Guest room management and control method and system based on guest room monitoring information

Abstract

The invention relates to the technical field of building environment control, in particular to a guest room management and control method and system based on guest room monitoring information, and the method comprises the steps: obtaining guest room three-dimensional space activity data, environment parameters and bed state information, and forming multi-source heterogeneous data; performing space-time alignment processing on the multi-source heterogeneous data; constructing an environment-equipment response model according to the data after time-space alignment to generate a control strategy; optimizing the control strategy, and feeding back in real time according to a control instruction; the regulation and control effect fed back in real time is calculated, and if the deviation of the fed-back regulation and control effect is larger than a threshold value, the model is updated online; the environment regulation and control precision is improved, the CO2 concentration control error is smaller than or equal to + / -35 ppm, the temperature fluctuation standard deviation is reduced to 0.4 DEG C, the energy efficiency optimization is remarkable, the invalid operation time of an air conditioner is shortened by 68%, the comprehensive energy consumption is reduced by 28.5%, the system robustness is enhanced, the basic control function is maintained when a sensor breaks down, and the equipment communication failure recovery time is smaller than or equal to 5.2 seconds.

Description

Technical Field

[0001] The present invention relates to the technical field of building environment control, and in particular to a guest room management and control method and system based on guest room monitoring information. Background Art

[0002] The existing guest room environment control systems have the following technical deficiencies: single data collection: for example, the infrared sensing-based guest room control system disclosed in CN109814488A relies only on human motion detection and lacks the coordinated analysis of environmental parameters and personnel behavior, resulting in frequent start and stop of the air conditioner; rigid control strategy: the threshold-triggered control method proposed in CN112230617A adopts fixed rule decision-making, which cannot adapt to the performance degradation of equipment and dynamic changes in the environment, and its energy efficiency decreases significantly after long-term use; weak exception handling: the hotel equipment monitoring system disclosed in CN113485245A directly switches to the preset mode when the sensor fails, lacks the ability of parameter self-learning, and leads to continuous deterioration of the control accuracy. Summary of the Invention

[0003] The purpose of the present invention is to provide a guest room management method and system based on guest room monitoring information, and to systematically solve the above technical problems by constructing a multi-source data spatiotemporal alignment mechanism, an equipment response efficiency matrix and a deep reinforcement learning decision model.

[0004] The present invention provides the following technical solutions: In a first aspect, a guest room management and control method based on guest room monitoring information is provided, comprising:

[0005] Step S1, obtaining guest room three-dimensional spatial activity data, environmental parameters and bed status information to form multi-source heterogeneous data;

[0006] Step S2, performing spatiotemporal alignment processing on the multi-source heterogeneous data;

[0007] Step S3, constructing an environment-device response model based on the spatiotemporally aligned data to generate a control strategy;

[0008] Step S4, optimizing the control strategy and providing real-time feedback based on the control instructions;

[0009] Step S5: Calculate the real-time feedback control effect. If the feedback control effect deviation is greater than a threshold, update the model online.

[0010] Preferably, in step S1, obtaining the activity data, environmental parameters and bed state information includes:

[0011] Deploy a three-dimensional millimeter-wave radar array to build a space activity perception network;

[0012] Configure environmental sensing units to form an environmental parameter detection matrix;

[0013] Bed status sensor is used to identify bed resting behavior.

[0014] Preferably, in step S2, the spatiotemporal alignment process includes:

[0015] Use dynamic time warping algorithm to align asynchronous data streams;

[0016] Compensate for the spatial position difference of sensors through Kalman filtering;

[0017] Perform cubic spline interpolation for missing data;

[0018] To construct the normalized feature tensor:

[0019] Preferably, in step S3, constructing an environment-device response model based on the spatiotemporally aligned data specifically includes:

[0020] S301, define the device impact weight matrix:

[0021]

[0022] in, It represents the efficiency coefficient of regulating the jth environmental parameter by the i-th type of equipment;

[0023] S302, calculate the current environmental parameter deviation vector ;

[0024] S303, by priority score = , sort the equipment control order.

[0025] Preferably, in step S3, generating a control strategy includes:

[0026] Define the objective function: ,in + =1;

[0027] The Q-learning algorithm is used to iteratively solve the optimal equipment start-stop combination.

[0028] Preferably, executing control instructions and providing real-time feedback adjustments include:

[0029] Calculate the actual control effect deviation every n seconds ;

[0030] like When it is greater than the threshold, the weight matrix is ​​triggered to be updated online.

[0031] Preferably, in step S2, a sliding window mechanism is used to cache sensor data of different sampling frequencies, a dynamic time warping algorithm is used to match the time axis of multiple sensors, and a cubic spline interpolation method is used to complete missing data frames.

[0032] Preferably, the method further includes an exception handling mechanism:

[0033] When it is detected that the sensor data is sampled out of the reasonable range for m consecutive times, the fault diagnosis process is triggered:

[0034] Switch to an alternate sensor data source;

[0035] Start the equipment self-test program and send the diagnostic code to the maintenance terminal;

[0036] Use historical data interpolation to maintain basic control functions;

[0037] When the device response times out:

[0038] Automatically downgrade to the default security mode;

[0039] Send hardware reset command through the control gateway

[0040] Record abnormal events to the cloud log library.

[0041] Preferably, in the feedback regulation, the threshold setting method includes:

[0042] Determine the basic threshold value based on the maximum allowable deviation of environmental parameters;

[0043] Determine the dynamic adjustment factor k to achieve day and night threshold fluctuations;

[0044] Determine the final threshold to ensure that larger control deviations are allowed at night to reduce equipment starts and stops.

[0045] Preferably, a multimodal sensor network in the guest room collects environmental data in real time, including:

[0046] The three-dimensional spatial coordinates and human movement trajectory data obtained by the millimeter-wave radar array;

[0047] Temperature and humidity, CO2 concentration, PM2.5, and TVOC values ​​collected by the multispectral sensor group;

[0048] The mattress's built-in flexible pressure sensor matrix detects bed status data.

[0049] The second aspect provides a guest room management and control system based on guest room monitoring information, including:

[0050] The data acquisition module is used to obtain guest room three-dimensional spatial activity data, environmental parameters and bed status information to form multi-source heterogeneous data;

[0051] A data processing module, configured to perform spatiotemporal alignment processing on the multi-source heterogeneous data;

[0052] A strategy generation module is used to construct an environment-device response model based on the spatiotemporal aligned data to generate a control strategy;

[0053] A strategy execution module is used to optimize the control strategy and provide real-time feedback based on the control instructions;

[0054] The feedback update module is used to calculate the control effect of real-time feedback, and if the deviation of the feedback control effect is greater than a threshold, the model is updated online.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] The system comprehensively assesses room activity (such as sleeping and leaving the room) and changes in air quality (such as elevated carbon dioxide levels) to automatically select the most appropriate combination of equipment to adjust the environment. For example, when someone leaves the room, instead of turning off the air conditioner, it first activates fresh air, minimizing energy waste and maintaining fresh air. By analyzing the actual performance of the air conditioner, fresh air system, and other equipment, the system prioritizes energy-efficient yet effective temperature control methods. For example, in summer, opening curtains to block sunlight reduces the burden on the air conditioner, rather than simply lowering the temperature. This saves energy while preventing the user from experiencing fluctuating temperatures. If a device malfunctions (such as an air conditioner malfunction), the system quickly detects and initiates a backup plan. For example, the fresh air system can be used to enhance ventilation to assist in temperature control, while also notifying maintenance to prevent the room environment from becoming uncontrolled. The system automatically learns and optimizes based on the actual performance of each adjustment. For example, if a particular air conditioner's cooling capacity slows, the system automatically adjusts its control strategy to prevent declining performance due to aging. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0058] The technical solutions in the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention.

[0059] The present invention provides a guest room management and control method and system based on guest room monitoring information.

[0060] refer to Figure 1 , the present invention provides the following technical solution: a guest room management and control method based on guest room monitoring information, comprising step S1, obtaining guest room three-dimensional spatial activity data, environmental parameters and bed status information to form multi-source heterogeneous data;

[0061] The acquisition of the activity data, environmental parameters and bed state information includes: deploying a three-dimensional millimeter wave radar array to build a spatial activity perception network; configuring environmental sensing units to form an environmental parameter detection matrix; and using a bed state sensor to identify bed rest behavior.

[0062] Step S2: Perform spatiotemporal alignment on the multi-source heterogeneous data. A sliding window mechanism is used to cache sensor data with different sampling frequencies. A dynamic time warping algorithm is used to align the time axes of multiple sensors. Missing data frames are then supplemented using cubic spline interpolation. This spatiotemporal alignment includes aligning asynchronous data streams using a dynamic time warping algorithm, compensating for sensor spatial disparity through Kalman filtering, and constructing a standardized feature tensor using cubic spline interpolation on missing data.

[0063] Step S3, constructing an environment-device response model based on the spatiotemporally aligned data to generate a control strategy;

[0064] The environment-device response model is constructed based on the spatiotemporal aligned data, including:

[0065] S301, define the device impact weight matrix:

[0066]

[0067] in, It represents the efficiency coefficient of regulating the jth environmental parameter by the i-th type of equipment;

[0068] S302, calculate the current environmental parameter deviation vector ;

[0069] S303, by priority score = , sort the equipment control order.

[0070] Generate a control strategy, including: defining the objective function: ,in + =1; Q-learning algorithm is used to iteratively solve the optimal equipment start-stop combination. Step S4, optimize the control strategy and provide real-time feedback according to the control instruction; Step S5, calculate the real-time feedback control effect. If the feedback control effect deviation is greater than the threshold, the model is updated online. Execute the control instruction and provide real-time feedback adjustment, including: calculating the actual control effect deviation every n seconds ;like When it is greater than the threshold, the weight matrix is ​​triggered to be updated online.

[0071] In one embodiment, the multimodal sensor network in the guest room collects environmental data in real time, including: three-dimensional spatial coordinates and human movement trajectory data obtained by the millimeter-wave radar array; temperature and humidity, CO2 concentration, PM2.5 and TVOC values ​​collected by the multispectral sensor group; and bed status data detected by the flexible pressure sensor matrix built into the mattress.

[0072] For example, millimeter-wave radar detected a person moving from a bed area (2.3m, 1.5m) to a bathroom (0.5m, 3.2m) at a speed greater than 0.5m / s. The flexible sensor detected a decrease in the standard deviation of mattress pressure from 85Pa to 8Pa (threshold σ = 10Pa). The CO2 sensor reading rose to 900ppm (target value 650ppm). 10Hz radar data was aligned with 1Hz environmental data (DTW path error < 30ms), and missing pressure sensor frames were supplemented (cubic spline interpolation).

[0073] decision making:

[0074] Calculate the deviation vector ;

[0075] Call weight matrix =0.75, generating strategy;

[0076] Start the fresh air system (50% power), with a priority score of 357.5; maintain the air conditioner in energy-saving mode, with a score of 126.

[0077] Execution Feedback:

[0078] After 10 minutes, CO2 dropped to 700ppm, the actual deviation ;

[0079] renew =0.75×(150 / 250)=0.45.

[0080] Example of the above score calculation: Fresh air system score:

[0081] S1=0.75×250 +0.85×0.2×1000=187.5+170=357.5.

[0082] Air conditioning system score S2=0.7×1.8×100=126.

[0083] The Q-learning model traverses possible action combinations and selects the optimal strategy.

[0084] In this example, the ineffective operation time of the air conditioner was reduced by 72% (actual measurement compared with traditional temperature control); the CO2 removal efficiency was increased by 40%, and the system energy consumption was reduced by 1.2kWh / day.

[0085] In one embodiment, the above method also includes an exception handling mechanism: when it is detected that the sensor data is sampled out of a reasonable range for m consecutive times, the fault diagnosis process is triggered: switching to a backup sensor data source; starting the equipment self-test program and sending a diagnostic code to the maintenance terminal; using historical data interpolation to maintain basic control functions; when the device response times out: downgrading to a preset safety mode; sending a hardware reset instruction through the control gateway to record the abnormal event to the cloud log library.

[0086] For example, fault detection and policy: If the edge node fails to receive air conditioner feedback for three consecutive times (>5s), and the Modbus CRC error rate exceeds 10%, the air conditioner is marked as faulty and the backup plan is activated. The fresh air system's ventilation frequency is increased to 8 minutes every 10 minutes, and the power windows are opened 15° (when the outdoor temperature is suitable). Self-repair: A hardware reset command (low level for 500ms) is sent, and the weight is reduced by 20% after recalibration.

[0087] The technical effect achieved in this example: temperature fluctuation is controlled within ±1.2°C (original setting ±0.5°C), and the self-repair success rate is 82% (30 simulation tests).

[0088] Among them, command issuance: the control gateway sends commands through the Modbus RTU protocol, and the response delay is <200ms.

[0089] In one embodiment, in feedback regulation, the threshold setting method includes: determining a basic threshold based on the maximum allowable deviation of environmental parameters; determining a dynamic adjustment factor k to achieve day and night threshold fluctuations; and determining a final threshold to ensure that a larger control deviation is allowed at night to reduce equipment start-up and shutdown.

[0090] In this embodiment, a sinusoidal adjustment factor is introduced to simulate the changes in the sensitivity of the human body's biological clock to the environment. Relaxing the threshold at night can reduce the frequency of device operation, and the actual measurement reduces the noise by 6dB.

[0091] The guest room management and control system based on guest room monitoring information includes a data acquisition module for acquiring three-dimensional spatial activity data, environmental parameters and bed status information of the guest room to form multi-source heterogeneous data; a data processing module for performing spatiotemporal alignment processing on the multi-source heterogeneous data; a strategy generation module for constructing an environment-equipment response model based on the spatiotemporal aligned data to generate a control strategy; a strategy execution module for optimizing the control strategy and providing real-time feedback according to control instructions; and a feedback update module for calculating the control effect of the real-time feedback. If the deviation of the feedback control effect is greater than a threshold, the model is updated online.

[0092] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same.

Claims

1. A guest room management and control method based on guest room monitoring information, characterized in that: The steps include: Step S1: Acquire guest room three-dimensional spatial activity data, environmental parameters, and bed status information to form multi-source heterogeneous data; Step S2, performing spatiotemporal alignment processing on the multi-source heterogeneous data; Step S3, constructing an environment-device response model based on the spatiotemporally aligned data to generate a control strategy; Step S4, optimizing the control strategy and providing real-time feedback based on the control instructions; Step S5: Calculate the real-time feedback control effect. If the feedback control effect deviation is greater than a threshold, update the model online.

2. The guest room management and control method based on guest room monitoring information according to claim 1, characterized in that: In step S1, the acquisition of the activity data, environmental parameters and bed state information includes: Deploy a three-dimensional millimeter-wave radar array to build a space activity perception network; Configure environmental sensing units to form an environmental parameter detection matrix; Bed status sensor is used to identify bed resting behavior.

3. The guest room management and control method based on guest room monitoring information according to claim 1, characterized in that: In step S2, the spatiotemporal alignment process includes: Use dynamic time warping algorithm to align asynchronous data streams; Compensate for the spatial position difference of sensors through Kalman filtering; Perform cubic spline interpolation for missing data; To construct the normalized feature tensor.

4. The guest room management and control method based on guest room monitoring information according to claim 3, characterized in that: In step S3, an environment-device response model is constructed based on the spatiotemporally aligned data, specifically including: S301, define the device impact weight matrix: in, It represents the efficiency coefficient of regulating the jth environmental parameter by the i-th type of equipment; S302, calculate the current environmental parameter deviation vector ; S303, by priority score = , sort the equipment control order.

5. The guest room management and control method based on guest room monitoring information according to claim 4, characterized in that: In step S3, a control strategy is generated, including: Define the objective function: ,in + =1; The Q-learning algorithm is used to iteratively solve the optimal equipment start-stop combination.

6. The guest room management and control method based on guest room monitoring information according to claim 1, characterized in that: Execute control instructions and provide real-time feedback adjustments, including: Calculate the actual control effect deviation every n seconds ; like When it is greater than the threshold, it triggers the online update of the weight matrix.

7. The guest room management and control method based on guest room monitoring information according to claim 1, characterized in that: In step S2, a sliding window mechanism is used to cache sensor data of different sampling frequencies, a dynamic time warping algorithm is used to match the time axis of multiple sensors, and cubic spline interpolation is used to complete the missing data frames.

8. The guest room management and control method based on guest room monitoring information according to claim 1, characterized in that: It also includes exception handling mechanisms: When it is detected that the sensor data is sampled out of the reasonable range for m consecutive times, the fault diagnosis process is triggered: Switch to an alternate sensor data source; Start the equipment self-test program and send the diagnostic code to the maintenance terminal; Use historical data interpolation to maintain basic control functions; When the device response times out: Automatically downgrade to the default security mode; Send hardware reset command through the control gateway Record abnormal events to the cloud log library.

9. The guest room management and control method based on guest room monitoring information according to claim 6, characterized in that: In feedback regulation, the threshold setting method includes: Determine the basic threshold value based on the maximum allowable deviation of environmental parameters; Determine the dynamic adjustment factor k to achieve day and night threshold fluctuations; Determine the final threshold to ensure that larger control deviations are allowed at night to reduce equipment starts and stops.

10. The guest room management and control system based on guest room monitoring information is characterized by: include: The data acquisition module is used to obtain guest room three-dimensional spatial activity data, environmental parameters and bed status information to form multi-source heterogeneous data; A data processing module, configured to perform spatiotemporal alignment processing on the multi-source heterogeneous data; A strategy generation module is used to construct an environment-device response model based on the spatiotemporal aligned data to generate a control strategy; A strategy execution module is used to optimize the control strategy and provide real-time feedback based on the control instructions; The feedback update module is used to calculate the control effect of real-time feedback, and if the deviation of the feedback control effect is greater than a threshold, the model is updated online.

Images