, etc., which comprehensively displays the building information of the hotel. Through the hotel building three-dimensional model, the layout, room distribution, passageway, equipment position, etc. of each floor can be directly observed. According to the hotel management precision, the size and shape of the grid are set, i.e. the grid for region division, which divides the hotel space into multiple small regions. The hotel management precision is usually determined according to the specifications of the hotel and the specific management needs.

[0030] Through the preset grid for region division, the hotel building three-dimensional model is floor segmented and regionally divided, i.e. each floor of the hotel building is spatially segmented and divided into several regions according to certain standards, obtaining a hotel floor regionally divided region set, including multiple regions of multiple floors. Floor segmentation refers to dividing the three-dimensional model of the hotel according to floors, and each floor becomes an independent management unit. Regional division refers to further dividing the region within each floor.

[0031] According to the function of each area (such as the guest room area, the restaurant area, the conference room, the hotel staff office area, etc.), the function identification is performed on the area information in each sub-area set of the hotel floor, and the use nature of each area of each floor is determined. The hotel floor area application function is identified according to the use function (such as the guest room, the restaurant, the conference room, etc.) of each floor or area, and is used to guide the equipment and service configuration in the corresponding area.

[0032] According to the function characteristics of the hotel floor area, a multi-level function label system is constructed, that is, different levels and dimensions of labels are used for classification and identification according to the hotel floor area application function, and the function of each area is refined and accurately identified. The label system needs to be flexibly designed according to the different needs of the hotel. For example, the guest room area can be divided into a double bed room area, a standard room area, a suite area, etc., and the specific use nature of the room is refined.

[0033] According to the multi-level function label system, the hotel floor area application function is identified at multiple levels, and the function attributes of each area are more accurately determined. Each area will be assigned a complete function label according to its specific application function. The hotel area multi-level function classification label refers to assigning a corresponding function classification label to each hotel floor area according to the multi-level function label system, and gradually dividing from macro to micro to identify the specific use and attributes of the area. According to the hotel area multi-level function classification label, the hotel floor sub-area set after floor segmentation and area division is labeled and valued, and the hotel floor area application function label is mapped to the corresponding area unit, ensuring that each area has accurate function identification, that is, each area is more detailed in function identification. The sub-area hotel function label set includes the detailed function label of each area, and each area is accurately monitored and managed.

[0034] For example, assuming that the 2nd floor of the hotel has multiple areas, such as the hotel staff office area and the guest room area, among which the guest room area includes a double bed room area and a luxury guest room area, and the hotel staff office area includes a marketing department, a logistics department, and a security department, therefore, they are respectively assigned multi-level function classification labels: 2nd floor-guest room area-double bed room area; 2nd floor-guest room area-luxury guest room area; 2nd floor-hotel staff office area-marketing department; 2nd floor-hotel staff office area-logistics department; 2nd floor-hotel staff office area-security department.

[0035] The preset function tag overlap threshold is used to determine the overlap degree of the function tags. When there is an overlap between the function tags of different areas, the overlap threshold can be used to determine whether to allow the merging of these areas. For example, if the overlap degree of two area tags is greater than or equal to 70%, the two areas can be merged into one area. According to the function tag overlap threshold, the overlap degree of the function tags in the equal-area hotel function tag set is determined, and a cosine similarity or other measurement method is usually used to measure the overlap degree between two function tag sets. Cosine similarity is used to measure the angle similarity of two vectors in a high-dimensional space. In the determination of the tag overlap degree, each tag can be regarded as a dimension, and the tag set is a multi-dimensional sparse vector. The overlap degree of the hotel function tags of each two areas is calculated by the cosine similarity formula, and compared with the function tag overlap threshold. The areas with an overlap degree greater than or equal to the function tag overlap threshold are merged to obtain a consistent equal-area tag set. After the overlap degree of all hotel areas is calculated, the areas with an overlap degree meeting the threshold requirement are screened out to form a consistent equal-area tag set. These areas have high functional consistency and can be merged for more unified management. For example, assuming that areas A, B, and C meet the following conditions after overlap degree calculation: the overlap degree between area A and area B is greater than 70%, so they are merged into one area; the overlap degree between area B and area C is greater than 70%, so areas B and C are also merged; finally, areas A, B, and C can be merged into one large area to form a consistent equal-area tag set.

[0036] According to the consistent equal-area tag set, the hotel floor equal-area set is merged and connected, the areas with function tags that can be merged are merged and connected, and finally the key areas of the hotel are determined. In other words, the areas with high tag overlap degree in the multiple areas divided by the hotel are merged and connected, thereby obtaining N key areas. Merging functionally similar areas simplifies the management process and improves management efficiency. For example, assuming that a hotel has 10 floors, each floor is divided into 100 equal areas, and there are 1000 areas in total. Through the multi-level function tag system, these areas are divided into categories such as guest rooms, restaurants, and public services, and further divided into small categories such as standard rooms, luxury rooms, Chinese restaurants, and western restaurants. The function tag overlap threshold is set to 70%. After overlap degree determination, it is found that the function tag overlap degree of 200 pairs of areas reaches 70%, and these areas are merged to finally determine 50 key areas.

[0037] Through the multi-level function tag system, the hotel is accurately divided, and the consistent areas are merged and connected. The hotel can manage the key areas centrally, reduce management complexity, and improve overall operational efficiency. The determination of the function tag overlap threshold allows the hotel to flexibly adjust the area range and function layout to adapt to actual operational needs.

[0038] The monitoring demand analysis of N hotel key areas is performed, that is, the demand analysis of monitoring and collecting data of various environments, equipment, personnel activities, etc. in the key areas of the hotel, to determine which data needs to be monitored in each area. On the basis of monitoring demand analysis, equipment deployment analysis is performed, and according to different regional monitoring requirements, the appropriate sensor type, number, installation location, etc. are evaluated to ensure that the monitoring requirements can be fully and accurately met. Select the appropriate sensor type, such as temperature and humidity sensor, pressure sensor, light sensor, air quality sensor, etc. According to the size, function and monitoring requirement of the area, determine the installation quantity and distribution of each equipment, ensure that the equipment deployment can cover all areas that need to be monitored, avoid blind area or data loss. Considering the energy consumption of the equipment, select appropriate low-power sensors, or manage the equipment remotely through wireless network.

[0039] Through monitoring demand analysis and deployment coverage analysis, N hotel area monitoring demand parameters and N hotel area equipment deployment parameters are obtained, including the monitoring demand and equipment deployment parameters of each area. According to the N hotel area monitoring demand parameters and the N hotel area equipment deployment parameters, the Internet of Things equipment selection and deployment is performed, that is, according to the monitoring demand and equipment deployment parameters of each area, appropriate Internet of Things sensors and equipment (such as temperature and humidity sensor, light sensor, pressure sensor, etc.) are selected and reasonably laid out and installed to ensure that the required data can be collected in real time. After the equipment deployment is completed, regular inspection and maintenance are required to ensure the normal operation of the equipment. Through data feedback and real-time monitoring, equipment failure or performance degradation can be detected in time for maintenance or adjustment. After the Internet of Things equipment selection and deployment, N hotel Internet of Things areas are obtained, that is, the areas in the hotel after equipment selection and deployment, which are covered by Internet of Things equipment and can collect and transmit monitoring data in real time.

[0040] Through the Internet of Things equipment of N hotel Internet of Things areas, N hotel area operation data streams are collected, that is, the equipment operation data streams of each area are collected, including environmental data (such as temperature, humidity, light, air quality, etc.), equipment status data (such as air conditioner, light, TV, smart door lock, etc.). Based on real-time data stream, intelligent management of energy is realized, and air conditioners, lighting and other equipment are automatically adjusted to reduce resource waste. Real-time monitoring of equipment status and environmental data helps hotel operation and maintenance personnel to discover faults in time and give early warning to reduce equipment failure rate and avoid customer complaints.

[0041] S200: A multi-modal perception network is laid out in the guest room department of the target smart hotel, and multi-user multi-modal data streams are collected in real time through the multi-modal perception network.

[0042] Specifically, a multi-modal perception network is deployed in the guest room department of the target smart hotel to collect multi-user multi-modal data streams in real time. The target smart hotel is an intelligent hotel that uses technologies such as Internet of Things, artificial intelligence, and big data to improve hotel operations, management, and customer experience. The guest room department refers to the area within the hotel dedicated to lodging and resting, typically including guest rooms, bathrooms, corridors, and the like. The multi-modal perception network is a perception network composed of multiple sensing technologies (such as visual, sound, temperature, humidity, pressure sensors, etc.). These sensors can work together to capture and process multi-modal data streams in real time, for comprehensive perception and monitoring of the environment and user status within the hotel. For example, visual and sound sensors are deployed in the hotel corridors, and humidity, temperature, light, and environmental noise sensors are deployed in the rooms to form a multi-modal perception network. In addition, guests can interact with devices in the room using smart cards to trigger specific operations (such as turning on the light, adjusting the air conditioner, etc.).

[0043] Through the multi-modal perception network, multi-modal data streams from multiple users in the guest room department are collected and transmitted in real time while ensuring guest privacy, resulting in multi-user multi-modal data streams that include different users' check-in habits information. Through the multi-modal perception network, more personalized services are provided to users, such as adjusting environmental parameters such as temperature and lighting based on user preferences.

[0044] S300: Load demand prediction and edge node configuration in the N hotel Internet of Things areas, building a distributed edge node set.

[0045] Further, the present application S300 includes:

[0046] Based on the N hotel Internet of Things areas, historical data mining is performed to obtain N hotel area historical operation data sets. The N hotel area historical operation data sets are calculated to extract computing tasks, obtaining N hotel area computing task sets and N hotel area computing volume sets. Demand quantization analysis is performed on the N hotel area computing task sets to determine N hotel area task computing demands, including computing precision, computing privacy level, and computing rate. Based on the N hotel area computing volume sets and the N hotel area task computing demands, load demand prediction and edge node configuration are performed on the N hotel Internet of Things areas to obtain a distributed edge node set.

[0047] Further, the present application further includes the following steps:

[0048] The distributed edge topology network is built, and according to the distributed edge topology network, the total distributed network computing resource is determined; the N hotel Internet of Things regions are clustered and analyzed based on the N hotel region computation amount set and the N hotel region task computation demand, to obtain a distributed edge computing region set; the total distributed network computing resource is taken as a constraint parameter, load demand prediction and computing resource allocation are performed on the distributed edge computing region set, to determine edge region computing resource allocation information; and the distributed edge computing region set is configured based on the edge region computing resource allocation information, to obtain the distributed edge node set.

[0049] Specifically, historical data mining is performed on the N hotel Internet of Things regions, that is, historical running data of each region is obtained, to obtain a hotel region historical running data set, including running states of each region of the hotel in the past period of time, such as environmental parameters (temperature and humidity, illumination, etc.), equipment running states, personnel activities, and the like. The historical data mining is performed by analyzing past data in the hotel Internet of Things region, to mine potential rules, trends, and patterns, including equipment running states, environmental parameters (such as temperature and humidity, air quality), energy consumption data, user behavior data, and the like. Before data mining, data preprocessing is required, including data cleaning (removing missing values, outliers, and duplicate data), data standardization (such as normalizing temperature and humidity data), and data completion (complementing missing data by interpolation method). The data mining usually relies on data mining algorithms, such as clustering analysis, to obtain the historical running data set of the N hotel regions.

[0050] The N hotel region historical running data set is subjected to computation task extraction, to extract computation tasks by identifying rules in the data, to obtain the N hotel region computation task set and the N hotel region computation amount set. The computation task extraction refers to extracting tasks that need to be computed or processed from the historical running data, such as temperature and humidity control tasks (analyzing environmental change trends, and calculating optimal settings of air conditioners or humidifiers), energy efficiency prediction tasks (extracting energy efficiency prediction tasks, to predict energy consumption patterns under different device combinations), and the like, to obtain the N hotel region computation task set and the N hotel region computation amount set. The N hotel region historical running data set is analyzed, to identify computation tasks that need to be performed, and to estimate computation amounts of each task. For example, energy consumption data of a guest room is analyzed, to identify a task that needs to be performed for energy consumption prediction.

[0051] Meanwhile, for each computation task, computation amounts required by the computation task need to be analyzed, including computation resource requirements (such as CPU / GPU usage), computation task complexity, time requirements, and the like. For example, processing a temperature and humidity control task may require different computation resources and data storage, and an energy efficiency analysis task may have higher requirements for computation accuracy.

[0052] The demand quantification analysis is performed on the set of N hotel area computing tasks to determine the N hotel area task computing demands, i.e., the specific technical requirements required by each computing task during execution, including computing accuracy, computing privacy level, and computing rate. The computing accuracy is the specific accuracy requirement set for each computing task, for example, the temperature control accuracy of the temperature and humidity control task is required to be ±1℃, and the humidity control accuracy is required to be ±5%. The computing privacy level is the privacy level analysis of tasks involving user data, for example, the user behavior monitoring task requires a high privacy level, and the data needs to be encrypted or desensitized to ensure the safety of user activity data. The computing rate is the computing rate requirement set for each computing task, which is quantified by analyzing the response time required by the task and allocating appropriate computing resources according to the nature of the task. For example, real-time energy efficiency analysis or temperature regulation tasks may require a response time of 1 second, while device maintenance prediction may tolerate longer response times.

[0053] According to the distribution of the hotel Internet of Things area, network demand and device characteristics, appropriate locations are selected to deploy edge computing nodes. The edge nodes can be deployed in various areas of the hotel, especially near the terminal devices (such as guest rooms, restaurants, conference rooms, etc.) and sensors, responsible for collecting, processing and storing data, reducing communication delay to the central cloud data center. For example, an intelligent gateway can be deployed in the guest room to collect and process data from room sensors (temperature and humidity, air quality, etc.). More computing nodes are deployed in public areas (such as lobbies, restaurants) to handle more complex analysis tasks.

[0054] The hardware resources of each edge node are configured, such as processing power (CPU, GPU), storage capacity (hard disk / SSD), memory (RAM), etc., as well as appropriate network connections (such as Wi-Fi, Bluetooth, LTE, etc.) to ensure communication between nodes. The network topology is designed to ensure that each edge node can effectively communicate and cooperate to complete tasks, such as using different topologies such as star, mesh, etc., and selecting the optimal solution according to the actual needs of the hotel. For example, if a star topology is selected, the central server acts as the master node, and all other edge nodes are connected to the master node. If a mesh topology is selected, each node can be directly connected to multiple nodes, increasing the reliability and fault tolerance of the network.

[0055] Determine the task type and computing load of each edge node, including data collection, preprocessing, preliminary analysis, data storage, real-time monitoring, etc. Each node should allocate different task loads according to its computing capacity to avoid some nodes being overloaded while others are idle. When building edge computing nodes, determine the computing capacity information of each node, including CPU performance (such as core number, clock frequency), memory size, storage capacity, network bandwidth, etc. For example, node 1 includes CPU (8 cores, 2.5 GHz), memory (16 GB), storage (500 GB SSD), bandwidth (1 Gbps); node 2 includes CPU (4 cores, 2.5 GHz), memory (8 GB), storage (256 GB SSD), bandwidth (500 Mbps); node 3 includes CPU (12 cores, 2.5 GHz), memory (32 GB), storage (1 TB SSD), bandwidth (1 Gbps). Aggregate the computing resources of all edge nodes to obtain the total computing resources of the entire distributed edge network, which is obtained by adding the CPU, memory, storage and bandwidth of all nodes. For example, according to the above, the total computing resources obtained are: CPU 24 cores, memory 56 GB, storage space 1780 GB, bandwidth 2 Gbps + 500 Mbps. Due to the possibility of node failure, uneven load, etc. in actual application, a certain amount of redundancy is usually reserved for each resource type when calculating the total resources.

[0056] The distributed edge topology network refers to a network architecture constructed by deploying edge computing nodes (such as servers, routers, gateways, etc.) in different physical locations, each node can independently handle a part of the computing task, and can work with other nodes to complete larger tasks. The total computing resources refer to the comprehensive computing capacity available to all computing nodes in the distributed edge network, including CPU, memory, storage space, etc. The determination of total computing resources is the basis for resource allocation and task scheduling of edge computing system.

[0057] To avoid the scale difference between the amount of calculation and the task requirement from biasing the clustering results, it is necessary to standardize the amount of calculation and the task calculation requirement first. The purpose of standardization is to convert the amount of calculation and the task calculation requirement of each region into a unified scale, so that data of different dimensions can be compared and clustered. According to the amount of calculation set of N hotel regions and the task calculation requirement of N hotel regions, the clustering analysis of N hotel Internet of Things regions is carried out, and the regions with similar calculation amount and task calculation requirement are clustered, so that the data in the same group has similarity, while the data in different groups has large difference. Randomly select multiple initial centers, each center represents a distributed computing region. According to the amount of calculation and the task requirement of each region, it is allocated to the nearest center cluster. According to the amount of calculation and the task requirement of each region in the cluster, the cluster center is recalculated. Continue to allocate and update until the cluster center no longer changes. Finally, the distributed edge computing region set is obtained, that is, a group of regions formed by clustering analysis, the calculation requirement and the amount of calculation in each region are similar, so the resource configuration and the calculation task allocation can be unified.

[0058] The total resource of distributed network calculation is taken as a constraint parameter, that is, the resource allocation cannot exceed the total resource. The load demand prediction of the distributed edge computing region set is carried out, the load demand prediction is carried out according to the historical data of each distributed edge computing region, and the future calculation demand of each distributed edge computing region is accurately predicted, that is, the calculation task load of the distributed edge computing region in the future period of time, including the calculation demand amount, the calculation accuracy, the task processing time and the like. The historical operation data of each distributed edge computing region is collected, including but not limited to calculation task data, running state data, external influencing factors and the like. The historical data is divided into training set and verification set, usually 80% of the data is used for training, and 20% of the data is used for verifying the model effect.

[0059] According to the load demand, select the LSTM model for prediction, build the LSTM model structure, including the input layer, the LSTM layer (the core of the model, used to capture the long-term dependence in the time series), the Dropout layer (to prevent overfitting, usually add the Dropout layer after the LSTM layer), the Dense layer (used to map the output of the LSTM to the predicted load value). In order to adapt to the input requirements of LSTM, the data needs to be constructed in the form of time series. Use the training set data to train the model, select the appropriate batch size and training rounds. For example, the batch size is 32, and the training rounds are 50. During the training process, use the mean square error as the loss function to measure the gap between the predicted load and the true load. During the training process, monitor the training loss and validation loss to prevent overfitting. If the validation loss starts to rise while the training loss continues to fall, the early stopping technique can be used to automatically stop training. Use the validation set data to evaluate the prediction effect of the model, including mean square error, etc. Set the convergence conditions of the model, such as the validation set loss changes less than 0.01 for 5 consecutive rounds or the training set accuracy reaches 95%. Use the trained model to predict the load demand of the distributed edge computing region set, and get the future load demand of each region.

[0060] According to the predicted load demand, reasonably allocate computing resources under the constraints of constraint parameters, and ensure that each region can run smoothly under limited resources: according to the importance of tasks and computing demand of each region, allocate more resources to regions with higher priority. Based on the predicted load demand, dynamically adjust the allocation of computing resources. All computing resources are summarized into a shared resource pool for each region to allocate on demand. Allocate resources to multiple nodes to avoid overloading a single node. For example, if a certain edge node carries too many computing tasks, automatically transfer a part of the tasks to other idle nodes. According to the resource allocation result, configure the edge nodes in the distributed edge computing region to ensure effective use of computing resources. According to the computing task demand of each region, select appropriate hardware configuration for the edge node. For example, for regions with large computing capacity, select servers or cloud computing resources with higher performance; while for regions with low computing demand, select lightweight edge devices (such as edge computing gateways or small servers).

[0061] The number and location of edge nodes in each region are determined to ensure smooth communication and data exchange between nodes. The load monitoring system of the nodes is configured to monitor the load status of each node in real time. If the computing load of a node is too high, the task scheduling is automatically adjusted, and part of the task is migrated to a node with lower load, thereby achieving dynamic load balancing. The distributed edge node set refers to a network set composed of multiple edge computing nodes, including resource allocation in each edge region. These nodes are distributed in different regions and work together to process computing tasks, forming a powerful computing resource pool.

[0062] By assigning computing tasks to edge nodes close to the physical location of the hotel, the delay of data transmission can be reduced, the speed of task processing can be improved, and the efficiency of hotel operation can be improved. According to the task demand and computing amount, the computing resources are reasonably configured to avoid waste or shortage of computing resources, and to ensure that the tasks in each region can be completed within the specified time. Through quantitative analysis of the accuracy, privacy level and computing rate of the computing task, it is ensured that the computing result meets the specific needs of hotel management, and the customer privacy protection is fully considered.

[0063] S400: associate and map the N hotel region operation data stream and the multi-user multi-modal data stream to the distributed edge node set for abnormal feature identification, and obtain a hotel abnormal feature set.

[0064] Further, the S400 of the present application comprises:

[0065] According to the hotel operation management standard, a hotel abnormal operation feature library is defined, which includes hotel equipment level abnormalities and user behavior abnormalities. Based on the hotel abnormal operation feature library, abnormal data association and abnormal identification training are performed to construct a hotel abnormal feature identifier, and the hotel abnormal feature identifier is stored in the distributed edge node set. The N hotel region operation data stream and the multi-user multi-modal data stream are associated and mapped to the distributed edge node set to obtain an edge node associated data stream set. The hotel abnormal feature identifier is called by the distributed edge node set to identify the abnormal features of the edge node associated data stream set, and the hotel abnormal feature set is obtained.

[0066] Based on the hotel abnormal operation feature library, abnormal data association is performed to obtain a hotel abnormal operation feature data set. The hotel abnormal operation feature data set is evaluated and labeled to obtain a hotel abnormal operation type sample set and a hotel abnormal operation level sample set. Based on the hotel abnormal operation type sample set and the hotel abnormal operation level sample set, identification training is combined to construct the hotel abnormal feature identifier.

[0067] Specifically, the hotel operation management standards of the target smart hotel are obtained, and it is determined what behaviors or states are abnormal, thereby defining a hotel abnormal operation feature library, including hotel equipment level abnormalities and user behavior abnormalities. The hotel equipment level abnormalities are faults or abnormal performances of various devices (such as air conditioners, lighting systems, elevators, televisions, etc.) in the hotel, such as long-term non-operation of air conditioners, elevator failures, etc. The user behavior abnormalities are abnormal phenomena in user behavior patterns, such as frequent room changes, abnormal use duration, abnormal check-in behaviors (such as frequent check-out, multiple people staying in a single room, etc.), abnormal demands (such as multiple requests for additional services in a short period of time), etc., which may imply potential safety hazards or service problems.

[0068] According to the hotel abnormal operation feature library, abnormal data correlation is performed. The collected abnormal data (device data and user behavior data) are matched with the definitions in the hotel abnormal operation feature library to identify specific abnormal types and levels, and a hotel abnormal operation feature data set is obtained. The hotel abnormal operation feature data set is evaluated and labeled. Each sample data in the hotel abnormal operation feature data set is automatically labeled according to historical data or expert experience to determine its corresponding abnormal type and abnormal level. The abnormal type labeling is to determine the abnormal type represented by each data sample, such as device failure (air conditioner failure) and abnormal check-in behavior (multiple people staying in a single room). The abnormal level labeling is to evaluate the abnormal level of each sample according to the influence degree of the abnormality, including mild abnormality (not affecting normal operation), moderate abnormality (having a negative impact on certain user experience, but not involving safety problems), and severe abnormality (causing serious impact on hotel operation, and even threatening user safety).

[0069] According to the hotel abnormal operation type sample set and the hotel abnormal operation level sample set, the training set and the test set are divided, usually the training set accounts for 80%, and the test set accounts for 20%. The training set is used for model training, and by inputting the hotel's equipment and user behavior data, the model will learn how to identify equipment-level anomalies and user behavior anomalies, and adjust its decision boundary according to the labeled anomaly level. The class of the input sample is predicted by establishing multiple trees. Each tree is constructed based on random feature extraction from the training data, and finally the final classification result of the sample is determined by the voting mechanism. The trained model is evaluated using the test set, and the accuracy, recall rate, F1 value and other indicators are calculated to evaluate the recognition performance of the model. The trained hotel anomaly feature recognizer is stored in the distributed edge node set, and each edge node is responsible for local processing and anomaly recognition of the received sensor data. The edge node can analyze the real-time collected equipment and user behavior data in real time. When an anomaly is identified, the edge node can immediately trigger an alarm or take appropriate measures. By deploying to edge computing nodes, the model can quickly perform inference and reduce latency. For example, when the air conditioning equipment in a certain area frequently fails, the edge node can immediately send an alarm and automatically repair or adjust within a certain tolerance range.

[0070] The N hotel area operation data streams and the multi-user multi-modal data streams are associated and mapped to the distributed edge node set, that is, the operation data stream of each hotel area is combined with its corresponding multi-user multi-modal data stream, and the device operation state and user behavior are associated, so as to better capture abnormal patterns. The associated data stream is input into the distributed edge node set to obtain the edge node associated data stream set, that is, the data set obtained after the hotel area operation data stream and the multi-user multi-modal data stream are associated and mapped. Each data stream represents the operation state and user behavior of a hotel area, and is stored in the distributed edge node after mapping and processing for further analysis.

[0071] According to the edge node associated data stream set in the distributed edge node set, the hotel anomaly feature recognizer is called for processing to identify the hotel anomaly feature set, that is, the abnormal data set obtained after the operation data stream of each area of the hotel is identified, which contains detailed information of various abnormal situations in the hotel operation, such as device failure, system performance degradation, user abnormal behavior, etc., including abnormal type, abnormal level, abnormal timestamp, impact range, etc. Through data processing and anomaly detection on the distributed edge node, almost instantaneous anomaly detection and response can be achieved, improving the operation efficiency of the hotel and enabling rapid problem discovery and intervention. By mapping the data stream to the distributed edge node, the pressure on the central server can be shared, and the data transmission and processing delay can be reduced. Each node only processes the data related to it, thereby improving the overall data processing capability.

[0072] S500: Construct a hotel operation strategy library, use the hotel operation strategy library to analyze the hotel abnormal feature set, determine the target hotel operation strategy parameter, and perform early warning operation management and control on the target smart hotel based on the target hotel operation strategy parameter.

[0073] Further, the present application S500 includes:

[0074] The hotel operation strategy library is used to analyze the hotel abnormal feature set, obtain a plurality of hotel operation strategy parameters, simulate and evaluate the plurality of hotel operation strategy parameters, obtain a plurality of hotel operation parameter effects, compare and optimize the plurality of hotel operation strategy parameters according to the plurality of hotel operation parameter effects, and determine the target hotel operation strategy parameter.

[0075] Specifically, the hotel operation strategy library is a database containing various operation strategies and methods, which is used to guide the daily operation and maintenance of the hotel. According to the actual operation needs and management experience of the hotel, various operation strategies are collected and defined, and stored in the hotel operation strategy library, including equipment maintenance plan, energy management strategy, customer service process, etc. For example, for the strategies related to air conditioning, lighting, elevator and other equipment failures, define equipment failure emergency strategy; for environmental problems such as temperature and humidity changes, noise pollution, define environmental abnormality emergency strategy; for user abnormal behavior, define user behavior abnormality emergency strategy; for fire, security alarm, personnel safety and other events, generate safety event emergency strategy, each strategy should contain information such as specific operation steps, execution sequence, resource allocation, personnel scheduling, etc. Each strategy defines how to respond to different abnormal situations, how to optimize resource use, how to improve customer experience, etc., including multiple key parameters such as response time, resource allocation, priority, etc., to obtain the hotel operation strategy library.

[0076] According to the hotel operation strategy library, the hotel abnormal feature set is analyzed by strategy matching, that is, according to the obtained hotel abnormal features (equipment abnormality and personnel abnormality), the strategy suitable for the abnormality is matched in the hotel operation strategy library. For each matched strategy, output the related strategy parameters. According to the strategy in the hotel operation strategy library, the abnormal conditions in the abnormal feature set are matched and analyzed to determine which strategy is suitable for the current abnormal condition. For example, if an air conditioner in a certain area fails, the strategy matching analysis will find the preset strategy to deal with the air conditioner failure, and give the specific response time, personnel configuration, resource configuration, priority, etc.

[0077] Using simulation tools, the target smart hotel is built into a simulation model, a simulation model that can reflect the actual structure of the target smart hotel. Some actual hotel abnormal scenes (such as air conditioner failure, guest room abnormality, fire alarm, etc.) are input into the simulation model to simulate different abnormal scenes. During the simulation process, different environments and variables are set, such as the time required for repairing equipment, the effect of personnel scheduling, resource consumption, etc., to test the performance of each strategy. Multiple hotel operation strategy parameters are input into the simulation model, and the execution efficiency and effect of each strategy in the real environment are verified through simulation simulation, the effect of each strategy is evaluated, and some standards are set according to the hotel operation management standard, such as the shortest repair time, the smallest resource consumption, the maximum customer satisfaction, the lowest cost, etc., according to which the effect of each strategy is evaluated, and multiple hotel operation parameter effects are obtained.

[0078] According to the results of simulation simulation and effect evaluation, multiple hotel operation strategy parameters are compared and optimized to determine the best operation strategy parameters. The evaluation results of different strategies are quantified into numerical values for subsequent comparison. The evaluation indexes of different strategies are standardized to make each index comparable. For example, repair time, resource consumption, etc. are converted into unified quantitative indexes (such as 0-1 normalized values) according to certain standardization methods. Different weights are set for each evaluation index to reflect their importance in hotel operation, which is usually determined according to the actual operation needs and strategic goals of the hotel. The standardized results of each strategy are weighted and averaged with the set weights to calculate a comprehensive score. Then, the strategy with the highest score is selected as the operation strategy of the target hotel.

[0079] After comparison and optimization, the best operation strategy parameter set is output. Comparison and optimization refers to comparing and analyzing the effects of different operation strategies and their parameters to find the best operation strategy configuration. Through comparison and optimization, the hotel can select the operation strategy that performs best in various situations, thereby significantly improving the operation efficiency of the hotel.

[0080] Further, the present application further includes the following steps:

[0081] According to the user historical check-in data and user biological characteristics, a user behavior portrait is constructed; based on the user behavior portrait, the adjustment degree of the target hotel operation strategy parameters is corrected, and the hotel operation personalized strategy parameters are determined.

[0082] Specifically, for each customer, historical check-in data and biometric features are obtained. The historical check-in data is various records of the user when they checked into the hotel in the past, including check-in time, check-in frequency, room preference, service request, etc., reflecting the user's behavior patterns in the hotel. User biometric features are physiological and biological data of the user, such as body temperature, heart rate, body shape, preferred environmental temperature, etc., which help to further refine the user's behavior portrait. The historical data of the user is cleaned and classified, and representative behavior patterns are extracted. For example, which users prefer a lower temperature environment, and which users often forget to turn off the lights, etc.

[0083] For each user, a personalized behavior portrait is constructed based on their behavior data in historical check-ins, combined with biometric features. For example, one user prefers a lower air conditioning temperature and often does not turn on the lights; another user is used to a higher temperature and often turns on the lights.

[0084] Based on the behavior portrait of each user, the hotel can automatically adjust its operation and maintenance strategy. For example, if it is identified that a user likes a lower temperature (such as 20℃), the system will automatically adjust the air conditioning temperature to this set value during the user's stay. If a user often forgets to turn off the lights, set up intelligent light management to automatically turn off the lights when the user leaves the room, or automatically dim the lights when there is no one in the room.

[0085] Based on the user's behavior portrait, the target hotel operation and maintenance strategy parameters are adjusted, some devices or services are automatically or intelligently adjusted, and some personalized services are recommended to the user. Each hotel room can customize specific operation and maintenance strategies according to the user portrait, such as energy optimization, device scheduling, temperature control, intelligent cleaning, etc. For example, if the user's behavior portrait shows that they usually need a lower air conditioning temperature at night, the hotel can automatically lower the temperature setting during this time period. After each user check-in, their behavior feedback is recorded and the user portrait is updated. Through feedback data, the hotel can continuously optimize the personalized strategy for each user. For example, if a user's temperature preference for the air conditioner changes (such as from 20℃ to 22℃), the system can automatically adjust it during their next stay.

[0086] Through personalized adjustment based on the user's behavior portrait, not only can the customer experience be improved, but also the configuration of hotel resources can be optimized. The system automatically adjusts according to historical data during each user's stay, reducing manual intervention and management workload, increasing customer satisfaction and loyalty, not only improving customer experience, but also optimizing resource use, improving efficiency and saving costs.

[0087] In summary, the intelligent hotel Internet of Things management method based on artificial intelligence provided by the present application has the following beneficial effects:

[0088] By constructing a three-dimensional model of a hotel building, dividing the hotel building three-dimensional model into regions and deploying Internet of Things devices, N hotel Internet of Things regions are obtained, and N hotel region operation data streams of the N hotel Internet of Things regions are collected; a multi-modal sensing network is arranged in a guest room of a target smart hotel, and multi-user multi-modal data streams are collected in real time through the multi-modal sensing network; load demand prediction and edge node configuration are performed in the N hotel Internet of Things regions, and a distributed edge node set is built; the N hotel region operation data streams and the multi-user multi-modal data streams are associated and mapped to the distributed edge node set for abnormal feature identification, and a hotel abnormal feature set is obtained; a hotel operation and maintenance strategy library is constructed, the hotel operation and maintenance strategy library is used for operation and maintenance strategy analysis of the hotel abnormal feature set, target hotel operation and maintenance strategy parameters are determined, and the target smart hotel is prewarned, operated, managed and controlled based on the target hotel operation and maintenance strategy parameters. That is, by constructing a three-dimensional model of a hotel, dividing the hotel into regions and deploying Internet of Things devices, N key regions are determined, a multi-modal sensing network is arranged in a guest room, a distributed edge node set is built, and the operation data streams of the N key regions of the hotel and the data streams collected by the multi-modal sensing network are associated and mapped to the distributed edge node set for abnormal feature identification, so that target hotel operation and maintenance strategies are determined, and the efficiency of hotel operation and maintenance management is improved.

[0089] In the second embodiment, based on the same inventive concept as the first embodiment, the present application also provides an intelligent hotel Internet of Things management system based on artificial intelligence. Please refer to the accompanying drawings Figure 2 The intelligent hotel Internet of Things management system based on artificial intelligence comprises:

[0090] The area division module 11 is configured to construct a hotel building three-dimensional model, divide areas of the hotel building three-dimensional model, and deploy Internet of Things devices to obtain N hotel Internet of Things areas, and collect N hotel area operation data streams of the N hotel Internet of Things areas; the sensing network layout module 12 is configured to layout a multi-modal sensing network in a guest room department of a target smart hotel, and collect multi-user multi-modal data streams in real time through the multi-modal sensing network; the edge node building module 13 is configured to perform load demand prediction and edge node configuration in the N hotel Internet of Things areas, and build a distributed edge node set; the abnormal feature identification module 14 is configured to associate and map the N hotel area operation data streams and the multi-user multi-modal data streams to the distributed edge node set for abnormal feature identification, and obtain a hotel abnormal feature set; and the operation and maintenance management module 15 is configured to construct a hotel operation and maintenance strategy library, analyze the hotel operation and maintenance strategy library for the hotel abnormal feature set, determine target hotel operation and maintenance strategy parameters, and perform early warning operation and maintenance management and control on the target smart hotel based on the target hotel operation and maintenance strategy parameters.

[0091] Further, the area division module 11 in the intelligent hotel Internet of Things management system based on artificial intelligence is further configured to: preset an area division grid according to hotel management precision, divide floors and divide areas of the hotel building three-dimensional model according to the area division grid, and obtain a hotel floor area division area set; identify the function of each area information in the hotel floor area division area set respectively to obtain hotel floor area application functions; merge and connect the hotel floor area division area set according to the hotel floor area application functions to determine N hotel key areas; analyze monitoring demand and deployment coverage of the N hotel key areas to obtain N hotel area monitoring demand parameters and N hotel area device deployment parameters; select and deploy Internet of Things devices based on the N hotel area monitoring demand parameters and the N hotel area device deployment parameters to obtain the N hotel Internet of Things areas.

[0092] Further, the area division module 11 in the intelligent hotel Internet of Things management system based on artificial intelligence is further configured to: construct a multi-level function label system, identify the hotel floor area application functions according to the multi-level function label system to obtain hotel area multi-level function classification labels; map and assign labels to the hotel floor area division area set according to the hotel area multi-level function classification labels to obtain an area division hotel function label set; preset a function label overlap threshold, and determine the overlap of the area division hotel function label set using the function label overlap threshold to obtain a consistent area division label set; merge and connect the hotel floor area division area set based on the consistent area division label set to determine the N hotel key areas.

[0093] Further, the edge node building module 13 in the intelligent hotel IoT management system based on artificial intelligence is further used for: based on the N hotel IoT areas, historical data mining is performed to obtain N hotel area historical operation data sets; the N hotel area historical operation data sets are calculated to extract task, and N hotel area calculation task sets and N hotel area calculation amount sets are obtained; demand quantization analysis is performed on the N hotel area calculation task sets to determine N hotel area task calculation demands, including calculation accuracy, calculation privacy level and calculation rate; based on the N hotel area calculation amount sets and the N hotel area task calculation demands, load demand prediction and edge node configuration are performed on the N hotel IoT areas to obtain a distributed edge node set.

[0094] Further, the edge node building module 13 in the intelligent hotel IoT management system based on artificial intelligence is further used for: building a distributed edge topology network, determining a distributed network calculation total resource according to the distributed edge topology network; based on the N hotel area calculation amount sets and the N hotel area task calculation demands, clustering analysis is performed on the N hotel IoT areas to obtain a distributed edge calculation area set; the distributed network calculation total resource is used as a constraint parameter to perform load demand prediction and calculation resource allocation on the distributed edge calculation area set to determine edge area calculation resource allocation information; based on the edge area calculation resource allocation information, edge node configuration is performed on the distributed edge calculation area set to obtain the distributed edge node set.

[0095] Further, the abnormal feature identification module 14 in the intelligent hotel IoT management system based on artificial intelligence is further used for: defining a hotel abnormal operation feature library according to hotel operation management standards, the hotel abnormal operation feature library including hotel equipment level abnormalities and user behavior abnormalities; based on the hotel abnormal operation feature library, abnormal data correlation and abnormal identification training are performed to build a hotel abnormal feature identifier, and the hotel abnormal feature identifier is stored in the distributed edge node set; the N hotel area operation data streams and the multi-user multi-modal data streams are associated and mapped to the distributed edge node set to obtain an edge node associated data stream set; the hotel abnormal feature identifier is called by the distributed edge node set to perform abnormal feature identification on the edge node associated data stream set to obtain the hotel abnormal feature set.

[0096] Further, the abnormal feature recognition module 14 in the intelligent hotel Internet of Things management system based on artificial intelligence is further used for: performing abnormal data correlation based on the hotel abnormal operation feature library to obtain a hotel abnormal operation feature data set; performing evaluation labeling on the hotel abnormal operation feature data set to obtain a hotel abnormal operation type sample set and a hotel abnormal operation level sample set; and performing identification training and merging based on the hotel abnormal operation type sample set and the hotel abnormal operation level sample set to construct the hotel abnormal feature recognizer.

[0097] Further, the operation and maintenance management module 15 in the intelligent hotel Internet of Things management system based on artificial intelligence is further used for: performing strategy matching analysis on the hotel abnormal feature set by using the hotel operation and maintenance strategy library to obtain a plurality of hotel operation and maintenance strategy parameters; performing simulation and effect evaluation on the plurality of hotel operation and maintenance strategy parameters to obtain a plurality of hotel operation and maintenance parameter effects; and performing comparison and optimization on the plurality of hotel operation and maintenance strategy parameters according to the plurality of hotel operation and maintenance parameter effects to determine the target hotel operation and maintenance strategy parameter.

[0098] Further, the operation and maintenance management module 15 in the intelligent hotel Internet of Things management system based on artificial intelligence is further used for: constructing a user behavior portrait according to historical check-in data and user biological features of a user; and performing regulation and correction on the target hotel operation and maintenance strategy parameter based on the user behavior portrait to determine a hotel operation and maintenance individualized strategy parameter.

[0099] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The foregoing Figure 1 The intelligent hotel Internet of Things management method based on artificial intelligence in Embodiment One and the specific examples are also applicable to the intelligent hotel Internet of Things management system based on artificial intelligence in the present embodiment. For the sake of brevity of the specification, no further description is given here.

[0100] The above description of the disclosed embodiments enables a person skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0101] Obviously, for those skilled in the art, without departing from the principles of the present application, the present application can be improved and modified in several ways, and these improvements and modifications also fall within the protection scope of the present application.

Claims

1. A smart hotel IoT management method based on artificial intelligence, characterized in that, include: A 3D model of the hotel building is constructed, and the 3D model of the hotel building is divided into regions and IoT devices are deployed to obtain N hotel IoT regions. The operation data streams of the N hotel IoT regions are collected and acquired. A multimodal sensing network is deployed in the guest room section of the target smart hotel to collect and acquire multi-user multimodal data streams in real time through the multimodal sensing network; Load demand forecasting and edge node configuration are performed within the N hotel IoT areas to build a distributed edge node set. The N hotel area operation data streams and the multi-user multimodal data streams are associated and mapped to the distributed edge node set for anomaly feature identification, resulting in a hotel anomaly feature set; A hotel operation and maintenance strategy library is constructed. The hotel operation and maintenance strategy library is used to analyze the abnormal feature set of the hotel to determine the operation and maintenance strategy parameters of the target hotel. Based on the operation and maintenance strategy parameters of the target hotel, early warning operation and maintenance management is carried out on the target smart hotel.

2. The smart hotel IoT management method based on artificial intelligence as described in claim 1, characterized in that, The obtained N hotel IoT zones include: Based on the preset area division grid according to the hotel management precision, the three-dimensional model of the hotel building is divided into floors and areas according to the area division grid to obtain a set of equally divided areas of the hotel floors. Each area in the set of equally divided areas of the hotel floor is functionally identified to obtain the application functions of the hotel floor area. According to the hotel floor area application function, the hotel floor equally divided area set is merged and connected to determine N key hotel areas; Monitoring demand analysis and deployment coverage analysis are performed on the N key areas of the hotels to obtain the monitoring demand parameters and equipment deployment parameters for the N hotel areas. Based on the monitoring requirement parameters and equipment deployment parameters of the N hotel areas, IoT devices are selected and deployed to obtain the N hotel IoT areas.

3. The smart hotel IoT management method based on artificial intelligence as described in claim 2, characterized in that, The determination of N key hotel areas includes: A multi-level functional labeling system is constructed, and the application functions of the hotel floor area are identified in a multi-level manner according to the multi-level functional labeling system to obtain multi-level functional classification labels for the hotel area. The hotel floor area set is mapped and assigned values ​​according to the multi-level functional classification labels of the hotel area to obtain the hotel functional label set of the equally divided area. A preset functional tag overlap threshold is used to determine the overlap of the hotel functional tag set in the equally divided area, thereby obtaining a consistent equally divided area tag set. Based on the consistent equally divided region label set, the hotel floor equally divided region set is merged and connected to determine the N key hotel regions.

4. The smart hotel IoT management method based on artificial intelligence as described in claim 1, characterized in that, The construction of the distributed edge node set includes: Based on the N hotel IoT areas, historical data mining is performed to obtain N hotel area historical operation datasets. The computational tasks are extracted from the historical operation datasets of the N hotel areas to obtain a set of computational tasks and a set of computational amounts for the N hotel areas. A demand quantification analysis is performed on the set of computing tasks for the N hotel areas to determine the computing requirements for the N hotel areas, including computing accuracy, computing privacy level, and computing speed. Based on the computational load set of the N hotel areas and the task computation requirements of the N hotel areas, load demand prediction and edge node configuration are performed on the N hotel IoT areas to obtain a distributed edge node set.

5. The smart hotel IoT management method based on artificial intelligence as described in claim 4, characterized in that, The process of obtaining the distributed edge node set includes: Construct a distributed edge topology network, and determine the total computing resources of the distributed network based on the distributed edge topology network; Based on the set of computational loads for the N hotel areas and the task computation requirements for the N hotel areas, cluster analysis is performed on the N hotel IoT areas to obtain a set of distributed edge computing areas. Using the total distributed network computing resources as a constraint parameter, load demand prediction and computing resource allocation are performed on the distributed edge computing region set to determine edge region computing resource allocation information. Based on the edge region computing resource allocation information, edge nodes are configured in the distributed edge computing region set to obtain the distributed edge node set.

6. The smart hotel IoT management method based on artificial intelligence as described in claim 1, characterized in that, The obtained set of abnormal hotel features includes: According to hotel operation management standards, a hotel abnormal operation feature database is defined, which includes hotel equipment-level abnormalities and user behavior abnormalities. Based on the hotel abnormal operation feature library, abnormal data association and abnormal identification training are performed to construct a hotel abnormal feature recognizer, and the hotel abnormal feature recognizer is stored in the distributed edge node set; The data streams from the N hotel areas and the multi-user multimodal data streams are associated and mapped to the distributed edge node set to obtain the edge node associated data stream set. The hotel anomaly feature identifier is invoked by the distributed edge node set to identify anomaly features in the data stream set associated with the edge nodes, thereby obtaining the hotel anomaly feature set.

7. The smart hotel IoT management method based on artificial intelligence as described in claim 6, characterized in that, The construction of the hotel anomaly feature identifier includes: Based on the aforementioned hotel abnormal operation feature library, abnormal data are correlated to obtain a hotel abnormal operation feature dataset. The abnormal operation feature dataset of the hotel is evaluated and labeled to obtain a sample set of abnormal operation types and a sample set of abnormal operation levels of the hotel. The hotel anomaly feature recognizer is constructed by combining the sample sets of hotel abnormal operation types and hotel abnormal operation levels through recognition training.

8. The smart hotel IoT management method based on artificial intelligence as described in claim 1, characterized in that, The parameters for determining the target hotel's operation and maintenance strategy include: The hotel operation and maintenance strategy library is used to perform strategy matching and parsing on the hotel anomaly feature set to obtain multiple hotel operation and maintenance strategy parameters. Simulation and effect evaluation were performed on the multiple hotel operation and maintenance strategy parameters to obtain the effects of the multiple hotel operation and maintenance parameters. The multiple hotel operation and maintenance strategy parameters are compared and optimized based on their effects to determine the target hotel operation and maintenance strategy parameters.

9. The smart hotel IoT management method based on artificial intelligence as described in claim 8, characterized in that, The method further includes: Based on users' historical check-in data and user biometrics, user behavior profiles are constructed; Based on the user behavior profile, the adjustment degree of the target hotel's operation and maintenance strategy parameters is modified to determine the personalized operation and maintenance strategy parameters for the hotel.

10. A smart hotel IoT management system based on artificial intelligence, characterized in that, The steps for implementing the AI-based smart hotel IoT management method according to any one of claims 1 to 9, wherein the AI-based smart hotel IoT management system comprises: The area division module is used to construct a 3D model of the hotel building, divide the 3D model of the hotel building into areas and deploy IoT devices to obtain N hotel IoT areas, and collect the N hotel area operation data streams of the N hotel IoT areas; The sensing network deployment module is used to deploy a multimodal sensing network in the guest room section of the target smart hotel, and to collect and acquire multi-user multimodal data streams in real time through the multimodal sensing network; The edge node construction module is used to predict load demand and configure edge nodes in the N hotel IoT areas, and to build a distributed edge node set. An anomaly feature identification module is used to associate and map the N hotel area operation data streams and the multi-user multimodal data streams to the distributed edge node set for anomaly feature identification, thereby obtaining a hotel anomaly feature set; The operation and maintenance management module is used to build a hotel operation and maintenance strategy library, use the hotel operation and maintenance strategy library to analyze the hotel's abnormal feature set, determine the target hotel's operation and maintenance strategy parameters, and perform early warning operation and maintenance control on the target smart hotel based on the target hotel's operation and maintenance strategy parameters.

Images