Intelligent property management system based on AI model

Abstract

The invention provides an intelligent property management system based on an AI model, which comprises a security and protection system, an employee attendance system, a work order monitoring system and an AI regulation and control system, and is characterized in that the security and protection system, the employee attendance system and the work order monitoring system carry out information intercommunication and interaction through the AI regulation and control system; and an AI data model for risk prediction and automatic work order distribution is formed through information interaction. During use, the prediction model is established through the AI regulation and control system, on-site images, audio data and the like are captured through equipment such as a camera and a sound pickup in the security and protection system, and whether dangerous factors exist in the site or not is judged in time by combining and comparing the on-site images, the audio data and the like with the prediction model. And the data of the point where the risk factor is located is transmitted to a work order monitoring system, and a temporary work order is generated to dispatch security personnel and other personnel to the site for processing, so that the response speed of emergencies in a property management area and the convenience of property management are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of property management systems, and in particular to an intelligent property management system based on an AI model. Background Art

[0002] Property management refers to the property owner (i.e. the actual owner of a place such as a business or community) hiring an external professional service management company through bidding, selection, etc. to provide maintenance, cleaning, security and other services to the owner's supporting facilities.

[0003] Property management generally encompasses patrols, security, equipment maintenance, and service management within the property area, a complex and complex task. Furthermore, property management companies often manage multiple service locations simultaneously, requiring a significant staffing burden. Existing practices often involve assigning responsible individuals to manage specific areas within the property management area, who then report to the parent company. This can be cumbersome and prone to management chaos.

[0004] In particular, security and equipment maintenance within the scope of property management are particularly cumbersome, and it is difficult for inexperienced managers to make reasonable arrangements and handle emergencies.

[0005] Therefore, this application proposes an intelligent property management system based on AI model to improve the convenience and efficiency of property management. Summary of the Invention

[0006] In view of the shortcomings of the prior art described above, the purpose of the present invention is to provide an intelligent property management system based on an AI model to solve the problem in the prior art of a large number of property management categories and a large amount of management time.

[0007] To achieve the above objectives and other related objectives, the present invention provides an AI-based intelligent property management system, comprising:

[0008] Security system: used to control, manage, and monitor security equipment in property management areas. It analyzes the security status of property management areas through video, audio, temperature, and humidity data, and issues early warnings based on the analyzed security status. At the same time, the security system generates data packets based on the above information.

[0009] Employee Attendance System: Used to schedule property staff and store relevant work skill data of property staff. It can generate a schedule list based on the work skills of property staff and make temporary schedule adjustments based on emergencies in the property management area. At the same time, the employee attendance system generates data packages based on the above information.

[0010] Work Order Monitoring System: This system is used to generate work orders based on instructions and delegate them to staff members. It also tracks the completion status of delegated work orders and generates data packages based on the work order's delegation, completion status, and completion process.

[0011] AI control system: The AI ​​control system is connected to the Internet of Things to receive data packets generated by the security system, employee attendance system, and work order monitoring system. At the same time, the AI ​​control system parses the received data packets and generates a data model. It continues to learn and improve the data model based on the continuously received data packets. The security system, employee attendance system, and work order monitoring system communicate with each other through the AI ​​control system.

[0012] The AI ​​control system will return the generated data model to the security system, employee attendance system and work order monitoring system according to type.

[0013] Preferably, the security system includes but is not limited to: a monitoring system, an access control system, and a fire protection system;

[0014] The monitoring system includes monitoring cameras distributed in every corner of the property management area;

[0015] The access control system includes access control locks and parking lot gate identification devices installed on each door in the property management area;

[0016] The surveillance cameras, access locks, and parking lot gate recognition devices are used to collect video and audio from the property management area, and transmit the video and audio data to the AI ​​control system for storage and data model comparison, and perform risk prediction based on the comparison results;

[0017] The fire protection system includes smoke sensors and temperature sensors distributed in every corner of the property management area;

[0018] The smoke sensor and temperature sensor are used to monitor the smoke and temperature in the property management area, and transmit the smoke and temperature data to the AI ​​control system for storage and data model comparison, and perform risk prediction based on the comparison results.

[0019] Preferably, the AI ​​control system pre-inputs data on sensitive actions, sensitive audio, key monitoring time periods, crowd gatherings, ambient smoke levels, and seasonal temperatures to form an early warning data model;

[0020] The real-time on-site data captured by surveillance cameras, access locks, parking lot gate recognition devices, smoke sensors, and temperature sensors are compared with the initial data model to generate early warning information. The early warning information is fed back to the AI ​​control system through data packets, and then transmitted to the work order monitoring system through the AI ​​control system to generate inspection work orders. The inspection work orders are delegated to the security department through the work order monitoring system.

[0021] Preferably, the sensitive actions include but are not limited to hijacking actions, falling actions, and collision actions;

[0022] Sensitive audio includes but is not limited to threatening information, stolen information, volume decibel recognition, and special command recognition.

[0023] Preferably, the employee attendance system includes a scheduling system and a skills database;

[0024] Scheduling system: used to schedule the working hours of property staff on a daily basis, and to make temporary shifts based on feedback from the security system and work order monitoring system;

[0025] Skill database: used to store the professional skills and skill level information of property staff;

[0026] The scheduling system schedules shifts based on the skills information of property employees, and makes temporary shifts based on the data packet information fed back by the security system and work order monitoring system.

[0027] Preferably, the work order monitoring system includes a work order generation module, a work order delegation module, and a work order status tracking module;

[0028] The work order generation module generates temporary work orders through manual instructions and forecast data;

[0029] After receiving the generated work order, the work order delegation module assigns the work order to the property staff according to the staff's skill data information. The property staff completes the work order according to the work order content.

[0030] The work order status tracking module calls the security system to track the employee who receives the work order, the work order site, and the status of the target equipment of the work order, determines the completion status of the work order, and generates a model data package based on the completion status of the work order to expand the data model in the AI ​​control system.

[0031] Preferably, the work order status tracking module determines the employee's arrival time, execution time, and end time by calling a surveillance camera at the work order execution location;

[0032] The work order status tracking module calls the operating status of the target equipment on the work order and estimates the completion status of the maintenance work order of the target equipment;

[0033] The work order status tracking module determines the execution status of the work order and the consumables consumption status of the work order based on the report of the work order executor.

[0034] Preferably, the work order status tracking module packages the maintenance data and transmits it to the AI ​​control system to expand the data model in the AI ​​control system;

[0035] The data model in the AI ​​control system adjusts the equipment information accordingly based on the feedback information, thereby generating an automatic work order generation model.

[0036] Preferably, the work order automatic generation model includes but is not limited to: theoretical service life of equipment, equipment working environment, maintenance records, equipment aging rate, equipment replacement time, equipment maintenance cycle, equipment maintenance interval and skilled personnel evaluation.

[0037] The method for establishing and learning a model of an intelligent property management system based on an AI model includes the following steps:

[0038] S1. Manually transmit data containing security, attendance, and work order delegation to the server for data feeding and for the AI ​​neural network to call;

[0039] S2. The AI ​​neural network screens and identifies data sets based on pre-set rules, classifies data sets based on data weights, data layers, and bias parameters, and uses backpropagation algorithms and different types of loss functions to understand and optimize data. It uses gradient descent, stochastic gradient descent, RM Prop, and Adam methods to form a preliminary AI model.

[0040] S3. Based on the AI ​​big model established above, AI can filter and improve the data generated by the security, attendance, and work order delegation systems at any time, and learn from human operating habits to expand the scale of the AI ​​big model and realize the construction of anthropomorphic thinking;

[0041] S4. Fine-tune AI models and logic through manual supervision and data correction to align AI logic with human expectations, learn preferences from human (or manual) feedback to reduce bias and review models;

[0042] S5. When AI neural networks are improving large models, they quantify data through the model's weights to reduce the model's storage cost and increase the model's computing speed.

[0043] S6. The AI ​​data model is deployed to local devices and the main server through the network or hardware devices. The main server obtains the data model from each local server through the network for deep learning and improves the model. The improved model is fed back to the corresponding local server through the network to optimize the local model.

[0044] As described above, the intelligent property management system based on the AI ​​model of the present invention has the following beneficial effects:

[0045] 1. The present invention establishes a prediction model through an AI control system, and captures on-site images, audio data, etc. through cameras, microphones and other equipment in the security system. By combining and comparing on-site images, audio and other data with the prediction model, it can timely determine whether there are dangerous factors on the scene, and transmit the data of the dangerous factors to the work order monitoring system, generate temporary work orders and dispatch security personnel and other personnel to the scene for processing, thereby achieving the response speed of emergencies in the property management area and the convenience of property management.

[0046] 2. The present invention connects the AI ​​control system with the employee attendance system, establishes an employee skill database, cooperates with the employee technical database and the work order monitoring system, and combines the historical work order records of the work order monitoring system to pre-judge the maintenance time of the equipment to arrange employees with corresponding skills to be on duty, so that when the equipment encounters an emergency, work orders can be dispatched in a timely and reasonable manner, thereby improving the efficiency of equipment maintenance and the work order completion rate.

[0047] At the same time, the employee skills database can be updated in real time based on employee self-feedback and big data diagnostic tests, so that the data is always up to date, ensuring the accuracy and effectiveness of AI work orders.

[0048] 3. The present invention monitors the completion of work orders through a work order monitoring system and establishes an equipment maintenance prediction model based on the equipment's maintenance cycle, historical maintenance status, and historical maintenance intervals. This allows the system to predict potential equipment failures based on the operating status of existing equipment and arrange maintenance personnel to inspect and repair the equipment in advance, thereby reducing equipment maintenance costs and improving maintenance efficiency. At the same time, it reduces the risk of equipment damage and the scope of impact caused by equipment damage.

[0049] 4. The present invention sets up a local server in each property management area, stores and processes the data of the property management area through the local server, and transmits the stored and processed data to upstream equipment through the Internet of Things for management, so as to improve the property company's monitoring of downstream property management areas and improve the timeliness of information exchange.

[0050] At the same time, storing and preprocessing information through local servers can reduce the pressure on the main server and improve the response speed of the main server.

[0051] Therefore, the present invention effectively overcomes various shortcomings in the prior art and has high industrial utilization value. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 Shown is a block diagram of the property management system of the present invention.

[0053] Figure 2 Shown is a flowchart of the security system of the present invention.

[0054] Figure 3 Shown is a flowchart of the noise identification and automatic work order generation process of the present invention.

[0055] Figure 4 Shown is a flowchart of the employee attendance system of the present invention.

[0056] Figure 5 Shown is a flowchart of the scheduling review process of the scheduling system of the present invention.

[0057] Figure 6 Shown is a flowchart of the work order types and their establishment according to the present invention.

[0058] Figure 7 Shown is a flowchart of the work order monitoring system of the present invention. DETAILED DESCRIPTION

[0059] The following describes the implementation of the present invention through specific embodiments. People skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification.

[0060] See also Figures 1 to 7 . It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the limiting conditions for the implementation of the present invention. Therefore, they have no substantive technical significance. Any modification of the structure, change in the proportional relationship or adjustment of the size should still fall within the scope of the technical content disclosed by the present invention without affecting the efficacy and purpose that can be achieved by the present invention. At the same time, the terms such as "upper", "lower", "left", "right", "middle" and "one" quoted in this specification are only for the convenience of description, and are not used to limit the scope of the implementation of the present invention. Changes or adjustments in their relative relationships should also be regarded as the scope of the implementation of the present invention without substantially changing the technical content.

[0061] like Figure 1 As shown, the present invention provides an intelligent property management system based on an AI model, including a security system, an employee attendance system, a work order monitoring system and an AI control system. The security system, the employee attendance system and the work order monitoring system communicate and interact with each other through the AI ​​control system.

[0062] Security system: Used to control, manage, and monitor security equipment in property management areas. This includes monitoring whether security equipment is online, whether data transmission is normal, and setting device parameters and status. The security system analyzes the security status of the property management area using video, audio, and temperature and humidity data, and issues early warnings based on this analysis. The security system also generates data packets based on this information, which are fed back to the AI ​​control system via the Internet of Things. The AI ​​control system then builds a large AI model based on the data packets and combines them with historical information, and continuously learns and refines this model.

[0063] Employee Attendance System: Used to schedule property staff and store their relevant work skills data. The system automatically generates rosters based on staff skills, placing staff with appropriate skills in appropriate positions. This improves employee efficiency and reduces workload for property management personnel. The system can also make temporary adjustments to shifts based on emergencies in the property management area. For example, after the AI ​​control system receives an early warning from the security system, it generates a temporary work order for the work order control system. This temporary work order is then sent to the employee attendance system via the AI ​​control system. Based on the work order content and employee skill data in the skills database, the system dispatches appropriate property staff to the site to handle the early warning event. Simultaneously, the system generates data packets based on this information. These data packets are fed back to the AI ​​control system via the Internet of Things (IoT) to inform and improve the AI ​​model, enabling it to better handle emergencies and predict and calculate their probability.

[0064] Work order monitoring system: used to generate work orders based on instructions and delegate work orders to staff. The instructions mentioned above are either manually issued or automatically generated by the system based on early warning information. At the same time, the work order monitoring system tracks the completion status of delegated work orders to analyze the completion of the work orders. In particular, for work orders for equipment maintenance, the completion status of the work orders needs to be tracked in more detail to analyze data such as the purpose of equipment maintenance, maintenance intervals, and causes of damage. Based on the delegation, completion status, and completion process of the work order, a data packet is generated and fed back to the AI ​​control system to improve the AI ​​big model. The AI ​​big model can predict the approximate time and cause of the next equipment failure based on the above information, making it easier for property management personnel to manage the equipment.

[0065] AI Control System: This AI control system is connected to the Internet of Things (IoT) to receive data packets generated by the security system, employee attendance system, and work order monitoring system. It analyzes and coordinates these data packets and transmits them by type among the security system, employee attendance system, and work order monitoring system. Simultaneously, the AI ​​control system parses the received data packets and generates an AI data model. The AI ​​data model continues to learn and improve based on the continuously received data packets, enabling it to predict new events and rationally arrange them based on historical events and how they were handled.

[0066] The AI ​​control system returns the generated data models to the security system, employee attendance system, and work order monitoring system based on their type, enabling local data processing, reducing the pressure on the central server and improving data response speed. Furthermore, the security system, employee attendance system, and work order monitoring system can exchange information through local servers.

[0067] like Figure 2 As shown, in some embodiments, the security system described in the present invention includes, but is not limited to, monitoring systems, access control systems, fire protection systems, and other security systems. The monitoring system includes surveillance cameras distributed throughout the property management area, which monitor the entire property management area through video and audio, and store the video and audio through hardware devices. The access control system includes access locks installed on each door in the property management area and parking lot gate recognition devices. The access locks manage people entering and exiting the property management area using pre-entered information. The parking lot gate recognition devices are used to manage vehicles.

[0068] The surveillance cameras, access locks, and parking lot gate recognition devices are used to capture video and audio from the property management area. The video and audio data is transmitted to the AI ​​control system for storage and data model comparison. The input data is compared with historical data, and risk predictions are made based on the comparison results. When newly input data poses a risk, the AI ​​control system transmits the data to the work order monitoring system. The work order monitoring system generates and issues different types of temporary work orders based on the risk type and continuously tracks and records the completion of temporary work orders. The final results of the work order completion are fed back to the AI ​​control system in the form of data packets to improve the AI ​​data model.

[0069] The fire protection system includes smoke and temperature sensors distributed throughout the property management area. These sensors monitor smoke and temperature within the property management area and transmit these data to the AI ​​control system for storage and comparison with data models. Risk prediction is then performed based on the comparison results. If a fire is detected, the data is immediately transmitted to the fire control center for an alarm. Simultaneously, this data is backed up and stored in the AI ​​control system for use in improving the AI ​​data model.

[0070] Specifically, the abnormal data of smoke and temperature sensors will first be transmitted to the fire control host. The fire control host will make warning actions (such as alarm sound, red light, display alarm area, etc.) based on simple logical judgment. At the same time, the transmission signal is transmitted to the AI ​​control system through the Internet of Things. The AI ​​control system quickly makes risk predictions through intelligent comparison and analysis.

[0071] When the prediction result is a fire, the following work order is generated at the same time:

[0072] 1. Temporary patrol work order - immediately dispatched to the patrol post closest to the fire point (Note: In this emergency, the AI ​​control system will call the order taker's mobile phone through the automatic voice system) to confirm whether the scene is a real fire;

[0073] 2. Distribute work orders to the on-duty personnel in the fire control room to prepare for fire monitoring and pre-operation;

[0074] 3. Send work orders to the project manager and make preparations for fire response;

[0075] 4. Based on the first step patrol work order feedback, if it is confirmed that the fire is out of control, the AI ​​control system will immediately call the 119 fire alarm number and report the location, fire status and other information;

[0076] 5. Send the second work order to the on-duty personnel in the fire control room to initiate the fire emergency operation.

[0077] like Figure 2 As shown, in some embodiments, the AI ​​control system of the present invention pre-inputs data on sensitive actions (such as: armed actions, hijacking actions, high-altitude climbing, etc.), sensitive audio (such as: threatening language keywords, fraud keywords, gambling information, etc.), key monitoring time periods (such as: midnight time period, etc.), crowd gathering, environmental smoke volume, seasonal temperature (indoor and outdoor temperatures are different in different seasons, and sufficient outdoor sunshine in summer is prone to high temperature misjudgment. Therefore, it is necessary to evaluate seasonal temperatures to reduce the occurrence of misjudgment) to form an early warning data model.

[0078] Real-time on-site data captured by surveillance cameras, access locks, parking lot gate identification devices, smoke sensors, and temperature sensors is compared with the initial data model. Based on the captured anomaly information, early warning information is generated and fed back to the AI ​​control system via data packets. This information is then transmitted to the work order monitoring system, which generates inspection work orders. The inspection work orders are then delegated to the security office through the work order monitoring system, allowing property personnel to promptly access the site for processing, improving the convenience of property management and the timeliness of emergency response.

[0079] like Figure 2 As shown, in some embodiments, the sensitive actions described in the present invention include but are not limited to hijacking actions, falling actions (actions such as falling into water can be recorded in water areas), collision actions, etc. The above action information can be learned from the Internet and imported and obtained through AI. And the geographical location of the above information (such as the channel number of the camera, the location information of the notes, etc.) can be obtained through devices such as cameras that currently capture information. When high-risk warnings such as hijacking actions occur, security personnel can be prompted to carry explosion-proof equipment for processing. When assistance is needed for early warnings such as falling actions, the AI ​​control system or local server will send assistance information to nearby staff through real-time communication to improve rescue efficiency. Collision actions are mainly used to monitor collision information of vehicles on roads and parking lots, for evidence preservation, and to facilitate the handling of subsequent disputes.

[0080] Sensitive audio includes but is not limited to threat information, theft information, volume decibel recognition, and special command recognition. The threat information and theft information keywords are used to improve the security of property management areas, enabling property security personnel to promptly identify potential risks.

[0081] like Figure 3 As shown, decibel volume recognition is used to monitor noise nuisances, enabling property management staff to promptly correct illegal renovations and other issues. For example, when the camera's microphone detects a noise level exceeding 60 decibels, the system compares the volume of multiple surrounding microphones to determine the location of the noise. Based on the noise's location, the system queries whether the location has a renovation application or falls within a rest period. If the noise occurs during a rest period or if renovations have not been reported, which violates property management regulations, the AI ​​control system sends a message to the work order monitoring system, which then issues a work order and dispatches staff to investigate and dissuade the renovations. If renovation procedures are in place, the system automatically ignores these warnings to reduce wasted personnel.

[0082] Special command recognition is used to ensure the safety of property personnel. By setting keyword commands in the server, special command alarms can be issued when property personnel themselves are coerced, so as not to aggravate the emotions of the criminals.

[0083] like Figure 4 and Figure 5 As shown, in some embodiments, the employee attendance system of the present invention includes a scheduling system and a skills database. The scheduling system is used to schedule the working hours of property employees on a daily basis and push the information to employees through an app or other means.

[0084] Specifically: Logistics and HR staff pre-determine positions and personnel arrangements based on the different property management projects, implementing pre-processing. The AI-controlled scheduling system automatically schedules shifts based on these pre-defined positions and personnel arrangements. The scheduling results are fed back to management for final review.

[0085] Employees work according to their daily schedules. These can be set manually by logistics staff or recommended by AI data models, which are then manually reviewed and distributed. The scheduling system can also create temporary shifts based on feedback from the security system and work order monitoring system to handle emergencies. The skills database stores information on the professional skills and skill levels of property management employees.

[0086] The scheduling system schedules shifts based on property management staff's skill information and makes temporary arrangements based on data packets fed back from the security system and work order monitoring system. Professional skills and skill levels serve as a reference for the scheduling system, ensuring that staff with the appropriate professional skills are assigned to the appropriate positions. Furthermore, in the event of an emergency within the property management area, the scheduling system can assign appropriate staff members based on data in the skills database, improving the efficiency and effectiveness of emergency handling.

[0087] Both the automatic and temporary shift schedules mentioned above can be manually modified and adjusted to enhance shift flexibility. Furthermore, the system automatically detects employees who are unable to report to work on time due to a temporary leave, based on their leave history and other data.

[0088] like Figure 6 and Figure 7 As shown, in some embodiments, the work order monitoring system of the present invention includes a work order generation module, a work order delegation module and a work order status tracking module.

[0089] The work order generation module generates temporary work orders based on manual instructions and forecast data, forming regular work orders and temporary emergency work orders issued by manual instructions, regular work orders and temporary emergency work orders automatically dispatched by the system.

[0090] Specifically, work orders can be divided into real-time work orders and planned work orders according to the type of work order. Real-time work orders are orders placed by customers (such as common events such as garbage removal and obstacle removal) and work orders placed by quality inspectors (such as events such as unqualified hygiene that need to be handled). After such work orders are submitted to the system, the system will immediately generate a work order based on the content of the work order. Planned work orders are work plan work orders made by staff based on nodes such as months, quarters, or years. After being formulated, such plans are stored in the server in a pre-processing state. When the work order processing time arrives, such work orders are assigned for execution.

[0091] After receiving the generated work order, the work order delegation module assigns the work order to a property management employee based on the employee's skill data. The property management employee then completes the work order according to the work order content. After completing the corresponding work order, the property management employee provides detailed feedback to the system for data storage and expansion of the AI ​​data model.

[0092] At the same time, after receiving a work order, the employee can forward the work order through the system based on his or her own situation at the time and the specific content of the work order. The forwarded work order will be completed and recorded by the recipient.

[0093] When an employee is executing a normal work order, if he receives a work order with a higher priority or the work order cannot be completed (such as temporary rescue, fire handling, lack of maintenance supplies, etc.), he can complete the closure of the work order according to the report and review to avoid the problem of missing attendance. This is especially true for patrol personnel. For example, if someone receives an inspection work order, the content of the work order is to inspect areas A, B, C, D, etc. When this person inspects area C, an emergency situation occurs and needs to be temporarily handled, resulting in the inability to execute subsequent inspection events and the inability to close the work order task. As a result, this person may have the problem of missing work order completion affecting his performance. To address this problem, when this person needs to perform an emergency task, he can report it through the system, and the above-mentioned unfinished work order will be closed after system or manual review.

[0094] At the same time, the system will regularly conduct comprehensive assessments of the completion of work orders to improve the completion quality of subsequent work orders and implement a reward and punishment system with logistics personnel and other departments.

[0095] The work order status tracking module tracks the status of employees who receive work orders, work order sites, and target equipment of work orders by calling the security system. Specifically, it calls surveillance cameras to confirm whether employees arrive at the designated location at the designated time, whether employees complete the work order in compliance, and determines whether the repaired equipment has returned to online status by calling equipment information. It is used to determine the completion status of work orders and to generate model data packets based on the completion status of work orders to expand the data model in the AI ​​control system. The data model in the AI ​​control system is combined with historical maintenance information to analyze and estimate the current status of the equipment, thereby estimating the next maintenance cycle of the equipment. It enables property management personnel to receive early warnings and obtain information about possible equipment failures at the same time, realize advance maintenance and targeted maintenance, reduce the cost of equipment maintenance and reduce the impact caused by equipment damage, and arrange appropriate maintenance personnel to perform maintenance during maintenance to improve maintenance efficiency.

[0096] like Figure 7 As shown, in some embodiments, the work order status tracking module of the present invention uses surveillance cameras at the work order execution site to determine the employee's arrival time, execution time, and completion time, thereby estimating the next maintenance cycle for the same issue and facilitating the development of appropriate contingency plans. Furthermore, combined with reports from the work order executors, it determines the consumables required for maintenance, thereby estimating the quarterly and annual maintenance costs for all equipment.

[0097] The work order status tracking module uses the operating status of the target device on the work order to estimate the completion status of the target device's maintenance work order based on the device's operating status, thereby determining the device's service life. This calculation is based on the device's theoretical service life, the device's maintenance level, the device's operating environment, and other factors. Using a large amount of historical maintenance data, a prediction model is established to estimate the device's next maintenance period.

[0098] The work order status tracking module determines the execution status of the work order and the consumables consumption status of the work order based on the reports submitted by the work order executor. This is used to estimate the next maintenance cost, thus facilitating the logistics department's consumables preparation and financial accounting.

[0099] like Figure 7 As shown, in some embodiments, the work order status tracking module of the present invention packages maintenance data and transmits it to the AI ​​control system to expand the data model in the AI ​​control system, continuously improving the data model in the AI ​​control system and reducing prediction errors. Furthermore, it can refine work orders based on information such as the type of equipment damage and the category of maintenance required, and select more suitable execution personnel, thereby ensuring more accurate and timely issuance of temporary work orders.

[0100] Moreover, the data model in the AI ​​control system can adjust and update the equipment information accordingly based on the feedback information, and can automatically generate a model based on the new equipment information and accurately generate work orders, thereby improving the prediction hit probability.

[0101] like Figure 7 As shown, in some embodiments, the automatic work order generation model of the present invention includes but is not limited to: theoretical service life of equipment, equipment working environment, maintenance records, equipment aging rate, equipment replacement time, equipment maintenance cycle, equipment maintenance interval and skilled personnel evaluation. Among them, the theoretical service life of equipment is the threshold of the equipment maintenance interval, and the predicted data will not be greater than this threshold, which is used to reduce the amount of calculation of the server. The equipment working environment is used for auxiliary evaluation, based on the historical working life of equipment in different environments, thereby reducing the estimation deviation caused by environmental factors. Maintenance records (maintenance techniques, quality of consumables used, completion of maintenance, etc.) and equipment aging rate are mainly provided by experienced maintenance technicians and are one of the important bases for estimating the next maintenance interval of the equipment. After the equipment is replaced, the data in the data model will be updated accordingly to determine the new maintenance time and interval. The equipment maintenance cycle and equipment maintenance interval are used to issue regular maintenance work orders, and the next maintenance time is estimated and predicted in combination with historical maintenance data and skilled personnel evaluation.

[0102] Temporary work orders are generated based on an automatic work order generation model package that predicts the past maintenance intervals, theoretical service life, maintenance time, equipment working environment, last maintenance status, and skilled personnel assessment data of existing equipment. This is used to assign employees to perform temporary maintenance or inspections of equipment during estimated high-risk time periods for emergencies, thereby reducing losses caused by sudden equipment failures and irreversible damage to the equipment.

[0103] Before automatically generating a work order, the automatic work order generation model transmits the data packet to the employee attendance system. The employee attendance system then matches property management personnel with appropriate skills based on the data from the automatic work order generation model. This allows for pre-emptive scheduling to address emergencies, preventing sudden staff changes from preventing the system from responding correctly.

[0104] The method for establishing and learning a model of an intelligent property management system based on an AI model includes the following steps:

[0105] S1. Manually transmit data containing security, attendance, and work order delegation to the server for data feeding and for AI neural network to call (such as how to determine whether security equipment is online, dangerous action models, sensitive voice, attendance rules, skill certificate weights, work order types, skills required for work orders, etc.);

[0106] S2. The AI ​​neural network screens and identifies data sets based on pre-set rules, classifies data sets based on data weights, data layers, and bias parameters, and independently understands and optimizes data through backpropagation algorithms and different types of loss functions, such as gradient descent, stochastic gradient descent, RM Prop, and Adam's method, thereby forming a preliminary AI large model;

[0107] S3. Based on the AI ​​big model established above, AI will filter and improve the data generated by the security, attendance, and work order delegation systems at any time (classifying them according to the security, attendance, and work order systems, and then eliminating duplicate and biased information). It will also perform fitting learning based on human operating habits, expand the scale of the AI ​​big model, and realize the construction of anthropomorphic thinking, so that the data model will be gradually improved and gradually have human thinking logic;

[0108] S4. Through manual supervision and fine-tuning of AI models and logic, data correction is performed to align AI logic with human expectations. By learning preferences from human (or manual) feedback, biases are reduced, models are audited, and events that do not conform to human thinking logic are avoided.

[0109] S5. When refining large models, AI neural networks quantize data using the model's weights to reduce storage costs and increase computational speed. For example, low-weight data stored using 16- or 32-bit representations can be converted to 4-bit representations for storage. While this approach reduces data accuracy, it significantly increases computational speed and reduces storage space. Furthermore, low-weight data is only used to refine the data model and does not play a decisive role in the AI ​​neural network's decision-making.

[0110] S6. AI data models are deployed to local devices and a central server via the network or hardware. The central server retrieves data models from each local server via the network for deep learning and model refinement. The refined models are then fed back to the corresponding local servers via the network to optimize the local models. Model integration is performed on a high-performance central server. The integrated data is then transferred to lower-performance local servers for fusion, accelerating data integration and reducing equipment costs.

[0111] To sum up, the intelligent property management system based on the AI ​​model of the present invention establishes a prediction model through the AI ​​control system, and captures on-site images, audio data, etc. through cameras, microphones and other equipment in the security system. By combining and comparing on-site images, audio and other data with the prediction model, it can timely determine whether there are dangerous factors on the scene, and transmit the data of the dangerous factors to the work order monitoring system, generate temporary work orders and dispatch security personnel and other personnel to the scene for processing, thereby achieving the response speed of emergencies in the property management area and the convenience of property management.

[0112] The present invention connects the AI ​​control system with the employee attendance system, establishes an employee skill database, cooperates with the employee technical database and the work order monitoring system, and combines the historical work order records of the work order monitoring system to pre-judge the maintenance time of the equipment to arrange employees with corresponding skills to be on duty. In this way, when an equipment emergency occurs, work orders can be dispatched in a timely and reasonable manner, thereby improving the efficiency of equipment maintenance and the work order completion rate.

[0113] At the same time, the employee skills database can be updated in real time based on employee self-feedback and big data diagnostic tests, so that the data is always up to date, ensuring the accuracy and effectiveness of AI work orders.

[0114] The present invention monitors the completion of work orders through a work order monitoring system and establishes an equipment maintenance prediction model based on the equipment's maintenance cycle, historical maintenance status, historical maintenance intervals, etc. This allows the system to predict potential equipment failures based on the operating status of existing equipment and arrange maintenance personnel to inspect and maintain the equipment in advance, thereby reducing equipment maintenance costs and improving maintenance efficiency. At the same time, it also reduces the risk of equipment damage and the scope of impact caused by equipment damage.

[0115] The present invention sets up a local server in each property management area, stores and processes the data of the property management area through the local server, and transmits the stored and processed data to upstream equipment for management through the Internet of Things, so as to improve the property company's monitoring of downstream property management areas and improve the timeliness of information exchange.

[0116] At the same time, storing and preprocessing information through local servers can reduce the pressure on the main server and improve the response speed of the main server.

[0117] Therefore, the present invention effectively overcomes various shortcomings in the prior art and has high industrial utilization value.

[0118] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Anyone skilled in the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by one of ordinary skill in the art without departing from the spirit and technical principles disclosed herein are intended to be covered by the claims of the present invention.

Claims

1. An intelligent property management system based on AI model, characterized in that: include: Security system: used to control, manage, and monitor security equipment in property management areas. It analyzes the security status of property management areas through video, audio, temperature, and humidity data, and issues early warnings based on the analyzed security status. At the same time, the security system generates data packets based on the above information. Employee Attendance System: Used to schedule property staff and store relevant work skill data of property staff. It can generate a schedule list based on the work skills of property staff and make temporary schedule adjustments based on emergencies in the property management area. At the same time, the employee attendance system generates data packages based on the above information. Work Order Monitoring System: This system is used to generate work orders based on instructions and delegate them to staff members. It also tracks the completion status of delegated work orders and generates data packages based on the work order's delegation, completion status, and completion process. AI control system: The AI ​​control system is connected to the Internet of Things to receive data packets generated by the security system, employee attendance system, and work order monitoring system. At the same time, the AI ​​control system parses the received data packets and generates a data model. It continues to learn and improve the data model based on the continuously received data packets. The security system, employee attendance system, and work order monitoring system communicate with each other through the AI ​​control system. The AI ​​control system will return the generated data model to the security system, employee attendance system and work order monitoring system according to type.

2. The AI ​​model-based intelligent property management system according to claim 1, characterized in that: The security system includes but is not limited to: monitoring system, access control system, fire protection system; The monitoring system includes monitoring cameras distributed in every corner of the property management area; The access control system includes access control locks and parking lot gate identification devices installed on each door in the property management area; The surveillance cameras, access locks, and parking lot gate recognition devices are used to collect video and audio from the property management area, and transmit the video and audio data to the AI ​​control system for storage and data model comparison, and perform risk prediction based on the comparison results; The fire protection system includes smoke sensors and temperature sensors distributed in every corner of the property management area; The smoke sensor and temperature sensor are used to monitor the smoke and temperature in the property management area, and transmit the smoke and temperature data to the AI ​​control system for storage and data model comparison, and perform risk prediction based on the comparison results.

3. The AI ​​model-based intelligent property management system according to claim 2, characterized in that: The AI ​​control system pre-inputs data on sensitive actions, sensitive audio, key monitoring time periods, crowd gatherings, ambient smoke levels, and seasonal temperatures to form an early warning data model; The real-time on-site data captured by surveillance cameras, access locks, parking lot gate recognition devices, smoke sensors, and temperature sensors are compared with the initial data model to generate early warning information. The early warning information is fed back to the AI ​​control system through data packets, and then transmitted to the work order monitoring system through the AI ​​control system to generate inspection work orders. The inspection work orders are delegated to the security department through the work order monitoring system.

4. The AI ​​model-based intelligent property management system according to claim 3, characterized in that: The sensitive actions include but are not limited to hijacking, falling, and collision; Sensitive audio includes but is not limited to threatening information, stolen information, volume decibel recognition, and special command recognition.

5. The AI ​​model-based intelligent property management system according to claim 1, characterized in that: The employee attendance system includes a scheduling system and a skills database; Scheduling system: used to schedule the working hours of property staff on a daily basis, and to make temporary shifts based on feedback from the security system and work order monitoring system; Skill database: used to store the professional skills and skill level information of property staff; The scheduling system schedules shifts based on the skills information of property employees, and makes temporary shifts based on the data packet information fed back by the security system and work order monitoring system.

6. The AI ​​model-based intelligent property management system according to claim 1, characterized in that: The work order monitoring system includes a work order generation module, a work order delegation module, and a work order status tracking module; The work order generation module generates temporary work orders through manual instructions and forecast data; After receiving the generated work order, the work order delegation module assigns the work order to the property staff according to the staff's skill data information. The property staff completes the work order according to the work order content. The work order status tracking module calls the security system to track the employee who receives the work order, the work order site, and the status of the target equipment of the work order, determines the completion status of the work order, and generates a model data package based on the completion status of the work order to expand the data model in the AI ​​control system.

7. The AI ​​model-based intelligent property management system according to claim 6, characterized in that: The work order status tracking module determines the employee's arrival time, execution time, and end time by calling the surveillance camera at the work order execution location; The work order status tracking module calls the operating status of the target equipment on the work order and estimates the completion status of the maintenance work order of the target equipment; The work order status tracking module determines the execution status of the work order and the consumables consumption status of the work order based on the report of the work order executor.

8. The AI ​​model-based intelligent property management system according to claim 7, characterized in that: The work order status tracking module packages the maintenance data and transmits it to the AI ​​control system to expand the data model in the AI ​​control system; The data model in the AI ​​control system adjusts the equipment information accordingly based on the feedback information, thereby generating an automatic work order generation model.

9. The AI ​​model-based intelligent property management system according to claim 8, characterized in that: The automatic work order generation model includes but is not limited to: theoretical service life of equipment, equipment working environment, maintenance records, equipment aging rate, equipment replacement time, equipment maintenance cycle, equipment maintenance interval and skilled personnel evaluation.

10. A method for establishing and learning a model for an intelligent property management system based on an AI model, characterized by: The following steps are included: S1. Manually transmit data containing security, attendance, and work order delegation to the server for data feeding and for the AI ​​neural network to call; S2. The AI ​​neural network screens and identifies data sets based on pre-set rules, classifies data sets based on data weights, data layers, and bias parameters, and uses backpropagation algorithms and different types of loss functions to understand and optimize data. It uses gradient descent, stochastic gradient descent, RM Prop, and Adam methods to form a preliminary AI model. S3. Based on the AI ​​big model established above, AI can filter and improve the data generated by the security, attendance, and work order delegation systems at any time, and learn from human operating habits to expand the scale of the AI ​​big model and realize the construction of anthropomorphic thinking; S4. Fine-tune AI models and logic through manual supervision and data correction to align AI logic with human expectations, learn preferences from human (or manual) feedback to reduce bias and review models; S5. When AI neural networks are improving large models, they quantify data through the model's weights to reduce the model's storage cost and increase the model's computing speed. S6. The AI ​​data model is deployed to local devices and the main server through the network or hardware devices. The main server obtains the data model from each local server through the network for deep learning and improves the model. The improved model is fed back to the corresponding local server through the network to optimize the local model.

Images