Intelligent control method and system for dynamic shielding of guest room

Abstract

The invention discloses an intelligent control method and system for dynamic shielding of a guest room, and belongs to the technical field of intelligent control. According to the method, distance data between personnel and equipment and environmental acoustic characteristic data are collected through a multi-modal sensor, a dynamic evaluation model is constructed, and the model comprises a linear fusion structure and an integrated architecture with graph structure modeling and uncertainty reasoning capabilities and is used for outputting a privacy risk evaluation value. And judging a privacy risk level according to the evaluation value, triggering a corresponding optical, acoustic or physical shielding response, and generating a shielding control instruction. The instruction is issued to compatible shielding execution equipment through a protocol adaptation mechanism to realize dynamic shielding control; and the system automatically enters a low-power-consumption mode when the low-risk condition is met. According to the invention, privacy level intelligent discrimination and multi-level shielding response control based on scene perception are realized, the adaptability, the control precision and the energy efficiency level of the system are improved, and the method is suitable for scenes with privacy requirements, such as intelligent guest rooms and conference rooms.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control technology, and in particular to a guest room dynamic shielding intelligent control method and system thereof. Background Art

[0002] With the development of smart home technology, particularly in smart hotels and guest room management systems, privacy protection within guest rooms has become a growing concern. Traditional guest room privacy protection methods (such as curtains and electrochromic glass) often rely on simple timed or manual controls, lacking intelligent sensing and response to environmental changes and occupant behavior. Existing control methods typically rely on preset thresholds for privacy control, such as using distance sensors to trigger curtain closing or adjusting glass transmittance at preset times. While this fixed-mode control approach can meet basic privacy requirements in certain scenarios, it often fails to effectively address evolving privacy needs in dynamic environments.

[0003] Existing technologies suffer from the following major issues: Static triggering mechanisms: Existing shielding control methods mostly rely on fixed trigger conditions (such as when a person's proximity or ambient acoustic characteristics reach a fixed threshold). These methods lack real-time perception and response to changes in the environment and person behavior. Consequently, shielding effectiveness may be insufficient in some cases and fail to meet actual privacy requirements. Low inter-device collaboration: Communication delays between existing shielding devices and perception modules often result in system response times exceeding ideal limits. Specifically, when a perception device detects a person approaching or a change in the environment, the device response delay can reach one second or longer, preventing the shielding device from executing actions in a timely manner, compromising privacy protection effectiveness. Energy efficiency management issues: Most existing shielding devices only support a single mode of operation, either fully on or fully off, lacking intelligent energy efficiency management. The system struggles to automatically switch to the appropriate operating mode based on actual needs, resulting in energy waste and an inability to achieve low-power operation in low-risk situations. Inadequate privacy risk assessment: Existing technologies are relatively simplistic in their privacy risk assessments, often based on a single data source and lacking the effective integration of multimodal data such as person behavior and ambient noise. Therefore, the assessment results are often delayed and cannot reflect the actual privacy risks in the current environment in a timely manner.

[0004] Therefore, the key to solving this problem is to provide a dynamic, real-time, and intelligent shielding control method that can accurately assess privacy risks based on environmental changes and personnel behavior in the guest room and flexibly control shielding equipment. Summary of the Invention

[0005] The purpose of the present invention is to provide a method and system for intelligent control of dynamic shielding in guest rooms, which solves the limitations of existing shielding control methods, such as static and slow response and lack of real-time privacy risk assessment. By fusing multimodal data and adopting a dynamic assessment model, privacy risks are assessed in real time and accurately based on environmental changes and personnel behavior, and the response level of shielding equipment is automatically adjusted, thus achieving efficient, intelligent and energy-saving privacy protection control.

[0006] The present invention achieves the above-mentioned purpose through the following technical solutions:

[0007] In one aspect, the present invention provides a method for intelligently controlling dynamic shielding of guest rooms, comprising the following steps:

[0008] Collecting multimodal data related to human behavior or environmental status in real time through a multimodal sensor, wherein the multimodal data includes distance data between the human and the device and environmental acoustic characteristic data;

[0009] Based on the multimodal data, a dynamic assessment model is constructed. The dynamic assessment model is used to fuse the multimodal data to perform privacy risk assessment. The dynamic assessment model structure includes a linear fusion structure and an integrated structure with graph structure modeling and uncertainty reasoning capabilities. A neural network is used to jointly model proximity and acoustic features, and a privacy risk assessment value is output within each assessment cycle.

[0010] According to the privacy risk assessment value, a privacy risk level is determined, and a corresponding shielding response level is initiated, wherein when the assessment value is greater than or equal to a preset first threshold, it is determined to be a high privacy risk and a high-level shielding response is initiated; when the assessment value is lower than a second threshold, it is determined to be a low privacy risk and a low-level shielding response is initiated;

[0011] generating corresponding control instructions according to the shielding response level, wherein the control instructions are used to control the shielding device to perform optical, acoustic or physical shielding operations;

[0012] The control instruction is sent to a compatible shielding execution device through a protocol adaptation mechanism to execute a corresponding shielding operation;

[0013] When the system detects that the privacy risk assessment value is lower than the second threshold and there is no obvious environmental change or human approach event, it automatically switches to low power mode. When it detects an increase in the privacy risk assessment value, environmental changes or human approach events, it exits the low power state.

[0014] A further improvement of the present invention is that the multimodal data includes distance data between the person and the device collected in real time by a millimeter-wave radar array sensor, and environmental acoustic feature data collected in real time by an acoustic sensor, wherein the environmental acoustic feature data includes spectral entropy values and voiceprint features;

[0015] The acoustic sensor includes a sound source localization module and a keyword recognition module. The sound source localization accuracy is better than ±3°, and the keyword recognition frequency band covers 300Hz to 3400Hz, which is used to assist in determining the degree of environmental changes and the activity status of people.

[0016] The millimeter-wave radar array is installed on the guest room ceiling in a regular hexagonal layout, with an array spacing of 1.2m, a downward tilt angle of 15°, a detection range of 0.5-5m, a data update frequency of no less than 100Hz, and a positioning accuracy better than ±5cm.

[0017] A further improvement of the present invention is that the dynamic evaluation model is based on the proximity normalization value and Environmental Change Index Constructing a privacy risk assessment , the calculation formula is:

[0018] ;

[0019] The proximity normalized value The calculation method is:

[0020] ;

[0021] in: The real-time distance measurement value between the user and the device. 0.5 represents the preset minimum privacy safety radius.

[0022] The environmental change index It is obtained by normalizing the acoustic spectrum entropy value, and the formula is:

[0023] ;

[0024] in: is the spectrum entropy value of the current sound source frame; is the historical maximum entropy reference value, used for standardization, The value range is ;

[0025] The coefficient It is optimized based on 200 sets of typical guest room scene training samples, and is used to improve the privacy assessment accuracy of the evaluation model in actual dynamic environments.

[0026] A further improvement of the present invention is that the dynamic assessment model is a structured uncertainty learning model that uses graph structure sensor feature propagation and multiple forward predictions to generate privacy risk assessment values. , and output the prediction uncertainty index;

[0027] The dynamic evaluation model includes the following steps:

[0028] Construct a graph structure based on the spatial deployment relationship of sensors, where each node represents the shielding-related sensor, and the edge connection relationship is determined by the distance between sensors and the layout topology;

[0029] Normalize the proximity value and Environmental Change Index as node input features;

[0030] The above features are input into the graph neural network structure for feature propagation, and the random inactivation mechanism is used to forward reasoning to obtain multiple privacy risk prediction values , and calculate its mean and uncertainty index as follows:

[0031] ;

[0032] in: For the model The privacy risk prediction value generated in the forward reasoning; It is an indicator of uncertainty in the risk assessment results; is the total number of forward propagations set; is the round index for forward prediction.

[0033] A further improvement of the present invention is that the privacy risk assessment value R is used to classify privacy risk levels, and the levels include:

[0034] When R ≥ 0.7, it is determined to be a high privacy risk level, triggering an optical shielding response, controlling the electrochromic glass to adjust the transmittance to 5% within 200ms, and simultaneously turning off the auxiliary lighting in the guest room;

[0035] When 0.6≤R<0.7, it is determined to be a medium privacy risk level, triggering the acoustic masking response. The white noise output sound pressure level is 65dB@1m, and the output frequency band is the current environment's main frequency ±200Hz. The frequency band is adjusted in real time through the frequency modulation system;

[0036] When 0.5≤R<0.6, it is determined to be a low privacy risk level, triggering a physical shielding response. The electric curtain is closed to 80%, which is executed by the PID closed-loop control module. The curtain closing accuracy is controlled within ±2cm. The PID parameters include the proportional coefficient Kp=1.2 and the integral time Ti=0.5 seconds.

[0037] After the shielding response is executed, the system monitors the current feedback of the corresponding device. If it detects that the execution current exceeds the preset threshold of 2A, the control module triggers the fault handling process, including cutting off the shielding power supply circuit and regenerating the shielding control instruction.

[0038] A further improvement of the present invention is that the ambient dominant frequency is determined by performing power spectral density analysis on the acoustic signal within the white noise region, extracting the main peak frequency as the current ambient dominant frequency. Based on this dominant frequency, the white noise output frequency band is dynamically set to ±200 Hz of the dominant frequency. The dynamic adjustment of the frequency band is achieved by controlling the frequency modulation module of the white noise generating device, and the frequency update cycle is synchronized with the privacy risk assessment cycle.

[0039] The frequency modulation module performs feedback detection on the acoustic output status after the white noise is output. If the detected signal-to-noise ratio is lower than 20dB, or the acoustic shielding device does not reach the set frequency band range within the specified time, the control system automatically terminates the current white noise output process according to the abnormal status and regenerates the frequency modulation instruction.

[0040] A further improvement of the present invention is that the control instruction is generated through a protocol adaptation mechanism, and the protocol adaptation mechanism supports the mutual conversion of the following six Internet of Things protocols, including MQTT, Modbus RTU, RS485, CoAP, OPC UA and Zigbee protocols, and the protocol conversion delay does not exceed 50ms; the protocol adaptation mechanism includes a backup protocol switching mechanism. When the default communication protocol fails to execute, it automatically switches to the backup protocol according to the preset priority order. The judgment conditions for the communication failure include a response timeout of more than 300ms or a communication verification error, and the core parameters of the control instruction are kept unchanged during the switching process, and only the communication format conversion is performed; after the backup protocol is successfully switched, the system continuously monitors the status of the default protocol channel. If it is detected that the communication has returned to normal, it automatically switches back to the default protocol within 10 seconds.

[0041] A further improvement of the present invention is that the triggering condition of the low-power mode is: when the privacy risk assessment value R is lower than 0.6 and the system does not detect significant environmental changes or human proximity events within a preset time period, it automatically switches to low-power mode; the environmental changes are comprehensively judged based on changes in acoustic spectrum characteristics and human proximity data, and the standby power consumption of the shielding system in the low-power mode is not higher than 5W.

[0042] A further improvement of the present invention is that the operating frequency of the dynamic assessment model is 100 Hz, and the system collects proximity data and environmental acoustic feature data at the same frequency, the proximity data including the real-time distance measurement value between the user and the shielding device, and the acoustic feature data including the spectrum entropy value of the current sound source;

[0043] The data is used as input to the dynamic assessment model to calculate the privacy risk assessment value in real time during each assessment cycle, and the shielding response level and corresponding control instructions are dynamically updated based on the privacy risk assessment value to achieve adaptive control of the shielding device to the current environmental status of the guest room.

[0044] In another aspect, the present invention provides a guest room dynamic shielding intelligent control system, which applies the guest room dynamic shielding intelligent control method described above, comprising:

[0045] A multimodal perception module, used to collect multimodal data related to human behavior or environmental status in real time. The multimodal data includes the distance between people and equipment and the acoustic characteristics of the environment. The multimodal perception module includes a millimeter-wave radar array and an environmental acoustic sensor array. The millimeter-wave radar array is installed on the guest room ceiling, with a detection accuracy better than ±5cm and an update frequency of no less than 100Hz.

[0046] A dynamic assessment module is configured to construct a dynamic assessment model based on the multimodal data, wherein the dynamic assessment model is configured to fuse the multimodal perception data to perform privacy risk assessment; the dynamic assessment model outputs a privacy risk assessment value based on a normalized proximity value and an acoustic change index;

[0047] a shielding control module, configured to determine a current shielding response level based on the privacy risk assessment value, including high risk, medium risk, and low risk levels, and to generate corresponding control instructions based on the level, the control instructions including execution parameters for optical shielding, acoustic shielding, and physical shielding;

[0048] A protocol adapter module, used to convert the control instructions into a communication protocol compatible with the shielding execution device, supporting MQTT, Modbus RTU, RS485, OPC UA, Zigbee and CoAP protocol conversion, with a protocol conversion delay of no more than 50ms, and equipped with a backup protocol switching and fallback mechanism;

[0049] A shielding execution module, including electrochromic glass, directional white noise equipment and electric curtain equipment, is used to execute corresponding shielding response operations;

[0050] The power consumption management module is used to automatically switch the system to low-power standby mode when the assessment result is continuously lower than the risk threshold and no human approach or environmental changes are detected. The trigger conditions include the privacy risk assessment value being lower than the set threshold and lasting for a period of time, and the shielding state remaining stable.

[0051] The present invention has the beneficial effect of constructing a multimodal fusion model for dynamic privacy risk assessment based on real-time multimodal data related to human behavior and environmental conditions, thereby achieving more accurate and real-time shielding response control. By fusing distance data with environmental acoustic feature data, the present invention effectively characterizes the relative positional relationship between the user and the device, as well as the environmental disturbance state. The constructed dynamic assessment model not only supports a linear fusion structure to meet conventional assessment requirements, but also incorporates a neural network integrated architecture with graph structure modeling and uncertainty reasoning capabilities to accurately model the complex interactions between multi-source data and identify privacy risk levels. Dynamically updated privacy risk assessment values are output within each assessment cycle. Based on this assessment value, the method adaptively determines the current privacy risk level and initiates a multi-level shielding response strategy, enabling hierarchical control of optical, acoustic, or physical shielding devices based on risk level. This overcomes the shortcomings of traditional shielding systems, such as fixed response levels, inaccurate regional coverage, and high energy consumption. Furthermore, through a protocol adaptation mechanism, the method seamlessly integrates control commands with various shielding devices, is compatible with mainstream IoT communication protocols, and ensures high response efficiency and communication reliability even when multiple devices are working together. In addition, when the system continuously detects a low-risk state and meets the static environmental characteristics, it can automatically switch to a low-power operation mode, significantly reducing standby energy consumption. It has strong energy-saving adaptability and environmental dynamic adaptability, and overall improves the intelligence level and operating efficiency of guest room privacy protection. It is suitable for the shielding control needs in a variety of smart hotels and privacy-sensitive spaces. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be derived from these drawings without inventive effort. Among them: Figure 1 is a flow chart of the method of the present invention; Figure 2 Schematic diagram of the structure of an embodiment of the present invention; Figure 3 This is the process intention of the dynamic shielding control method in an embodiment of the present invention; Figure 4 Schematic diagram of shielding execution and effect verification structure in an embodiment of the present invention; Figure 5 Schematic diagram of a multi-protocol fault-tolerant architecture of a shielding control system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0053] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of the present invention.

[0054] Example 1

[0055] like Figure 1-Figure 3 FIG. 1 is an embodiment of the present invention, which provides a method for intelligently controlling dynamic shielding of guest rooms, including:

[0056] Step 1: Using a multimodal sensor to collect multimodal data related to human behavior or environmental status in real time, wherein the multimodal data includes distance data between the human and the device and environmental acoustic feature data;

[0057] In one feasible implementation, the multimodal data includes distance data between a person and a device collected in real time by a millimeter-wave radar array sensor, and environmental acoustic feature data collected in real time by an acoustic sensor, wherein the environmental acoustic feature data includes spectral entropy values and voiceprint features;

[0058] The acoustic sensor includes a sound source localization module and a keyword recognition module. The sound source localization accuracy is better than ±3°, and the keyword recognition frequency band covers 300Hz to 3400Hz, which is used to assist in determining the degree of environmental changes and the activity status of people.

[0059] The millimeter-wave radar array is installed on the guest room ceiling in a regular hexagonal layout, with an array spacing of 1.2m, a downward tilt angle of 15°, a detection range of 0.5-5m, a data update frequency of no less than 100Hz, and a positioning accuracy better than ±5cm.

[0060] To improve assessment accuracy and response speed, this embodiment denoises and preprocesses distance and acoustic data during data collection, enhancing the system's sensitivity to environmental changes and human behavior. By optimizing sensor placement and data transmission stability, errors caused by interference are reduced, ensuring the timeliness and accuracy of multimodal data, providing more reliable input for subsequent dynamic assessment models.

[0061] Step 2: Based on the multimodal data, a dynamic assessment model is constructed. The dynamic assessment model is used to fuse the multimodal data to perform privacy risk assessment. The dynamic assessment model structure includes a linear fusion structure and an integrated structure with graph structure modeling and uncertainty reasoning capabilities. A neural network is used to jointly model proximity and acoustic features, and a privacy risk assessment value is output in each assessment cycle.

[0062] In one embodiment, the dynamic evaluation model is based on the proximity normalization value and Environmental Change Index Constructing a privacy risk assessment , the calculation formula is:

[0063] ;

[0064] The proximity normalized value The calculation method is:

[0065] ;

[0066] in: The real-time distance measurement value between the user and the device. 0.5 represents the preset minimum privacy safety radius.

[0067] The environmental change index It is obtained by normalizing the acoustic spectrum entropy value, and the formula is:

[0068] ;

[0069] in: is the spectrum entropy value of the current sound source frame; is the historical maximum entropy reference value, used for standardization, The value range is ;

[0070] The coefficient It is optimized based on 200 sets of typical guest room scene training samples, and is used to improve the privacy assessment accuracy of the evaluation model in actual dynamic environments.

[0071] In this embodiment, the dynamic assessment model is a structured uncertainty learning model that uses graph structure sensor feature propagation and multiple forward predictions to generate privacy risk assessment values. , and output the prediction uncertainty index;

[0072] The dynamic evaluation model includes the following steps:

[0073] Construct a graph structure based on the spatial deployment relationship of sensors, where each node represents the shielding-related sensor, and the edge connection relationship is determined by the distance between sensors and the layout topology;

[0074] Normalize the proximity value and Environmental Change Index as node input features;

[0075] The above features are input into the graph neural network structure for feature propagation, and the random inactivation mechanism is used to forward reasoning to obtain multiple privacy risk prediction values , and calculate its mean and uncertainty index as follows:

[0076] ;

[0077] in: For the model The privacy risk prediction value generated in the forward reasoning; It is an indicator of uncertainty in the risk assessment results; is the total number of forward propagations set; is the round index for forward prediction.

[0078] In this embodiment, through this joint modeling process, the system not only enables high-frequency privacy risk assessment based on real-time sensor data, but also dynamically evaluates uncertainty indicators based on historical prediction sequences, assisting in adaptive judgment of subsequent response levels. Compared to traditional single-assessment methods that rely on fixed thresholds, this implementation integrates dynamic calculation based on multimodal features with multiple rounds of forward reasoning, significantly improving the recognition sensitivity and response accuracy to complex environmental changes, and enhancing the model's adaptability and robustness in diverse scenarios.

[0079] Step 3: Determine the privacy risk level based on the privacy risk assessment value and initiate a corresponding shielding response level. When the assessment value is greater than or equal to a preset first threshold, it is determined to be a high privacy risk and a high-level shielding response is initiated; when the assessment value is lower than a second threshold, it is determined to be a low privacy risk and a low-level shielding response is initiated.

[0080] In one possible implementation, Figure 4 As shown, the privacy risk assessment value R is used to classify privacy risk levels, which include:

[0081] When R ≥ 0.7, it is determined to be a high privacy risk level, triggering an optical shielding response, controlling the electrochromic glass to adjust the transmittance to 5% within 200ms, and simultaneously turning off the auxiliary lighting in the guest room;

[0082] When 0.6≤R<0.7, it is determined to be a medium privacy risk level, triggering the acoustic masking response. The white noise output sound pressure level is 65dB@1m, and the output frequency band is the current environment's main frequency ±200Hz. The frequency band is adjusted in real time through the frequency modulation system;

[0083] When 0.5≤R<0.6, it is determined to be a low privacy risk level, triggering a physical shielding response. The electric curtain is closed to 80%, which is executed by the PID closed-loop control module. The curtain closing accuracy is controlled within ±2cm. The PID parameters include the proportional coefficient Kp=1.2 and the integral time Ti=0.5 seconds.

[0084] After the shielding response is executed, the system monitors the current feedback of the corresponding device. If it detects that the execution current exceeds the preset threshold of 2A, the control module triggers the fault handling process, including cutting off the shielding power supply circuit and regenerating the shielding control instruction.

[0085] Step 4: generating corresponding control instructions according to the shielding response level, wherein the control instructions are used to control the shielding device to perform optical, acoustic or physical shielding operations;

[0086] In one embodiment, the ambient main frequency is determined by performing power spectral density analysis on the acoustic signal within the white noise region to extract the main peak frequency as the current ambient main frequency. Based on this main frequency, the white noise output frequency band is dynamically set to ±200 Hz of the main frequency. The dynamic adjustment of the frequency band is achieved by controlling the frequency modulation module of the white noise generating device, and the frequency update cycle is synchronized with the privacy risk assessment cycle.

[0087] The frequency modulation module performs feedback detection on the acoustic output status after the white noise is output. If the detected signal-to-noise ratio is lower than 20dB, or the acoustic shielding device does not reach the set frequency band range within the specified time, the control system automatically terminates the current white noise output process according to the abnormal status and regenerates the frequency modulation instruction.

[0088] Step 5: Send the control instruction to a compatible shielding execution device through a protocol adaptation mechanism to execute the corresponding shielding operation;

[0089] In one embodiment, Figure 5 As shown, the control instructions are generated by a protocol adaptation mechanism, which supports the conversion of at least six IoT protocols, including MQTT, Modbus RTU, RS485, CoAP, OPC UA and Zigbee protocols, and the protocol conversion delay does not exceed 50ms, as shown in Table 1:

[0090] Table 1

[0091] The protocol adaptation mechanism includes a backup protocol switching mechanism. When the default communication protocol fails to execute, it automatically switches to the backup protocol according to the preset priority order. The judgment conditions for the communication failure include a response timeout exceeding 300ms or a communication verification error. During the switching process, the core parameters of the control instructions remain unchanged, and only the communication format conversion is performed. After the backup protocol is successfully switched, the system continuously monitors the status of the default protocol channel. If it is detected that the communication has returned to normal, it automatically switches back to the default protocol within 10 seconds.

[0092] Step 6: When the system detects that the privacy risk assessment value is lower than the preset threshold for a long time and there is no obvious environmental change or human approach event, it automatically switches to low-power operation mode and exits the low-power state when the set conditions are restored.

[0093] In one embodiment, the triggering condition of the low power consumption mode is: when the privacy risk assessment value R is lower than 0.6 and the system does not detect significant environmental changes or human proximity events within a preset time period, it automatically switches to low power consumption mode; the environmental changes are comprehensively judged based on the changes in acoustic spectrum characteristics and human proximity data, and the standby power consumption of the shielding system in the low power consumption mode is not higher than 5W.

[0094] In one embodiment, the dynamic assessment model operates at a frequency of 100 Hz, and the system collects proximity data and environmental acoustic feature data at the same frequency. The proximity data includes a real-time distance measurement between the user and the shielding device, and the acoustic feature data includes a spectrum entropy value of the current sound source.

[0095] The data is used as input to the dynamic assessment model to calculate the privacy risk assessment value in real time during each assessment cycle, and the shielding response level and corresponding control instructions are dynamically updated based on the privacy risk assessment value to achieve adaptive control of the shielding device to the current environmental status of the guest room.

[0096] Another embodiment of the present invention provides a guest room dynamic shielding intelligent control system, which applies the guest room dynamic shielding intelligent control method described above, including:

[0097] A multimodal perception module is used to collect multimodal data related to human behavior or environmental status in real time. The multimodal data includes the distance between people and equipment and the environmental acoustic characteristics. The multimodal perception module includes a millimeter-wave radar array and an environmental acoustic sensor array. The millimeter-wave radar array is installed on the guest room ceiling, with a measurement accuracy better than ±5cm and an update frequency of no less than 100Hz.

[0098] A dynamic assessment module is configured to construct a dynamic assessment model based on the multimodal data, wherein the dynamic assessment model is configured to fuse the multimodal perception data to perform privacy risk assessment; the dynamic assessment model outputs a privacy risk assessment value based on a normalized proximity value and an acoustic change index;

[0099] a shielding control module, configured to determine a current shielding response level based on the privacy risk assessment value, including high risk, medium risk, and low risk levels, and to generate corresponding control instructions based on the level, the control instructions including execution parameters for optical shielding, acoustic shielding, and physical shielding;

[0100] A protocol adapter module, used to convert the control instructions into a communication protocol compatible with the shielding execution device, supporting MQTT, Modbus RTU, RS485, OPC UA, Zigbee and CoAP protocol conversion, with a protocol conversion delay of no more than 50ms, and equipped with a backup protocol switching and fallback mechanism;

[0101] A shielding execution module, including electrochromic glass, directional white noise equipment and electric curtain equipment, is used to execute corresponding shielding response operations;

[0102] The power consumption management module is used to automatically switch the system to low-power standby mode when the assessment result is continuously lower than the risk threshold and no human approach or environmental changes are detected. The trigger conditions include the privacy risk assessment value being lower than the set threshold and lasting for a period of time, and the shielding state remaining stable.

[0103] Example 2

[0104] To validate the performance of the proposed intelligent control method and system for dynamic guest room shading, a standard test scenario was constructed for experimental verification. The test environment was a standard guest room (approximately 25 square meters, 3 meters high). A millimeter-wave radar array sensor (TI IWR6843ISK, with a detection range of 0.5 to 5 meters) and a ring-shaped array acoustic sensor (ReSpeaker Core v2) were deployed within the environment to monitor occupant proximity and ambient acoustic characteristics in real time. The system operated at a 100Hz frequency, with an evaluation period of 200ms.

[0105] Under the above configuration, the dynamic assessment model described in the present invention is used to integrate distance standardization parameters and environmental change indicators to calculate the privacy risk assessment value, and drive a three-level response mechanism based on the risk level:

[0106] When R≥0.7, the system activates optical shielding within 200ms to control the transmittance of the electrochromic glass to below 5%;

[0107] When 0.6≤R<0.7, the acoustic shielding and physical shielding linkage is triggered, where the white noise generator outputs a sound pressure of 65dB@1m and the curtains are closed to 50%;

[0108] When 0.5≤R<0.6, basic physical shielding is performed (curtains are closed to 80%), and the system is switched to low-power operation mode (power consumption ≤5W).

[0109] In addition, the system uses a six-protocol conversion matrix (MQTT, Modbus RTU, RS485, CoAP, OPC UA, Zigbee) to achieve protocol adaptation and conversion of shielding control instructions, and automatically activates the alternative protocol switching mechanism when the channel communication delay exceeds 300 ms or verification is abnormal to ensure uninterrupted command execution.

[0110] The experiment performed 200 shielding response operations. Statistics show that the average system shielding response delay was 180ms, the control command conflict rate was 2.8%, and the protocol command switching success rate reached 98.7%. Furthermore, compared to traditional shielding schemes based on static threshold triggering, the proposed method can reduce overall energy consumption by approximately 42.3% in actual dynamic environments, significantly improving the system's energy efficiency and response consistency in complex scenarios. See Table 2 below:

[0111] Table 2

[0112] The above experimental results show that the shielding method proposed in the present invention has good comprehensive performance in terms of shielding timeliness, multimodal control coordination and protocol compatibility, has feasibility for practical engineering application, and is suitable for highly sensitive privacy protection scenarios such as hotel rooms and medical monitoring.

[0113] In summary, this invention provides a method and system for intelligent control of dynamic guest room shading, overcoming the bottlenecks of existing shading technologies, such as fixed response, delayed recognition, and high energy consumption. By constructing a graph-based uncertainty modeling and assessment model, this solution effectively improves the accuracy of identifying privacy risks and the adaptability of shading strategies, enabling precise responsive control under dynamic environmental changes. This method offers significant advantages, including flexible structural deployment, strong command compatibility, and high operational energy efficiency, and possesses promising engineering implementation prospects and market application potential.

[0114] In the description of this specification, the reference terms "one embodiment," "some embodiments," "example," "specific example," or "some examples" mean that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the specific features, structures, materials, or characteristics described may be combined in any appropriate manner in any one or more embodiments or examples. In addition, those skilled in the art may combine and combine different embodiments or examples described in this specification, as well as features of different embodiments or examples, unless they are mutually inconsistent.

[0115] Any process or method description in the flowchart or otherwise described herein can be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process. The scope of the preferred embodiments of the present application includes additional implementations in which the functions may be performed in a different order than shown or discussed, including in a substantially simultaneous manner or in a reverse order depending on the functions involved.

[0116] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any person skilled in the art can easily conceive of various modifications or substitutions within the technical scope disclosed in this application, and such modifications or substitutions should be included within the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A guest room dynamic shielding intelligent control method, characterized in that: The following steps are involved: Collecting multimodal data related to human behavior or environmental status in real time through a multimodal sensor, wherein the multimodal data includes distance data between the human and the device and environmental acoustic characteristic data; Based on the multimodal data, a dynamic assessment model is constructed. The dynamic assessment model is used to fuse the multimodal data to perform privacy risk assessment. The dynamic assessment model structure includes a linear fusion structure and an integrated structure with graph structure modeling and uncertainty reasoning capabilities. A neural network is used to jointly model proximity and acoustic features, and a privacy risk assessment value is output within each assessment cycle. According to the privacy risk assessment value, a privacy risk level is determined, and a corresponding shielding response level is initiated, wherein when the privacy risk assessment value is greater than or equal to a preset first threshold, it is determined to be a high privacy risk and a high-level shielding response is initiated; when the privacy risk assessment value is lower than a second threshold, it is determined to be a low privacy risk and a low-level shielding response is initiated; generating corresponding control instructions according to the shielding response level, wherein the control instructions are used to control the shielding device to perform optical, acoustic or physical shielding operations; The control instruction is sent to a compatible shielding execution device through a protocol adaptation mechanism to execute a corresponding shielding operation; When the system detects that the privacy risk assessment value is lower than the second threshold and there is no obvious environmental change or human approach event, it automatically switches to low power mode. When it detects an increase in the privacy risk assessment value, environmental changes or human approach events, it exits the low power state.

2. A guest room dynamic shielding intelligent control method according to claim 1, characterized in that: The multimodal data includes distance data between people and equipment collected in real time by a millimeter-wave radar array sensor, and environmental acoustic feature data collected in real time by an acoustic sensor, the environmental acoustic feature data including spectral entropy values and voiceprint features; The acoustic sensor includes a sound source localization module and a keyword recognition module. The sound source localization accuracy is better than ±3°, and the keyword recognition frequency band covers 300Hz to 3400Hz, which is used to assist in determining the degree of environmental changes and the activity status of people. The millimeter-wave radar array is installed on the guest room ceiling in a regular hexagonal layout, with an array spacing of 1.2m, a downward tilt angle of 15°, a detection range of 0.5-5m, a data update frequency of no less than 100Hz, and a positioning accuracy better than ±5cm.

3. The intelligent control method for dynamic shielding of guest rooms according to claim 1, characterized in that: The dynamic evaluation model is based on the normalized value of proximity and Environmental Change Index Constructing a privacy risk assessment , the calculation formula is: ； The proximity normalized value The calculation method is: ； in: The real-time distance measurement value between the user and the device. 0.5 represents the preset minimum privacy safety radius. The environmental change index It is obtained by normalizing the acoustic spectrum entropy value, and the formula is: ； in: is the spectrum entropy value of the current sound source frame; is the historical maximum entropy reference value, The value range is ; The coefficient The optimization is based on 200 sets of typical guest room scene training samples.

4. A guest room dynamic shielding intelligent control method according to claim 3, characterized in that: The dynamic assessment model is a structured uncertainty learning model that uses graph-structured sensor feature propagation and multiple forward predictions to generate privacy risk assessment values. , and output the prediction uncertainty index; The dynamic evaluation model includes the following steps: Construct a graph structure based on the spatial deployment relationship of sensors, where each node represents the shielding-related sensor, and the edge connection relationship is determined by the distance between sensors and the layout topology; Normalize the proximity value and Environmental Change Index as node input features; The above features are input into the graph neural network structure for feature propagation, and the random inactivation mechanism is used to Forward reasoning, we get Privacy risk prediction value , and calculate its mean and uncertainty index as follows: ； in: For the model The privacy risk prediction value generated in the forward reasoning; It is an indicator of uncertainty in the risk assessment results; is the total number of forward propagations set; is the round index for forward prediction.

5. The intelligent control method for dynamic shielding of guest rooms according to claim 4, characterized in that: The privacy risk assessment value Used to classify privacy risk levels, including: When R ≥ 0.7, it is determined to be a high privacy risk level, triggering an optical shielding response, controlling the electrochromic glass to adjust the transmittance to 5% within 200ms, and simultaneously turning off the auxiliary lighting in the guest room; When 0.6≤R<0.7, it is determined to be a medium privacy risk level, triggering the acoustic masking response. The white noise output sound pressure level is 65dB@1m, and the output frequency band is the current environment's main frequency ±200Hz. The frequency band is adjusted in real time through the frequency modulation system; When 0.5≤R<0.6, it is determined to be a low privacy risk level, triggering a physical shielding response. The electric curtain is closed to 80%, which is executed by the PID closed-loop control module. The curtain closing accuracy is controlled within ±2cm. The PID parameters include the proportional coefficient =1.2, integration time Ti=0.5 seconds; After the shielding response is executed, the system monitors the current feedback of the corresponding device. If it detects that the execution current exceeds the preset threshold of 2A, the control module triggers the fault handling process, including cutting off the shielding power supply circuit and regenerating the shielding control instruction.

6. A guest room dynamic shielding intelligent control method according to claim 5, characterized in that: The ambient main frequency is determined by performing power spectral density analysis on the acoustic signal within the white noise region, extracting the main peak frequency as the current ambient main frequency. Based on this main frequency, the white noise output frequency band is dynamically set to ±200 Hz of the main frequency. The dynamic adjustment of the frequency band is achieved by controlling the frequency modulation module of the white noise generating device, and the frequency update cycle is synchronized with the privacy risk assessment cycle. The frequency modulation module performs feedback detection on the acoustic output status after the white noise is output. If the detected signal-to-noise ratio is lower than 20dB, or the acoustic shielding device does not reach the set frequency band range within the specified time, the control system automatically terminates the current white noise output process according to the abnormal status and regenerates the frequency modulation instruction.

7. The intelligent control method for dynamic shielding of guest rooms according to claim 5, characterized in that: The control instructions are generated through a protocol adaptation mechanism that supports interconversion between the following six IoT protocols: MQTT, Modbus RTU, RS485, CoAP, OPC UA, and Zigbee, with a protocol conversion delay of no more than 50ms. The protocol adaptation mechanism includes a backup protocol switching mechanism that automatically switches to a backup protocol in a preset priority order when the default communication protocol fails to execute. The judgment conditions for communication failure include a response timeout exceeding 300ms or a communication verification error. During the switching process, the core parameters of the control instructions remain unchanged, and only the communication format conversion is performed. After the backup protocol is successfully switched, the system continuously monitors the status of the default protocol channel. If it detects that communication has returned to normal, it will automatically switch back to the default protocol within 10 seconds.

8. The intelligent control method for dynamic shielding of guest rooms according to claim 1, characterized in that: The triggering conditions for the low-power mode are: when the privacy risk assessment value R is lower than 0.6 and the system does not detect significant environmental changes or human proximity events within a preset time, it automatically switches to low-power mode; the environmental changes are comprehensively judged based on changes in acoustic spectrum characteristics and human proximity data, and the standby power consumption of the shielding system in the low-power mode is no more than 5W.

9. The intelligent control method for dynamic shielding of guest rooms according to claim 1, characterized in that: The dynamic assessment model operates at a frequency of 100 Hz. The system collects proximity data and environmental acoustic feature data at the same frequency. The proximity data includes the real-time distance measurement between the user and the shielding device, and the acoustic feature data includes the spectral entropy value of the current sound source. The data is used as input to the dynamic assessment model to calculate the privacy risk assessment value in real time during each assessment cycle, and the shielding response level and the corresponding control instructions are dynamically updated according to the privacy risk assessment value.

10. A dynamic shielding intelligent control system for guest rooms, characterized in that: The method for intelligently controlling dynamic shielding of guest rooms according to any one of claims 1 to 9 is applied, comprising: A multimodal perception module, used to collect multimodal data related to human behavior or environmental status in real time. The multimodal data includes the distance between people and equipment and the acoustic characteristics of the environment. The multimodal perception module includes a millimeter-wave radar array and an environmental acoustic sensor array. The millimeter-wave radar array is installed on the guest room ceiling, with a detection accuracy better than ±5cm and an update frequency of no less than 100Hz. A dynamic assessment module is configured to construct a dynamic assessment model based on the multimodal data, wherein the dynamic assessment model is configured to fuse the multimodal perception data to perform privacy risk assessment; the dynamic assessment model outputs a privacy risk assessment value based on a normalized proximity value and an acoustic change index; a shielding control module, configured to determine a current shielding response level based on the privacy risk assessment value, including high risk, medium risk, and low risk levels, and to generate corresponding control instructions based on the level, the control instructions including execution parameters for optical shielding, acoustic shielding, and physical shielding; A protocol adapter module, used to convert the control instructions into a communication protocol compatible with the shielding execution device, supporting MQTT, Modbus RTU, RS485, OPC UA, Zigbee and CoAP protocol conversion, with a protocol conversion delay of no more than 50ms, and equipped with a backup protocol switching and fallback mechanism; A shielding execution module, including electrochromic glass, directional white noise equipment and electric curtain equipment, is used to execute corresponding shielding response operations; The power consumption management module is used to automatically switch the system to low-power standby mode when the risk assessment result is continuously lower than the risk threshold and no human approach or environmental changes are detected. The trigger conditions include the privacy risk assessment value being lower than the set threshold for a period of time and the shielding state remaining stable.

Images