Smart home control system

Abstract

The invention discloses an intelligent home control system, which relates to the technical field of intelligent home and comprises an illumination sensor, an infrared human body sensor, an auxiliary sensor, an edge analysis module, an illumination controller and an illumination strategy optimization module. The multi-source sensing data is subjected to fusion analysis through an edge analysis module, so that accurate light control based on multi-condition judgment is realized; and the lighting strategy optimization module dynamically adjusts a logic weight or a threshold value based on the abnormal behavior characteristics of the user, so that the control strategy has an adaptive optimization capability, and the response accuracy, stability and individuation level of the system are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of smart home technology, and in particular to a smart home control system. Background Art

[0002] A smart home control system is a system that uses advanced Internet of Things technology to connect various devices in the home (such as lighting, air conditioning, security, and home appliances) to a unified platform, allowing users to centrally control and manage them through mobile phones, voice, or automated scenarios. This system not only improves the comfort and convenience of living, but also enables energy consumption management and remote monitoring, enhancing the safety and intelligence of the home. For example, in smart lighting control, the system is usually equipped with light sensors, infrared human sensors, and networked control terminals. Users can set scene modes, such as "home mode," in the mobile phone app. When the user returns home in the evening, the door lock automatically unlocks the user's identity through Bluetooth recognition. The system senses human movement and determines that the indoor light is dim, automatically turning on the entrance and living room lights, and adjusting the light brightness and color temperature according to the user's preferences.

[0003] The existing technology has the following shortcomings:

[0004] Existing intelligent lighting control systems, based on the linkage control of "ambient light + personnel sensing", have the problem of sensor misjudgment leading to abnormal triggering. Especially in a multi-device deployment environment, infrared sensors are easily interfered with by heat signals from outside the space, causing lights to be frequently turned on when no one is around. This in turn leads to problems such as increased energy consumption, decreased user trust, and disordered system response. Summary of the Invention

[0005] The purpose of the present invention is to provide a smart home control system to solve the shortcomings of the background technology.

[0006] To achieve the above-mentioned object, the present invention provides the following technical solutions: a smart home control system, comprising a light sensor, an infrared human body sensor, an auxiliary sensor, an edge analysis module, a lighting controller, and a lighting strategy optimization module;

[0007] Light sensor, used to detect the ambient brightness of the target space;

[0008] Infrared human body sensor, used to sense human activity signals in the target space;

[0009] Auxiliary sensors, including door magnetic sensors and sound sensors;

[0010] an edge analysis module, connected to the light sensor, infrared human body sensor, and auxiliary sensor, respectively, for receiving and fusing various sensor data, and determining whether light-on conditions are met according to preset logic rules; if so, outputting a light-on signal to the lighting controller; otherwise, outputting a light-on or light-off signal;

[0011] A lighting controller, connected to the edge analysis module, for controlling the turning on or off of the light according to the analysis results;

[0012] The lighting strategy optimization module is connected to the edge analysis module and the lighting controller, and dynamically adjusts the judgment weight or threshold of the preset logic rules based on the user's abnormal behavior characteristics to optimize the lighting control strategy.

[0013] Preferably, the light sensor continuously collects light intensity in the target space through a photosensor, converts the collected light signal into a voltage or digital signal, and transmits it to the edge analysis module through a communication bus.

[0014] Preferably, the infrared human body sensor is used to cause changes in local thermal infrared distribution when a human body enters the sensing area; the changed heat signal is filtered and amplified to form a recognizable human activity signal; and the presence or absence status is output to the edge analysis module.

[0015] Preferably, the door magnetic sensor realizes opening and closing detection based on the Hall effect or reed switch contact induction. When the door is closed, the two parts are close to each other, and under the action of the magnetic field, the door closed state is output; when the door is opened, the two parts are separated, the magnetic field is disconnected, and the sensor outputs a door opening signal; the state change is reported to the edge analysis module in real time; the sound sensor is based on a capacitive microphone or a MEMS micro-microphone array to realize the perception of sound intensity and characteristics: the sound generates sound waves, causing the microphone diaphragm to vibrate and convert it into an electrical signal; after amplification and filtering, the signal is converted into a numerical sound pressure level; it is compared with the set sound pressure threshold to determine whether there is valid sound, and the judgment result is reported to the edge analysis module in real time.

[0016] Preferably, the edge analysis module processes the perception data collected by the light sensor, infrared human body sensor and auxiliary sensor through a fuzzy logic control algorithm, including:

[0017] Converting sensory data into fuzzy linguistic variables;

[0018] Match the preset fuzzy control rules and output the fuzzy control results;

[0019] Defuzzify the fuzzy control results and generate specific lighting control instructions;

[0020] The control instruction is transmitted to the lighting controller to execute the corresponding lighting operation.

[0021] Preferably, the lighting strategy optimization module dynamically adjusts the judgment weight or threshold of the preset logic rule based on the user's abnormal behavior characteristics, wherein the user's abnormal behavior characteristics include the frequency of lighting intervention and the degree of inconsistency between environmental conditions and behaviors.

[0022] Preferably, the method for obtaining the lighting intervention frequency is as follows: set the initial intervention frequency to S0, and in each monitoring cycle, if the user manually modifies the lighting status, record x t =1; otherwise x t =0, calculate the lighting intervention frequency, the expression is: S t =α·x t +(1-α)·S t-1 Where S t is the lighting intervention frequency at time t, S t-1 is the lighting intervention frequency at time t-1; α is the smoothing factor.

[0023] Preferably, the method for obtaining the degree of contradiction between environmental conditions and behaviors is as follows: set a set of preset rules, record the moment when the preset rule conditions are met, record whether the actual behavior of the user at this time is equal to the expected behavior of the rule, and record the update trigger times N. trigger and the number of hits N correct ; For each rule, calculate the environmental condition behavior contradiction degree V of the i-th rule during the observation period i , the expression is:

[0024] Preferably, the lighting intervention frequency and the environmental condition behavior contradiction degree are normalized so that they are both between [0, 1], and the normalized lighting intervention frequency and environmental condition behavior contradiction degree are weighted averaged and summed to obtain the matching deviation analysis value between the current lighting strategy and the user expectation.

[0025] Preferably, the obtained matching deviation analysis value between the current lighting strategy and the user's expectation is compared with a predetermined threshold. If the matching deviation analysis value between the current lighting strategy and the user's expectation is greater than or equal to the predetermined threshold, it indicates that there is a significant deviation between the system's current lighting control strategy and the user's actual expectation, and the strategy adjustment operation is immediately performed, including dynamically lowering the light trigger threshold, increasing the weight of the infrared or sound sensor in the decision logic, updating the rule confidence model, and eliminating long-term contradictory rules; if the matching deviation analysis value between the current lighting strategy and the user's expectation is less than the predetermined threshold, it indicates that the system control strategy is consistent with the user's expectation, and the current control strategy is kept unchanged, and the existing parameter configuration is continued to be observed and maintained.

[0026] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0027] 1. This invention builds a multi-source perception network consisting of light sensors, infrared human motion sensors, door sensors, and sound-assisted sensors, combined with a fuzzy logic control algorithm in the edge analysis module. This enables a multi-dimensional, dynamic, and integrated assessment of environmental conditions and human behavior. This effectively addresses the existing issue of abnormal lighting activation caused by infrared misjudgment. The system's accuracy and robustness are significantly improved, particularly in multi-device deployment scenarios. Compared to traditional single-threshold control logic, this invention offers greater scenario adaptability and fault tolerance, ensuring that lighting responses are more aligned with actual usage scenarios.

[0028] 2. This invention introduces a lighting strategy optimization module. Based on abnormal user behavior characteristics, such as the frequency of lighting interventions and the degree of conflict between environmental behavior, it establishes a user behavior model and implements adaptive strategy optimization. This enables the lighting control system to continuously perceive and learn user preferences, dynamically adjust judgment weights and control parameters, and form an intelligent closed-loop. By analyzing matching deviations and introducing a dynamic adjustment mechanism, this invention not only significantly reduces false trigger rates and energy consumption, but also improves user trust and interactive experience, achieving a more intelligent, personalized, low-power, and long-term stable home lighting control system. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0030] Figure 1 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0031] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in 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. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0032] For examples, see Figure 1 As shown, the smart home control system described in this embodiment includes a light sensor, an infrared human body sensor, an auxiliary sensor, an edge analysis module, a lighting controller, and a lighting strategy optimization module;

[0033] Light sensor, used to detect the ambient brightness of the target space;

[0034] Infrared human body sensor, used to sense human activity signals in the target space;

[0035] Auxiliary sensors, including door magnetic sensors and sound sensors;

[0036] an edge analysis module, connected to the light sensor, infrared human body sensor, and auxiliary sensor, respectively, for receiving and fusing various sensor data, and determining whether light-on conditions are met according to preset logic rules; if so, outputting a light-on signal to the lighting controller; otherwise, outputting a light-on or light-off signal;

[0037] A lighting controller, connected to the edge analysis module, for controlling the turning on or off of the light according to the analysis results;

[0038] The lighting strategy optimization module is connected to the edge analysis module and the lighting controller, and dynamically adjusts the judgment weight or threshold of the preset logic rules based on the user's abnormal behavior characteristics to optimize the lighting control strategy.

[0039] Light sensors are key environmental sensing components in smart home control systems. They detect the current ambient light intensity in a target space and use this as a basis for intelligent lighting control. These detection results help the system determine whether the space is "dim" or "sufficiently lit," thus deciding whether to automatically turn lights on or off.

[0040] Light sensors work based on the photoelectric effect. Common types include:

[0041] Light-dependent resistor (LDR) type: resistance changes with ambient light intensity;

[0042] Photodiode type: generates current after receiving light and has a fast response speed;

[0043] CMOS photoelectric sensor type: Based on the imaging element design, it can provide more accurate light intensity values;

[0044] Digital light intensity sensor (such as BH1750, TSL2561): directly outputs digital light value in lux.

[0045] A typical light sensor module includes: a photosensitive element (such as LDR or photodiode); a signal conditioning circuit (for amplification and filtering); an analog-to-digital converter (ADC) (if it is an analog output type); a communication interface (I 2 C, SPI, UART, etc.); housing package, with dust-proof and UV-proof design.

[0046] The workflow includes: the photosensitive element continuously collects the light intensity in the target space; the collected light signal is converted into a voltage or digital signal; and it is transmitted to the edge analysis module via the communication bus; the edge analysis module compares the light value with the set threshold (such as 50 Lux); if the brightness is lower than the threshold and other perception conditions are met, the lighting turn-on logic is triggered.

[0047] Usage methods include: installing it at the top of the room or near the window to avoid obstruction and reflection interference; deploying multiple light sensors in a large space to obtain more accurate light distribution; and automatically adjusting the judgment threshold according to day / night and seasonal changes to improve environmental adaptability.

[0048] Infrared human sensors are one of the core sensing units in intelligent lighting control systems, used to detect human activity within a target space. By identifying changes in natural infrared thermal radiation emitted by the human body, they dynamically detect a person's entry, stay, and movement, providing a key basis for the automatic triggering of intelligent scenarios.

[0049] Infrared human body sensors generally use passive infrared technology. Its basic principles are as follows:

[0050] The surface temperature of the human body will emit thermal radiation to the surrounding in the form of infrared band;

[0051] The PIR sensor itself does not actively emit infrared, but detects the dynamic changes of background infrared radiation caused by human activities through pyroelectric elements;

[0052] The sensor will only output a change signal when the infrared heat source (such as the human body) moves or changes its position;

[0053] A stationary human body will not be continuously sensed, so the system needs to combine a "continuous judgment mechanism" to improve judgment accuracy.

[0054] A standard infrared human body sensor usually includes the following components:

[0055] Pyroelectric sensing element: senses changes in infrared radiation; Fresnel lens array: focuses and partitions infrared signals to enhance the sensing range; signal processing circuit: amplifies and filters pyroelectric signals to remove clutter; control output module: outputs high and low levels or analog signals for system use; housing: has an infrared-transmitting window and has anti-static, dust-proof and other protective capabilities.

[0056] The workflow includes: after the system is started, the PIR module enters the monitoring state; when a person enters the sensing area, the local thermal infrared distribution changes; the changing heat signal is filtered and amplified to form a recognizable human activity signal; the "person / no person" status is output to the edge analysis module as one of the decision-making bases for lighting control; the system usually sets a "hold time" so that the lighting remains on even if a person is still for a short time.

[0057] Deployment and application strategies: Installed at the top or corner of the room, facing the door or high-activity area; sensing angle is 120° to 150° horizontally and 30° to 60° vertically, with an adjustable sensing distance of 510 meters; the sensing hold time can be set (such as 30 seconds or 60 seconds) to prevent the light from turning off when leaving for a short time; combined with a filtering algorithm, it can block interference sources such as hot air from air conditioners, pets, and small heat sources.

[0058] Door magnetic sensors are used to detect the open / close status of doors in target spaces. By determining whether a door is open, they can indirectly infer whether a person has entered the space. This, combined with signals like infrared sensors, can create a more accurate "entry and exit" logic chain.

[0059] The door magnetic sensor uses Hall effect or reed switch contact induction to detect opening and closing. It consists of a pair of components: a magnetic body (usually installed on the door) and a sensor (installed on the door frame). When the door is closed, the two parts are close to each other, and under the action of the magnetic field, the sensor outputs the "door closed" state. When the door is opened, the two parts separate, the magnetic field is disconnected, and the sensor outputs the "door open" signal. The state change is reported to the edge analysis module in real time and used as one of the conditions for lighting linkage.

[0060] The structural components include: magnetic component: high-magnetism permanent magnet; induction component: reed switch or Hall element + trigger circuit; packaging shell: with dust-proof and anti-disassembly design; signal output port: usually digital signal output (high / low level), some models support wireless (Zigbee / LoRa).

[0061] Installed at the main entrance and exit of the space (such as bedroom door, living room door); installation height is consistent to ensure sensitive sensing; the system can set time tags to record the time nodes of door opening behavior, assisting in behavioral path analysis; it can be associated with PIR infrared signals to confirm that the person has indeed "entered through the door" rather than "passed through the outside."

[0062] The sound sensor is used to detect whether there is acoustic activity in the target space, serving as an auxiliary basis for judging the actual presence of people. It can provide effective judgment supplements, especially when the infrared signal is silent or the people are stationary.

[0063] Sound sensors are usually based on capacitive microphones or MEMS micro-microphone arrays to perceive sound intensity and characteristics: sound generates sound waves, causing the microphone diaphragm to vibrate and convert into electrical signals; after amplification and filtering, the signal is converted into a numerical sound pressure level (unit: dB); this is compared with the set sound pressure threshold to determine whether there is "valid sound"; specific models can be used with voice recognition chips to identify voice-like features (such as conversation sounds) rather than background noise (such as air conditioning sounds).

[0064] The structural components include: microphone sensing unit: MEMS or electret condenser microphone; pre-amplifier and filtering circuit: processing weak signals; signal sampling and conversion module: output analog / digital sound intensity data; shell and noise protection processing: contains wind noise suppression and EMI shielding layer to improve recognition accuracy.

[0065] Deployment methods and usage strategies include: installing in the center of the space or in a quiet area to avoid strong sound sources at close range (such as TVs and speakers); setting an adjustable sound pressure threshold (such as 40dB to 60dB) to adapt to different environmental background sounds; combining a time window to determine the "duration of sound activity"; and can be used for "presence confirmation" or "activity evidence" in static infrared signal scenarios.

[0066] This invention utilizes door magnetic and sound sensors as auxiliary sensing units, combining light and infrared information for multi-source joint judgment. This effectively addresses the issues of single-sensor misjudgment and poor scene adaptability. The "event context information" provided by the auxiliary sensors significantly enhances the system's ability to identify "real human activity" and is a key support for robust lighting control.

[0067] The edge analysis module is the core processing unit in this invention. It is mainly responsible for receiving the original perception data from the light sensor, infrared human body sensor and auxiliary sensors (door magnet, sound), and performing real-time fusion analysis and logical judgment at the local end (edge), and finally outputting control instructions to the lighting controller to realize intelligent lighting management.

[0068] The structure includes: support I 2 C, SPI, GPIO and other communication methods; stable connection with multiple sensors such as light, infrared, door magnetic, sound, etc.; support sensor status acquisition frequency setting (such as 1 time / second to 10 times / second).

[0069] Embedded MCU, ARM Cortex-M, RISC-V and other low-power computing cores; with local data caching, real-time processing, logical judgment and output control capabilities; integrated lightweight rule engine or FSM state machine.

[0070] Time-synchronize and normalize multi-sensor data; apply preset logic rules to determine whether the "light-on conditions" are met;

[0071] Unlike existing intelligent lighting systems based on Boolean logic or single-threshold control, this invention incorporates a fuzzy logic control algorithm into the edge analysis module to achieve intelligent fusion and dynamic response to multi-source sensor data. Through this comprehensive analysis mechanism, the system not only determines the user's current behavior but also has the ability to predict behavioral trends, significantly reducing the false trigger rate in complex or ambiguous scenarios and improving the intelligence, stability, and personalization of the lighting system. Specifically:

[0072] In the target space, the system continuously collects the following environmental and behavioral signals:

[0073] The light sensor obtains the current ambient brightness (unit: Lux); the infrared motion sensor detects the frequency and intensity of human activity (outputting an activity status code or level change frequency); the sound sensor records the ambient sound intensity (unit: dB); and the door magnetic sensor obtains the current open / closed state of the door (binary switch value). All sensor data is transmitted in real time to the edge analysis module and sampled at a set frequency (for example, once per second).

[0074] The system converts this raw data into fuzzy set input values ​​using a "fuzzification function." The goal of fuzzification is to convert "continuous values" or "discrete states" into understandable language variables, such as "dark," "very dark," "slight movement," and "silent."

[0075] The specific conversion method is as follows:

[0076] Light intensity: Mapped from a continuous Lux value to "very bright", "normal", "dim", and "very dark";

[0077] Infrared activity level: The number of sensing times per unit time is mapped to "no one", "slight movement" or "active";

[0078] Sound intensity: mapped from sound pressure values ​​to "silent", "soft", "normal speech", and "noisy";

[0079] Door sensor status: Reserved as binary logic, used to determine whether entry or exit occurs.

[0080] This linguistic fuzzy variable processing makes the system closer to human judgment and supports tolerance of sensor errors and fluctuations.

[0081] Before system deployment or during the training phase, engineers set or learn a set of "fuzzy control rules." These rules, similar to empirical reasoning, describe "under what combination of perceptions should the light be turned on."

[0082] For example:

[0083] If [Light is "Very Dim"], [Infrared is "Active"], and [Sound is "Normal Talking"], then [Light should be "Full Bright"];

[0084] If the light is dim, the infrared is slightly moving, and the sound is soft, then the light should be half bright.

[0085] If [Light is "Normal"] or [Infrared is "No One"], then [Light should be "Off"].

[0086] The "if-then" logical statements in each rule can overlap and complement each other to form various situational judgments.

[0087] The edge analysis module matches the current fuzzy input variables with all the rules in the rule base through a fuzzy inference engine (such as the Mamdani model or the Sugeno model) to determine which rules are activated.

[0088] Each activated rule will generate a "fuzzy output" based on the membership of the input variable, that is, a fuzzy suggestion of what state the light should be in, such as "brightness is 0.7" or "off degree is 0.3".

[0089] The system will generate a comprehensive fuzzy output on "whether to turn on the light" and "the degree of turning on" based on the superposition results of multiple rules.

[0090] The fuzzy output results are converted into specific executable instructions through the "defuzzification" operation:

[0091] Map the fuzzy output value to a light brightness value between 0 and 100%;

[0092] If the comprehensive brightness recommendation value is higher than the set threshold (e.g. 40%), the system decides to “turn on the light”;

[0093] If the integrated value is close to zero or below a certain lower limit, the system decides to "turn off the lights".

[0094] This method ensures that the control instructions have gradual and flexible response capabilities, avoiding frequent switching of lights due to small fluctuations in a certain instantaneous signal.

[0095] The analysis results are sent to the lighting controller in the form of control instructions, which include:

[0096] Whether to turn on the light;

[0097] Specific brightness value (such as 70%, 50%, off);

[0098] Control the duration (such as setting a delay to automatically turn off).

[0099] After receiving the control signal, the lighting controller immediately performs the corresponding lighting adjustment.

[0100] The system continues to operate, updating input signals in real time and rerunning the fuzzy judgment process periodically (e.g., every 1 or 30 seconds). Incorporating user feedback or long-term statistical data, the system dynamically adjusts the shape and boundary values ​​of membership functions, the weights of control rules, and lighting response thresholds, achieving personalized adaptation. This mechanism enables the system to self-learn and self-optimize, adapting to long-term changes in factors such as seasonal changes and occupant habits.

[0101] The lighting controller receives control instructions (including light status, brightness value, etc.) from the edge analysis module and controls the working status of the connected lighting devices according to the instructions. Its core principles include:

[0102] Digital input control: determines whether the light should be turned on by receiving high / low level or communication instructions;

[0103] PWM dimming control: adjust the brightness of the lamp through pulse width modulation to achieve flexible lighting;

[0104] Smart protocol support: supports Zigbee, Z-Wave, BLE, Wi-Fi and other protocols to achieve remote control and local response;

[0105] Status retention and feedback: It can record the current lighting status and feed it back to the system master control to achieve two-way communication.

[0106] A typical lighting controller module includes the following parts:

[0107] Communication receiving unit: receives control data from edge analysis module; supports GPIO, UART, I 2 C. Wired interfaces such as RS485, or wireless communication methods such as Zigbee and Wi-Fi.

[0108] Logic Control Unit (MCU): parses control commands and executes state logic; controls output ports to connect to dimming circuits or switch drive circuits.

[0109] Execution drive unit: controls relays, MOS tubes, or LED driver chips to achieve switching or brightness adjustment; supports compatibility and adaptation of different types of lamps (incandescent lamps, LED lamps, thyristor dimming lamps, etc.).

[0110] Power module and protection circuit: Provides stable voltage for the system; has functions such as short circuit protection, overload protection, lightning protection and anti-interference.

[0111] The lighting controller can receive the following types of control commands:

[0112] Switch signal: logic high level → turn on the light; logic low level → turn off the light;

[0113] Brightness adjustment signal: Receives PWM signal duty cycle (such as 50%, 75%) to control LED light brightness; or receives digital brightness value (such as 0 to 255) for smooth dimming control.

[0114] Delay command (timed off): When receiving an on command with a timed attribute, the controller will turn off the lighting at a fixed time.

[0115] Intelligent scene commands (combination control): Combination commands such as "movie mode" and "home mode" can control multiple groups of lights at the same time to achieve multi-area linkage.

[0116] After each logical judgment, the edge analysis module immediately sends the result to the lighting controller, which responds in milliseconds. After the light status changes, the controller confirms the execution result through a receipt signal. The controller also has a power-off memory function, restoring the previous state after a power outage and restart, ensuring a consistent user experience.

[0117] Deployment and adaptation include: local deployment: usually installed near the lamp power line or in the low-voltage box; modular form: can be embedded module, DIN rail mounting box or wall panel integration; strong compatibility: adapt to the mainstream lamp types on the market, support single-fire, zero-fire, dual-control and other wiring methods.

[0118] The lighting strategy optimization module is used to dynamically adjust the judgment weights or logic thresholds used in the edge analysis module based on the user's abnormal behavior characteristics to adapt to changing actual usage scenarios and optimize the accuracy, comfort and personalized performance of lighting control decisions.

[0119] Among them, the user's abnormal behavior characteristics include the frequency of lighting intervention and the degree of inconsistency between environmental conditions and behavior.

[0120] Lighting intervention frequency: The frequency with which users manually modify system control results per unit time (e.g., manually turning lights on / off via the app or voice assistant). This value reflects the "decision-making deviation" between the current lighting strategy and user expectations. This value is obtained using:

[0121] Set the initial intervention frequency to S0, such as S0 = 0. In each monitoring cycle (for example, every 5 minutes): if the user manually modifies the light status (such as turning the light on or off), then record x t =1; otherwise x t = 0. Calculate the lighting intervention frequency, the expression is: S t =α·x t +(1-α)·S t-1 Where S t is the lighting intervention frequency at time t, S t-1is the frequency of light intervention at time t-1; α is the smoothing factor, which controls the sensitivity to new data; the larger α is, the faster the response to the current behavior and the more sensitive the system is. The recommended value range is: general scenario: α = 0.2 ~ 0.4; scenario with frequent changes in user behavior: α = 0.5 ~ 0.7; the calculated S t Compare with the set strategy optimization threshold θ: If S t ≥θ, indicating that users have frequently intervened recently and the system strategy needs to be adjusted; if S t <θ, indicating that the current strategy is stable and maintains the existing control logic. θ∈[0.3,0.5] (indicating that the intervention behavior density of 30% to 50% is the upper tolerance limit).

[0122] Environmental Condition Behavior Inconsistency: This indicator measures the degree of inconsistency between the environmental conditions detected by the system (e.g., "dark" + "quiet" + "no one") and the user's actual behavior (e.g., frequently turning on the light). This indicator reveals situations where the system misjudged "no one" or "inactivity," resulting in a failure to respond to lighting correctly.

[0123] Set a set of preset rules, for example: R1: dim + people → turn on the lights; R2: bright + no one → turn off the lights; R3: night + sound > 60dB → turn on the auxiliary lighting. In a set time period (such as every 24 hours), record the time when the preset rule conditions are met, record whether the user's actual behavior at this time is equal to the expected behavior of the rule; record the update trigger number N trigger and the number of hits N correct ; For each rule, calculate the environmental condition behavior contradiction degree V of the i-th rule during the observation period i , the expression is: V i The higher the value, the worse the system's prediction accuracy under the rule, and the more the user behavior "violates" the system's expectations, that is, the greater the degree of contradiction.

[0124] The lighting intervention frequency and the environmental condition behavior contradiction degree are normalized so that they are both between [0, 1]. The weighted average sum of the normalized lighting intervention frequency and environmental condition behavior contradiction degree is calculated to obtain the matching deviation analysis value between the current lighting strategy and user expectations.

[0125] The obtained matching deviation analysis value between the current lighting strategy and the user's expectations is compared with the predetermined threshold. If the matching deviation analysis value between the current lighting strategy and the user's expectations is greater than or equal to the predetermined threshold, it means that there is a significant deviation between the system's current lighting control strategy and the user's actual expectations, the user frequently intervenes in the system behavior or there are a large number of rule violations. The system should immediately perform strategy adjustment operations, including: dynamically lowering the light trigger threshold to improve the brightness response sensitivity; increasing the weight of infrared or sound sensors in the decision logic; updating the rule confidence model to eliminate long-term contradictory rules; and establishing user preference modes (such as customized brightness strategies for time periods).

[0126] If the match deviation analysis value between the current lighting strategy and user expectations is less than a predetermined threshold, it indicates that the system control strategy is generally consistent with user expectations, user intervention frequency is low, behavior is stable, and system rules are trustworthy. The system can maintain the current control strategy, continue to observe and maintain the existing parameter configuration. If the match deviation analysis value between the current lighting strategy and user expectations is less than a predetermined threshold for multiple consecutive cycles, the system can determine that the strategy has entered a "stable phase" and enter a low-frequency update state. If there are significant fluctuations within a short period of time but they do not exceed the threshold, buffering can be performed to avoid frequent strategy switching.

[0127] It should be noted here that the predetermined threshold can be set according to the user's tolerance for the intelligent system. For example, if the tolerance for deviation is low, the threshold is set to 0.3. In high-level implementations, the threshold can be dynamically adjusted (such as using a dynamic learning model update).

[0128] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A smart home control system, characterized by: Includes light sensors, infrared human body sensors, auxiliary sensors, edge analysis modules, lighting controllers, and lighting strategy optimization modules; Light sensor, used to detect the ambient brightness of the target space; Infrared human body sensor, used to sense human activity signals in the target space; Auxiliary sensors, including door magnetic sensors and sound sensors; an edge analysis module, connected to the light sensor, infrared human body sensor, and auxiliary sensor, respectively, for receiving and fusing various sensor data, and determining whether light-on conditions are met according to preset logic rules; if so, outputting a light-on signal to the lighting controller; otherwise, outputting a light-on or light-off signal; A lighting controller, connected to the edge analysis module, for controlling the turning on or off of the light according to the analysis results; The lighting strategy optimization module is connected to the edge analysis module and the lighting controller, and dynamically adjusts the judgment weight or threshold of the preset logic rules based on the user's abnormal behavior characteristics to optimize the lighting control strategy.

2. The smart home control system according to claim 1, characterized in that: The light sensor continuously collects the light intensity in the target space through a photosensor, converts the collected light signal into a voltage or digital signal, and transmits it to the edge analysis module through a communication bus.

3. The smart home control system according to claim 1, characterized in that: The infrared human body sensor is used to cause changes in the local thermal infrared distribution when a human body enters the sensing area; the changed heat signal is filtered and amplified to form a recognizable human activity signal; and the presence or absence status is output to the edge analysis module.

4. The smart home control system according to claim 1, characterized in that: The door magnetic sensor uses Hall effect or reed switch contact induction to detect opening and closing. When the door is closed, the two parts are close to each other, and under the action of the magnetic field, the door is output as a closed state; when the door is opened, the two parts separate, the magnetic field is disconnected, and the sensor outputs a door open signal; state changes are reported to the edge analysis module in real time; The sound sensor is based on a capacitive microphone or a MEMS micro-microphone array to perceive the intensity and characteristics of sound: the sound generates sound waves, causing the microphone diaphragm to vibrate and convert it into an electrical signal; after amplification and filtering, the signal is converted into a numerical sound pressure level; it is compared with the set sound pressure threshold to determine whether there is valid sound, and the judgment result is reported to the edge analysis module in real time.

5. The smart home control system according to claim 1, characterized in that: The edge analysis module processes the perception data collected by the light sensor, infrared human body sensor and auxiliary sensor through a fuzzy logic control algorithm, including: Converting sensory data into fuzzy linguistic variables; Match the preset fuzzy control rules and output the fuzzy control results; Defuzzify the fuzzy control results and generate specific lighting control instructions; The control instruction is transmitted to the lighting controller to execute the corresponding lighting operation.

6. The smart home control system according to claim 1, characterized in that: The lighting strategy optimization module dynamically adjusts the judgment weight or threshold of the preset logic rule based on the user's abnormal behavior characteristics, wherein the user's abnormal behavior characteristics include the frequency of lighting intervention and the degree of inconsistency between environmental conditions and behaviors.

7. The smart home control system according to claim 6, characterized in that: The method for obtaining the lighting intervention frequency is as follows: set the initial intervention frequency to S0, and in each monitoring cycle, if the user manually modifies the lighting status, then record x t =1; otherwise x t =0, calculate the lighting intervention frequency, the expression is: S t =α·x t +(1-α)·S t-1 Where S t is the lighting intervention frequency at time t, S t-1 is the lighting intervention frequency at time t-1; α is the smoothing factor.

8. The smart home control system according to claim 7, characterized in that: The method for obtaining the degree of contradiction between environmental conditions and behaviors is as follows: set a set of preset rules, record the moment when the preset rule conditions are met, record whether the user's actual behavior at this time is equal to the expected behavior of the rule, and record the update trigger times N trigger and the number of hits N correct ; For each rule, calculate the environmental condition behavior contradiction degree V of the i-th rule during the observation period i , the expression is:

9. The smart home control system according to claim 8, characterized in that: The lighting intervention frequency and the environmental condition behavior contradiction degree are normalized so that they are both between [0, 1]. The weighted average sum of the normalized lighting intervention frequency and environmental condition behavior contradiction degree is calculated to obtain the matching deviation analysis value between the current lighting strategy and user expectations.

10. The smart home control system according to claim 9, characterized in that: The obtained matching deviation analysis value between the current lighting strategy and the user's expectation is compared with the predetermined threshold. If the matching deviation analysis value between the current lighting strategy and the user's expectation is greater than or equal to the predetermined threshold, it indicates that there is a significant deviation between the system's current lighting control strategy and the user's actual expectation. The strategy adjustment operation is immediately performed, including dynamically lowering the light trigger threshold, increasing the weight of the infrared or sound sensor in the decision logic, updating the rule confidence model, and eliminating long-term contradictory rules. If the matching deviation analysis value between the current lighting strategy and the user's expectation is less than the predetermined threshold, it indicates that the system control strategy is consistent with the user's expectation. The current control strategy remains unchanged, and the existing parameter configuration continues to be observed and maintained.

Images