A lighting control false trigger filtering method and system based on multi-sensor fusion

Abstract

The application discloses a kind of lighting control false trigger filtering method and system based on multi-sensing fusion, belong to intelligent lighting control technical field.Method includes: the phase value of synchronous acquisition millimeter wave radar echo signal and the output voltage value of passive infrared sensor, and the sampling frequency of two is same;Respectively calculate the change direction of phase value and output voltage value at adjacent time;In the proportion of the number of sampling time that two change directions are same in the total number of sampling time in preset time window;When proportion is greater than or equal to preset threshold, output lighting control signal.The application compares the consistency of radar phase change direction and passive infrared sensor voltage change direction, changes from instantaneous feature matching to time domain change direction consistency verification, filters interference source such as pet, curtain flutter without increasing hardware cost, while maintaining lighting not off in human body stationary state.

Description

Technical Field

[0001] This invention relates to the field of intelligent lighting control technology, specifically to a method and system for filtering false triggers in lighting control based on multi-sensor fusion. Background Technology

[0002] Passive infrared (PIR) sensors are the most widely used lighting control sensing solution on the market due to their low cost and mature technology. Their working principle is to detect changes in 9-10μm infrared radiation emitted by the human body to determine the presence of people in the detection area. However, PIR sensors are essentially thermal change detectors, not image sensors. Any external factor that causes a sudden change in thermal radiation in the scene can lead to sensor misjudgment. For example, when a pet passes through the detection area, the resulting thermal radiation change is highly similar to human characteristics, easily causing lights to turn on for no apparent reason; temperature fluctuations caused by hot and cold airflow from air conditioner vents will also frequently trigger PIR sensors; when wind blows curtains, the difference in surface temperature between the curtains and the room temperature causes a sudden change in thermal radiation, which is often misinterpreted as human activity. Furthermore, PIR sensors can only sense dynamic changes in thermal radiation and cannot identify people in a stationary state.

[0003] To address the aforementioned shortcomings of single-sensor systems, Chinese patent application CN121703955A discloses a human detection method, device, equipment, and medium based on dual-technology sensors. The specific solution involves simultaneously acquiring radar and pyroelectric infrared signals, decoupling and fusing their features, and then using a decision tree for real-time classification. The system determines whether the target is a real person based on the target's inherent micro-motion-thermal radiation coupling fingerprint. While this solution improves system reliability, it still has significant drawbacks: the computational load of feature extraction combined with decision tree inference is substantial, requiring considerable computing power from the microcontroller (MCU), which is difficult for ordinary low-cost microcontrollers to meet real-time operational needs. Furthermore, this solution primarily addresses relatively simple situations such as the presence of micro-motion without a heat source or the presence of a heat source without micro-motion. Real-life scenarios are more complex. For example, when curtains sway, radar sensors can detect micro-motion signals, and PIR sensors can collect changes in thermal radiation, but this event is not considered human activity. Air conditioning vents can also trigger both types of sensors. In addition, these schemes share a common flaw: they rely solely on instantaneous signal characteristics to make judgments, failing to fully explore the temporal evolution process and change patterns of the signal over a period of time.

[0004] Furthermore, some products on the market use an AND logic between radar and PIR sensors, determining the presence of a person in the detection area only when both types of sensors are triggered simultaneously. While this approach reduces the false trigger rate, it also simultaneously reduces the detection sensitivity for real human targets. The system requires the person to make a significant movement or be at a closer distance to trigger the illumination. When the person is stationary or far from the sensor, the system cannot respond properly. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method and system for filtering false triggering in lighting control based on multi-sensor fusion. By acquiring the phase value of the millimeter-wave radar echo signal and the output voltage value of the passive infrared sensor, the method calculates the direction of change of both signals at adjacent moments. Within a preset time window, it counts the proportion of sampling moments where the two signals have the same direction of change to the total number of sampling moments. When this proportion reaches a preset threshold, a lighting control signal is output. This transforms the judgment criterion from instantaneous feature matching to temporal domain direction consistency verification, achieving the technical effect of accurately filtering various types of interference sources such as pets, swaying curtains, and air conditioning airflow without increasing hardware costs or computational complexity.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for filtering false triggers in lighting control based on multi-sensor fusion, comprising:

[0008] The phase value of the millimeter-wave radar echo signal and the output voltage value of the passive infrared sensor are acquired simultaneously, and the sampling frequency of both is the same.

[0009] Based on the phase value, calculate the first direction of change of the phase value at adjacent time points; based on the output voltage value, calculate the second direction of change of the output voltage value at adjacent time points;

[0010] Within a preset time window, the proportion of sampling times where the first change direction and the second change direction are the same to the total number of sampling times within the preset time window is calculated.

[0011] When the ratio is greater than or equal to a preset threshold, a lighting control signal is output.

[0012] Preferably, the first direction of change of the phase value at adjacent moments is calculated based on the phase value; the second direction of change of the output voltage value at adjacent moments is calculated based on the output voltage value, including: calculating the difference between the phase value or output voltage value at the current moment and the phase value or output voltage value at the previous moment; if the difference is positive, the corresponding first direction of change or the second direction of change is positive; if the difference is negative, the corresponding first direction of change or the second direction of change is negative.

[0013] Preferably, the length of the preset time window is configured to cover at least one human respiratory cycle. The respiratory rate of a human body at rest is 12 to 20 breaths per minute, corresponding to a respiratory cycle of 3 to 5 seconds. Setting the preset time window to cover at least one complete respiratory cycle can capture the periodic changes in human respiratory signals.

[0014] Preferably, when simultaneously acquiring the phase value of the millimeter-wave radar echo signal and the output voltage value of the passive infrared sensor, the process includes: acquiring the amplitude value of the millimeter-wave radar echo signal; taking the arithmetic mean of the amplitude values ​​at multiple consecutive sampling times to obtain the amplitude average; and terminating the lighting control false trigger filtering process when the amplitude average is less than a preset amplitude threshold. This scheme utilizes the positive correlation between radar echo amplitude and target radar cross section (RCS). The RCS of small pets such as cats and small dogs is much smaller than that of humans, and interference from small pets can be quickly eliminated through amplitude threshold prediction.

[0015] Preferably, the lighting control false triggering filtering method based on multi-sensor fusion further includes: after outputting the lighting control signal, determining the spatial partition where the human body is located based on the human body azimuth angle and target distance measured by the millimeter-wave radar, and outputting a drive signal to the lighting fixtures in the corresponding partition.

[0016] Preferably, the human body azimuth angle is obtained by phase comparison angle measurement method, which uses the phase difference of the echo signals received by two receiving antennas, and calculates the target angle according to the correspondence between the phase difference and the target azimuth angle.

[0017] Preferably, determining the spatial partition where the human body is located based on the human body azimuth angle and target distance measured by the millimeter-wave radar includes: dividing the room plane into four quadrants with the millimeter-wave radar installation position as the origin, and determining the partition where the human body is located based on the quadrant to which the azimuth angle belongs and the distance.

[0018] Preferably, the target distance is compared with a preset distance threshold. If the distance is greater than the preset distance threshold, it is determined that the human body is not in the room or is at the edge of the room, and the lighting is not triggered or the nearest partition is triggered. The preset distance threshold is half the length of the room's diagonal.

[0019] Preferably, the sampling time interval between adjacent moments is configured such that the preset time window contains at least 20 sampling moments. Ensuring at least 20 sampling points makes the statistical results statistically significant. When the sample size is too small, there is a greater chance of randomness, making it impossible to reliably distinguish between real signals and random noise; when the sample size reaches 20 or more, the statistical results tend to be stable and can meet the basic requirements of the law of large numbers.

[0020] Preferably, when the lighting control false triggering filtering method based on multi-sensor fusion is run in parallel in multiple different areas, each area independently performs human presence detection. The lighting in the area where the target is detected leaving is turned off after a delay, and the lighting in the area where the target is detected entering remains on.

[0021] In a second aspect, the present invention provides a lighting control false triggering filtering system based on multi-sensor fusion, for implementing the method described in any of the above claims, comprising:

[0022] The signal acquisition unit has a built-in analog-to-digital converter and is used to simultaneously acquire the phase value of the millimeter-wave radar echo signal and the output voltage value of the passive infrared sensor.

[0023] A direction comparator is configured with a change direction comparison algorithm to calculate the change direction of the phase value and the output voltage value at adjacent time points;

[0024] A statistical analyzer is used to count the proportion of sampling times with the same direction of change within the preset time window to the total number of sampling times within the preset time window;

[0025] The detector is used to output a lighting control signal when the ratio is greater than or equal to a preset threshold.

[0026] A zone controller, connected to the determiner, is used to determine the spatial zone where the human body is located based on the human body azimuth angle and target distance measured by the millimeter-wave radar, and output drive signals to the lighting fixtures of the corresponding zone.

[0027] The output of the signal acquisition unit is connected to the input of the direction comparator, the output of the direction comparator is connected to the input of the statistician, the output of the statistician is connected to the input of the decision-maker, and the output of the decision-maker is connected to the input of the partition controller.

[0028] Compared with the prior art, the beneficial effects of the present invention are:

[0029] 1. This invention utilizes the physical law that the radar phase change caused by human respiration and the PIR thermal radiation change are in the same direction. It can filter out multiple types of interference sources such as pet activity, curtain swaying, and air conditioning airflow without increasing hardware costs, thus significantly reducing the false trigger rate of lighting control.

[0030] 2. This invention only requires subtraction and sign judgment, with extremely low computational load. It can run in real time on low-cost 8-bit or 32-bit MCUs without the need for dedicated AI chips or digital signal processors (DSPs).

[0031] 3. This invention uses millimeter-wave radar to detect micro-movements in the chest cavity caused by breathing, solving the problem that PIR sensors cannot detect effective signals when the human body is stationary, thus causing the lighting to automatically turn off, and realizing continuous lighting when the human body is stationary. Attached Figure Description

[0032] Figure 1 This is a structural block diagram of a lighting control false triggering filtering system based on multi-sensor fusion according to the present invention;

[0033] Figure 2 This is a flowchart illustrating a method for filtering false triggering in lighting control based on multi-sensor fusion according to the present invention.

[0034] Figure 3 This is a schematic diagram of the installation and spatial partitioning of a lighting control false triggering filtering device based on multi-sensor fusion according to the present invention. Detailed Implementation

[0035] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof.

[0036] Example 1

[0037] Please see Figure 1 This embodiment provides a lighting control false triggering filtering system based on multi-sensor fusion. The system includes: a millimeter-wave radar module, a passive infrared (PIR) sensor, a signal preprocessing unit, an MCU processor, an LED driver circuit, and an LED lamp group.

[0038] The millimeter-wave radar module is a 24 GHz frequency-modulated continuous wave (FMCW) radar module. FMCW radar calculates the target's range and velocity by transmitting a linear frequency-modulated continuous wave signal and analyzing the frequency difference between the reflected and transmitted signals. Specifically, the radar module continuously transmits a linear frequency-modulated continuous wave signal into the target area in the 24 GHz to 24.25 GHz frequency band. The transmitted signal can be represented as:

[0039]

[0040] in, For time variables, For the amplitude of the transmitted signal, The starting frequency is set to 24GHz in this embodiment. In this embodiment, the frequency modulation bandwidth is set to 250MHz. The duration of a single chirp. When the transmitted signal encounters a target object, it produces a reflected echo, which is captured by the radar receiving antenna. The transmitted signal and the echo signal are mixed in a mixer to generate an intermediate frequency (IF) signal, where the frequency of the IF signal is... Distance from target satisfy:

[0041]

[0042] in, The speed is the speed of light. By performing analog-to-digital conversion and fast Fourier transform on the intermediate frequency signal, the range gate information of the target can be obtained. After determining the range gate of the target, phase information is extracted from the complex baseband signal corresponding to the range gate. Specifically, for the first... The complex baseband signal of a chirped signal at the corresponding range gate can be expressed as:

[0043]

[0044] in, The imaginary unit, For in-phase components, For orthogonal components, For signal amplitude, For the goal in the Radial distance at each chirp This is the radar's operating wavelength. From this, the phase value can be extracted:

[0045]

[0046] The phase value Distance from target Between When the human chest cavity undergoes millimeter-level periodic displacement due to respiration, the target distance... A minute change occurs, and this change is directly reflected in the phase value. Above. The phase detection accuracy of a 24GHz radar can reach the order of 1 / 1000th of the wavelength, making it fully capable of detecting millimeter-level chest cavity displacement caused by human respiration. The radar module simultaneously outputs the amplitude value of the radar echo signal. and phase value Amplitude value The amplitude of the echo is positively correlated with the radar cross section (RCS) of the target. That is, the larger the RCS, the larger the echo amplitude. There is a significant difference in the RCS between the human body and small pets. Therefore, the amplitude value can be used to quickly eliminate interference from small pets.

[0047] PIR sensors, such as the RE200B model, are used to detect infrared thermal radiation signals from a target area and output a voltage signal. The PIR sensor contains two pyroelectric elements connected in series, forming a differential structure. When a human enters the detection area, the 9-10μm infrared radiation emitted by the human body is focused onto the pyroelectric elements through a Fresnel lens. Due to the temperature difference between the human body and the ambient temperature, the infrared radiation energy received by the pyroelectric elements changes, leading to a change in the charge distribution on the element surface. The two pyroelectric elements receive different amounts of heat and cannot cancel each other out, thus generating a changing voltage signal at the output. This voltage signal is then output after impedance transformation by an internal field-effect transistor. The PIR sensor outputs an analog voltage signal, the rate and amplitude of which are related to the rate and amplitude of change in the target's thermal radiation. When a human breathes, the expansion and contraction of the chest cavity causes a periodic change in the body's surface heat dissipation area, resulting in a slight fluctuation in the PIR sensor's output voltage at the same frequency.

[0048] The signal preprocessing unit is electrically connected to the millimeter-wave radar module and the PIR sensor, and is used to filter, amplify, and perform analog-to-digital conversion (ADC) on the radar signal and the PIR sensor signal. Specifically, the radar intermediate frequency signal is first filtered by a low-pass filter to remove high-frequency noise, then sent to a programmable gain amplifier for signal amplification, and finally converted into a digital signal by an ADC for the MCU processor to read. The PIR sensor output voltage signal is also filtered, amplified, and converted by an ADC before being sent to the MCU processor.

[0049] An MCU processor, such as an STM32F103, is electrically connected to the signal preprocessing unit and has a built-in change direction comparison algorithm for performing the false trigger filtering judgment of this invention. The MCU processor uses a set sampling frequency, 20Hz in this embodiment. This sampling frequency is chosen because the human respiratory rate is 0.2~0.33Hz, and the 20Hz sampling frequency is much higher than the respiratory rate, enabling complete capture of the waveform changes of the respiratory signal. Simultaneously, corresponding to a 50ms sampling time interval, the phase value output by the radar module is read through the serial peripheral interface (SPI) or integrated circuit bus (I2C) interface. and amplitude value Simultaneously, the output voltage value of the PIR sensor is read through the ADC channel. These values, as a time series, are fed into the MCU processor and used as continuously sampled data. The MCU processor internally runs a change direction comparison algorithm that performs the following operations: calculates the change direction of the radar module's output phase at adjacent sampling times, i.e., calculates the phase value at the current time. Phase value from the previous moment The difference is calculated; if the difference is positive, the direction of change at that moment is positive; if the difference is negative, the direction of change at that moment is negative. The direction of change of the PIR sensor output voltage at adjacent sampling moments is calculated, i.e., the voltage value at the current moment is calculated. voltage value at the previous moment The difference is calculated, and if the difference is positive, the direction of change at that moment is positive; if the difference is negative, the direction of change at that moment is negative. Within a preset time window, the proportion of sampling moments in which the two directions are consistent is counted, the statistical proportion is compared with a preset threshold, and the judgment result is output.

[0050] The LED driver circuit is electrically connected to the MCU processor and is used to drive the LED group according to the lighting control signals output by the MCU processor. The LED driver circuit includes four independent pulse width modulation (PWM) output channels, each of which can independently adjust its duty cycle to control the brightness of the corresponding zone's LED lights. For example, when the brightness of the LED lights in zone A needs to be adjusted to 50%, the MCU processor outputs a PWM signal with a 50% duty cycle to the PWM channel corresponding to zone A; when it needs to be adjusted to 100% brightness, it outputs a PWM signal with a 100% duty cycle. Each channel can also output independent color temperature control signals, achieving color temperature adjustment by adjusting the current ratio of cool white LEDs and warm white LEDs, with a color temperature range of 2700K~6500K. In this embodiment, the nominal color temperature of the cool white LED is 6500K, and the nominal color temperature of the warm white LED is 2700K. The color temperature of the mixed light is determined by the ratio of the luminous flux emitted by the cool white LED and the warm white LED, and the luminous flux ratio is positively correlated with the driving current ratio. When the target color temperature is 3000K, which falls between 2700K and 6500K and is closer to 2700K, the driving current of the warm white LED should be greater than that of the cool white LED. Specifically, the luminous flux of both cool white and warm white LEDs is approximately proportional to their driving current. To achieve a mixed color temperature of 3000K, the required luminous flux of the warm white LED is about three times that of the cool white LED. Since the luminous efficacy of the two types of LEDs is similar, setting the driving current ratio of the cool white LED to the warm white LED to 1:3 will make the luminous flux of the warm white LED about three times that of the cool white LED, resulting in a mixed color temperature of approximately 3000K. Similarly, when the target color temperature is 6000K, which is closer to 6500K, the required luminous flux of the cool white LED is about three times that of the warm white LED. Therefore, setting the driving current ratio of the cool white LED to the warm white LED to 3:1 will result in a mixed color temperature of approximately 6000K.

[0051] The LED light groups consist of multiple units, each corresponding to a different spatial lighting zone. Each LED light group has its color temperature and illuminance adjusted independently. Each zone contains one set of cool white LEDs and one set of warm white LEDs, driven by independent PWM channels. Specifically, the cool white LEDs are driven by the first PWM channel, and the warm white LEDs are driven by the second PWM channel. By adjusting the duty cycle of the two channels, the brightness of the cool white and warm white light is controlled respectively, thereby achieving independent adjustment of color temperature and illuminance.

[0052] Please see Figure 2 This embodiment provides a method for filtering false triggers in lighting control based on multi-sensor fusion, including the following steps:

[0053] S1: Synchronously acquire the phase value of the millimeter-wave radar echo signal and the output voltage value of the passive infrared sensor, and the sampling frequency of both is the same.

[0054] Specifically, the MCU processor synchronously acquires the phase value of the millimeter-wave radar echo signal at a sampling frequency of 20Hz through the signal preprocessing unit. and the output voltage value of the PIR sensor Synchronous acquisition is achieved by triggering ADC and SPI read operations through the MCU processor's timer. The timer generates an interrupt with a period of 50ms, which corresponds to a sampling frequency of 20Hz. That is, the sampling period is the reciprocal of the sampling frequency. In the interrupt service routine, ADC conversion and SPI read operations are started simultaneously to ensure that the two signals are strictly aligned on the time axis, providing an accurate time reference for comparison of changing directions.

[0055] In this step, the MCU processor also acquires the amplitude value of the millimeter-wave radar echo signal. Amplitude values ​​at multiple consecutive sampling times The arithmetic mean is used to obtain the mean amplitude. This is used to characterize the overall level of the echo signal amplitude over a period of time. Amplitude value During the sampling process and phase value Synchronous acquisition. The MCU processor will... With preset amplitude threshold Comparison, It is an empirical value calibrated based on the minimum radar cross section (RCS) of the human body, usually determined experimentally: in a typical indoor environment, people of different body types stand at different distances, and the radar echo amplitude is recorded. The lower limit of the human body's RCS is taken as the baseline. . and Both are normalized amplitude values, and since they are on the same dimension, they can be directly compared. If Less than If the target is not identified or is a small pet, the subsequent steps are terminated directly without triggering illumination. The RCS of a small pet is much smaller than that of a human body, and its radar echo amplitude is significantly lower than that of a human body. Therefore, interference from small pets can be quickly eliminated by using the amplitude threshold.

[0056] S2: Based on the phase value, calculate the first direction of change of the phase value at adjacent time intervals; based on the output voltage value, calculate the second direction of change of the output voltage value at adjacent time intervals.

[0057] Specifically, setting the sampling time interval This corresponds to a sampling frequency of 20Hz. The MCU processor calculates the first change direction indicator of the radar phase value in adjacent time intervals. Indicator of the direction of change of the PIR sensor output voltage value at adjacent time points :

[0058]

[0059]

[0060] in, exist The value is +1, indicating that the phase value is increasing, i.e., positive. When the value is -1, it indicates a decrease in phase, i.e., a negative phase. The value is 0, indicating no change.

[0061] Taking a specific numerical value as an example: Suppose that the data was collected at time t=0. rad, ;exist Real-time collection rad, .but If it is positive, then it is positive. Both are positive, and are considered to be in the same direction. Conversely, if... Decline and Rise, or Rise and If it is downward, it is recorded as a misalignment in direction.

[0062] This step involves only subtraction and sign determination. Subtraction is used to calculate the difference, and sign determination is used to determine the sign of the difference. The computational load is minimal and can be completed in real time on a low-cost MCU. In embedded implementations, sign determination only requires checking the highest bit of the binary number, without any additional arithmetic operations.

[0063] S3: Within a preset time window, count the proportion of sampling times where the first change direction and the second change direction are the same to the total number of sampling times within the preset time window.

[0064] Specifically, a preset time window T = 3 seconds is set. Within this 3-second window, 60 sets of data are collected. The statistical analyzer has a circular buffer of length 60 to store the data from the most recent 60 sampling times. and Value. Whenever a new pair... and After the calculation is complete, the statistician removes the oldest data from the buffer, stores the new data in the buffer, and then iterates through all 60 sets of data in the buffer, comparing each sample time frame by frame. and Count the number of sampling times when both values ​​are the same. That is, both are +1 or both are -1. Calculate the consistency ratio. .

[0065] The physical meaning of the same direction of change is that the radar echo phase and the PIR sensor output voltage change in the same direction at the same time, either rising or falling simultaneously. During real human respiration, the expansion of the chest cavity increases the radar reflection path, causing the radar echo phase to rise. Simultaneously, the area of ​​the body radiating heat increases, leading to an increase in the PIR sensor output voltage. Conversely, when the chest cavity contracts, the radar reflection path shortens, causing the radar echo phase to fall. Simultaneously, the area of ​​the body radiating heat decreases, causing the PIR sensor output voltage to fall. Therefore, in real-world human scenarios, the directions of change for the two signals are highly consistent in time. However, interference sources, such as swaying curtains, can cause the radar phase change and the PIR sensor's heat radiation change to be asynchronous or even opposite. When the curtains sway outwards, the change in the radar reflection path causes the phase to rise, but the curtain surface temperature may be lower than room temperature, causing a decrease in the heat radiation detected by the passive infrared sensor. The two changes in opposite directions, resulting in a lower consistency rate.

[0066] The preset time window T=3 seconds is based on the physiological characteristics of human respiratory rate. The human body's resting respiratory rate is 12-20 breaths per minute, corresponding to a respiratory cycle of 3-5 seconds. The 3-second preset time window precisely covers approximately one complete respiratory cycle, allowing the system to capture the complete waveform of the respiratory signal and reliably determine whether the changes in the two signals are consistent. If the preset time window is too short, such as less than 1 second, it may not cover a complete respiratory cycle, resulting in insufficient judgment data; if the preset time window is too long, such as exceeding 10 seconds, the system response delay will be too large, affecting the user experience.

[0067] S4: When the number of sampling times with the same first change direction and the second change direction accounts for a proportion of the total number of sampling times within the preset time window that is greater than or equal to a preset threshold, a lighting control signal is output.

[0068] Specifically, setting a preset threshold .like The MCU processor determines that a real human body exists in the target area and outputs a lighting control signal to the LED driver circuit, illuminating the corresponding LED group. If... The MCU processor determines that the light was falsely triggered and does not output a lighting control signal, so the LED lights remain off.

[0069] threshold The meaning is: within a preset time window of 3 seconds, the critical value is 70% of the sampling moments in which the direction of radar phase change is consistent with the direction of PIR sensor voltage change. The selection of this threshold is based on the following: Experimental tests were conducted in a typical indoor environment (4m×4m room, temperature 25±2℃, humidity 50%±10%). The test subjects were 10 adult volunteers, in sitting, standing, and slightly active states. Interference source tests included scenarios such as curtains swaying, wind force 3-4, a pet cat weighing approximately 4kg, and air conditioning airflow with an outlet wind speed of 2-3 m / s. Experimental results showed that in the real-person breathing scenario, the consistency rate of the two signal change directions ranged from 80% to 95%, with an average of 88%; in the interference source scenario, the consistency rate ranged from 25% to 55%, with an average of 40%. Therefore, 70% was selected as the threshold for distinguishing between real people and interference sources. Of the 60 sampling points, 70% correspond to at least 42 sampling points where the two signals change in the same direction. This value is calculated from 60 × 70% = 42.

[0070] Optionally, after outputting the lighting control signal, the MCU processor also determines the spatial zone where the human body is located based on the azimuth angle of the human body measured by the millimeter-wave radar and the target distance, and outputs drive signals to the lighting fixtures in the corresponding zone. Specifically, the radar module obtains the azimuth angle of the human body by using the phase difference of the two receiving antennas to calculate the target angle. Through the linear relationship between intermediate frequency signal frequency and distance Calculate human distance The MCU processor is based on and Determine the partition where the human body is located, and only output drive signals to the PWM channel of the corresponding partition.

[0071] Taking a user entering a room as an example, the overall workflow of the system described in this embodiment is as follows: The signal acquisition unit continuously acquires radar phase values ​​and PIR sensor output voltage values ​​at a sampling frequency of 20Hz, and transmits each set of data to the direction comparator. The direction comparator performs a change direction calculation every 50ms and outputs two direction flags to the statistical unit, i.e., +1 or -1. The statistical unit has a circular buffer of length 60. Each time a new set of data is received, the buffer is updated, and the statistical consistency ratio is re-traversed. ,Will The value is transmitted to the decision-maker in real time. The decision-maker will... Compared with the preset threshold of 70%, if The system then outputs a high-level control signal to the LED driver circuit. Simultaneously, the zone controller receives the azimuth angle measured by the millimeter-wave radar. With distance The azimuth angle is obtained through phase comparison angle measurement, and the distance is calculated based on the correspondence between the intermediate frequency signal frequency, frequency modulation bandwidth, chirp duration, and speed of light. After determining the zone where the human body is located, a drive signal is output to the corresponding zone's PWM channel. The entire process completes one cycle within 50ms, realizing real-time detection of the human body's presence and zoned lighting control.

[0072] Example 2

[0073] Please see Figure 2 and Figure 3 This embodiment takes a scenario of accidental triggering caused by a pet cat in a home environment as an example to illustrate the practical application effect of the amplitude prediction step in the method of the present invention.

[0074] Scenario Setting: A 4m x 4m living room, with the lighting control device of this invention installed in the center of the ceiling. A cat weighing approximately 4kg is kept in the house. When the cat walks around the living room, electromagnetic waves emitted by the millimeter-wave radar strike the cat, generating a reflected echo. A PIR sensor detects changes in the thermal radiation emitted by the cat. In step S1, the MCU processor acquires the amplitude value of the radar echo signal and calculates its arithmetic mean to obtain the amplitude average. Because a cat's radar cross-section (RCS) is much smaller than a human's, based on typical radar cross-section data, a cat's RCS is approximately 0.01~0.05 m², while a human's RCS is approximately 0.5~1.0 m², a difference of about an order of magnitude. Significantly less than the preset amplitude threshold ,in The calibration is typically performed near the amplitude value corresponding to the minimum RCS of the human body, for example, taking the amplitude value corresponding to the RCS of a child's human body model at a distance of 3m. The MCU processor determines this. < If it is identified as a small pet, the subsequent steps will be terminated and the lighting will not be triggered.

[0075] If a cat is large, such as a large cat weighing over 10kg, its RCS may be close to that of a small human. Slightly higher The system cannot be excluded by the amplitude prediction step. At this point, the system enters the complete discrimination process from steps S2 to S4. When a large cat walks around the room, the phase change detected by the radar is caused by the dog's body movement. When the dog walks, its limbs move alternately, and its center of gravity swings periodically, causing the radar phase to fluctuate at high frequency and irregularly. To obtain the specific value of this fluctuation frequency, the radar phase signal when the cat walks normally in the room is collected, and the collected phase signal is subjected to fast Fourier transform spectrum analysis. The spectrum analysis results show that the radar phase fluctuation frequency caused by the cat's movement is mainly distributed in the range of 1~3Hz. This frequency range is significantly higher than the human breathing frequency of 0.2~0.33Hz. Therefore, the frequency characteristics can be used to preliminarily distinguish between cat activity and human breathing. At the same time, the thermal radiation change detected by the PIR sensor is caused by the difference between the cat's surface temperature and the ambient temperature. During the movement, different parts of the cat's body alternately enter and leave the detection field of the PIR sensor, and the output voltage of the PIR sensor also shows high-frequency and irregular fluctuations. Because the movement frequency and thermal radiation change frequency of a cat are much higher than the human breathing frequency (0.2~0.33Hz), and there is no physical coupling relationship between them like that of human breathing, the radar phase change and PIR sensor voltage change have no predictable direction. To obtain the time lag relationship between radar phase change and PIR sensor voltage change, cross-correlation analysis was performed on the radar phase signal and PIR sensor voltage signal synchronously collected during the cat's rapid running. The cross-correlation analysis results show that there is a significant time difference between the moment of drastic radar phase change and the moment of corresponding PIR sensor voltage change, which is approximately 100~200ms. That is, the PIR sensor voltage change lags behind the radar phase change by about 100~200ms. Therefore, the two changes often occur in opposite directions. Within a preset time window of 3 seconds, 60 sampling moments were analyzed. and Statistical analysis was conducted. To obtain a specific value for the consistency rate in the pet cat scenario, 10 independent tests were performed on a cat weighing approximately 4kg in the aforementioned 4m×4m living room environment. Each test lasted 3 seconds, and data were collected at 60 sampling points. and The data was used to calculate the consistency rate for each test. The consistency rates for the 10 tests were 32%, 28%, 35%, 30%, 38%, 33%, 29%, 36%, 31%, and 34%, with an average of 32.6%. All these rates were below 40%, far below the preset threshold of 70%. The system determined this to be a false trigger and did not output a lighting control signal.

[0076] Example 3

[0077] Please see Figure 2 and Figure 3This embodiment uses a scenario of false triggering caused by the fluttering of curtains in a home environment as an example to illustrate the filtering effect of the method of the present invention on complex interference that has a signal but is not from a human source.

[0078] Scenario: The window is open, and the outdoor wind force is approximately level 3-4. The wind blows the curtains, causing them to flutter periodically for about 1-2 seconds. As the curtains flutter, millimeter-wave radar detects the echo signal caused by the mechanical movement of the curtains. The curtain fabric, typically made of polyester or cotton, has a certain electromagnetic wave reflectivity. Its back-and-forth movement under the wind causes a periodic change in the radar reflection path length, resulting in a change in the radar echo phase. Simultaneously, a PIR sensor detects a sudden change in thermal radiation caused by the temperature difference between the curtain surface and room temperature. While the curtains are normally at indoor temperature, when blown by the wind, the air exchange between the inside and outside of the curtain causes a rapid change in surface temperature, especially when the curtains are near the window where there is a temperature difference between the window glass and room temperature. The surface temperature of the curtains quickly approaches the window glass temperature after being blown up; this temperature change is detected by the PIR sensor as a sudden change in thermal radiation.

[0079] In step S2, the MCU processor calculates the radar phase change direction respectively. and the direction of voltage change of the passive infrared sensor When the curtains flutter outwards, they move closer to the window, causing the radar reflection path to shorten or lengthen, depending on the curtains' initial position relative to the radar and the phase value. The temperature changes in one direction; simultaneously, the direction of change in the curtain's surface temperature depends on the relative relationship between the curtain's initial temperature and the outside air temperature. If the outside air temperature is lower than the indoor temperature, the curtain's surface temperature decreases after it is blown up, reducing the heat radiation detected by the PIR sensor, and thus decreasing the PIR sensor's output voltage. The radar phase change direction is often opposite to the PIR sensor voltage change direction; one rises while the other falls, or vice versa. In rare cases, they may change in the same direction, but because there is no physical coupling between the mechanical motion of the curtain swaying and the change in thermal radiation as in human breathing, their changes are asynchronous in time. The radar detects the instantaneous speed of the mechanical sway, which is related to the swaying speed, while the PIR sensor detects the cumulative change in thermal radiation, which is related to the integral of the temperature difference. To obtain the time lag relationship between the two, cross-correlation analysis was performed on the radar phase signal and the PIR sensor voltage signal synchronously acquired when the curtain swayed. The cross-correlation analysis results show that there is a significant time difference between the moment of drastic radar phase change and the moment of corresponding change in PIR sensor voltage, which is approximately 200-500 ms. That is, the PIR sensor voltage change lags behind the radar phase change by about 200-500 ms. This time lag is greater than the time lag of the two signals in the human breathing scenario, and therefore can be used to distinguish between curtain swaying and human breathing.

[0080] Within a preset time window of 3 seconds, 60 sampling moments were performed. and Statistical analysis was conducted. Because the directions of change for both are often opposite or there is a time lag, to obtain a specific numerical value for the consistency ratio in the curtain swaying scenario, 10 independent tests were performed in the aforementioned 4m×4m room environment under wind force 3-4 conditions. Each test lasted 3 seconds, and 60 sampling times were collected. and Data, calculate the consistency ratio for each test. The consistency rates for the 10 tests were 42%, 38%, 45%, 40%, 48%, 43%, 39%, 46%, 41%, and 44%, with an average of 42.6%, all below 50% and far below the preset threshold of 70%. The system determined this to be a false trigger and did not output a lighting control signal.

[0081] Example 4

[0082] Please see Figure 3 This embodiment uses device installation and spatial partitioning as an example for explanation.

[0083] The device is installed in the center of the room's ceiling. The installation height is approximately 2.8m above the ground. The radar antenna's elevation angle coverage range is -60° to +60° relative to the antenna's normal direction, and the PIR sensor's detection field of view is approximately 110° × 90°. Together, they cover the entire room area.

[0084] The room is divided into four lighting zones, each corresponding to a set of independently driven LED lights. Each zone can independently adjust its color temperature and illuminance. Zone A corresponds to the front left area of ​​the room, Zone B to the front right area, Zone C to the rear left area, and Zone D to the rear right area. Each zone is equipped with a set of LED ceiling lights or downlights, controlled via an independent PWM channel.

[0085] Millimeter-wave radar module detects the azimuth angle of the human body in real time. and distance The radar obtains the azimuth angle through phase comparison angle measurement and utilizes the phase difference of the echo signals received by two receiving antennas. Let the distance between the two receiving antennas be... The radar's operating wavelength is The target azimuth is The phase difference between the received echo signals from the two receiving antennas is Then the phase difference and the target azimuth angle satisfy the following: From this, the target azimuth angle can be calculated. Radar obtains distance by using intermediate frequency (IF) signals, based on... calculate .

[0086] MCU processor according to and Determine the lighting zone where the human figure is located. The determination rule is as follows: Divide the room plane into four quadrants with the installation location of the device as the origin. The first quadrant, i.e., The 0°~90° range corresponds to partition B, the second quadrant, i.e. The 90°~180° range corresponds to partition D, the third quadrant, i.e. The area corresponding to -180° to -90° is zone C, the fourth quadrant, i.e. Partition A corresponds to -90° to 0°. Also, based on distance... To determine if a person is outside the room's boundaries, if... If the distance is half the diagonal of the room, the person may not be in the room or may be at the edge of the room. In this case, the lighting will not be triggered or the nearest zone will be triggered.

[0087] Example 5

[0088] Please see Figure 1 This embodiment describes the workflow of the lighting control false triggering filtering system based on multi-sensor fusion of the present invention from the perspective of the overall system logic.

[0089] After system power-on initialization, the signal acquisition unit operates continuously at a sampling frequency of 20Hz. The analog-to-digital converter inside the signal acquisition unit converts the analog intermediate frequency signal output from the millimeter-wave radar module into a digital signal. After range-dimensional FFT processing, the target range gate information is obtained, and the phase value is extracted from the complex baseband signal corresponding to the range gate. Simultaneously, the analog voltage signal output by the PIR sensor is converted into a digital voltage value. The signal acquisition device will and Data is transmitted to the direction comparator via the internal data bus.

[0090] Direction comparator received and Then, the configuration change direction comparison algorithm begins execution. The direction comparator internally has two registers, one storing the previous time step's... and Value. Whenever a new and Upon arrival, the direction comparator calculates... and and the calculation results and The data is transmitted to the statistics unit separately.

[0091] The statistician has a circular buffer of length 60 to store the data from the most recent 60 sampling times. and Value. Whenever a new and Upon arrival, the statistician updates the buffer, iterates through all 60 data sets, and performs statistics. and Number of times the same value is taken Calculate the consistency ratio and will The value is transmitted to the decision maker.

[0092] The determiner will receive the consistency ratio With preset threshold Compare. If The detector outputs a high-level lighting control signal to the LED driver circuit; if The detector maintains a low-level output and does not trigger the lighting.

[0093] After receiving a high-level control signal, the LED driver circuit outputs drive current to the PWM channel of the corresponding zone according to the zone selection signal provided by the zone controller, thus illuminating the LED lights in that zone. The zone controller uses the azimuth angle provided by the millimeter-wave radar module... Distance to target The human body is located in the partition according to the partition judgment rules described in Example 4.

[0094] The entire system executes the above process in a 50ms cycle to achieve real-time detection of human presence and lighting control.

[0095] Example 6

[0096] Please see Figure 2 This embodiment uses a complex situation in an office setting where multiple people are present simultaneously as an example to illustrate the robustness of the method of the present invention in practical applications.

[0097] Setting: A 6m x 5m office is equipped with two sets of lighting control devices from this invention, installed on the ceiling at opposite ends of the room, covering the entire office area. There are two employees in the office: Employee A is sitting at a workstation using a computer, while Employee B is moving between the filing cabinet and the workstation.

[0098] In this multi-person scenario, each device operates independently, covering its corresponding half-zone. For the half-zone where employee A is located, the device detects the breathing signal of a stationary human. The radar phase exhibits a periodic change of 0.2Hz, which corresponds to the lower limit of a human's resting respiratory rate of 12 breaths per minute (12 breaths / 60 seconds = 0.2Hz), a typical value within the normal human breathing rate range. The PIR sensor voltage shows a slight fluctuation at the same frequency. Within a preset time window of 3 seconds, a total of 60 sampling moments are collected. Approximately 48 of these sampling moments show the same direction of change for both signals, indicating a consistency ratio of... The calculation process is as follows: statistical analysis of 60 sampling times. and Number of times the same value is taken , ,when hour, The device detects the presence of a person and keeps the lighting on in that half-segment. In the half-segment where employee B is located, the device detects a moving human signal; the radar phase shows a large and rapid change, and the PIR sensor output voltage shows a large thermal radiation change. Both changes are in the same direction. Within a preset time window of 3 seconds, a total of 60 sampling moments are collected. Approximately 51 of these sampling moments show the two signals changing in the same direction, indicating a consistency ratio of... The device detects the presence of someone and keeps the lighting on in that half of the area.

[0099] When employee B moves from their own half of the radar field to employee A's half, the detection results of both devices change dynamically. The device in employee B's original half detects the target's departure, and the radar phase and PIR sensor voltages gradually decrease until they disappear, resulting in a decrease in the consistency ratio. If the light level drops below 50%, the lighting in that half-zone will be turned off after a 3-second delay; if employee B enters the half-zone, the device will detect a new target and keep the lighting on.

[0100] In summary, this invention acquires the phase value of the millimeter-wave radar echo signal and the output voltage value of the PIR sensor, calculates the direction of change of both at adjacent moments, and statistically analyzes the proportion of the two signals with the same direction of change within a preset time window. When this proportion reaches a preset threshold, a lighting control signal is output. This invention changes the judgment criterion from the instantaneous feature value magnitude of existing technologies to the consistency of the direction of change in the time domain, fully utilizing the physical law that the radar phase change caused by human respiration and the thermal radiation change of the PIR sensor are necessarily consistent in direction of change. Compared with existing technologies, this invention does not rely on complex feature extraction and decision tree models, but only requires subtraction operations and sign judgment; it can simultaneously filter multiple types of interference sources such as pet activity, curtain swaying, and air conditioning airflow; it can maintain continuous lighting when the human body is stationary; and the hardware cost is controllable, the computational complexity is extremely low, and it is suitable for various low-cost embedded lighting control platforms.

[0101] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments under the guidance of the present invention without departing from the spirit and scope of the present invention. All of these variations are within the protection scope of the present invention.

Claims

1. A method for filtering false triggering in lighting control based on multi-sensor fusion, characterized in that, include: The phase value of the millimeter-wave radar echo signal and the output voltage value of the passive infrared sensor are acquired simultaneously, and the sampling frequency of both is the same. Based on the phase value, calculate the first direction of change of the phase value at adjacent time points; based on the output voltage value, calculate the second direction of change of the output voltage value at adjacent time points; Within a preset time window, the proportion of sampling times where the first change direction and the second change direction are the same to the total number of sampling times within the time window is counted. When the ratio is greater than or equal to a preset threshold, a lighting control signal is output.

2. The lighting control false triggering filtering method based on multi-sensor fusion as described in claim 1, characterized in that, Based on the phase value, calculate the first direction of change of the phase value at adjacent time points; Based on the output voltage value, calculate the second direction of change of the output voltage value at adjacent time points, including: Calculate the difference between the phase value or output voltage value at the current moment and the phase value or output voltage value at the previous moment. If the difference is positive, the corresponding first change direction or second change direction is positive; if the difference is negative, the corresponding first change direction or second change direction is negative.

3. The lighting control false triggering filtering method based on multi-sensor fusion as described in claim 1, characterized in that, The length of the preset time window is configured to cover at least one human respiratory cycle; the sampling time interval between adjacent moments is configured to include at least 20 sampling moments within the preset time window.

4. The lighting control false triggering filtering method based on multi-sensor fusion as described in claim 1, characterized in that, When synchronously acquiring the phase value of the millimeter-wave radar echo signal and the output voltage value of the passive infrared sensor, the process includes: acquiring the amplitude value of the millimeter-wave radar echo signal, taking the arithmetic mean of the amplitude values ​​at multiple consecutive sampling times to obtain the amplitude average, and terminating the lighting control false trigger filtering process when the amplitude average is less than a preset amplitude threshold.

5. The lighting control false triggering filtering method based on multi-sensor fusion as described in claim 1, characterized in that, Also includes: After outputting the lighting control signal, the spatial zone where the human body is located is determined based on the human body azimuth angle and target distance measured by the millimeter-wave radar, and a drive signal is output to the lighting fixtures in the corresponding zone.

6. The lighting control false triggering filtering method based on multi-sensor fusion as described in claim 5, characterized in that, The human body azimuth angle is obtained by phase comparison angle measurement method, which uses the phase difference of the echo signal received by two receiving antennas, and calculates the target angle according to the correspondence between the phase difference and the target azimuth angle.

7. The lighting control false triggering filtering method based on multi-sensor fusion as described in claim 5, characterized in that, Based on the human body azimuth angle and target distance measured by the millimeter-wave radar, the spatial zone in which the human body is located is determined, including: Using the millimeter-wave radar installation location as the origin, the room plane is divided into four quadrants. The zone where the human body is located is determined based on the quadrant to which the azimuth angle belongs and the target distance.

8. The lighting control false triggering filtering method based on multi-sensor fusion as described in claim 7, characterized in that, Also includes: The target distance is compared with a preset distance threshold. If the distance is greater than the preset distance threshold, it is determined that the human body is not in the room or is at the edge of the room, and the lighting is not triggered or the nearest partition is triggered.

9. The lighting control false triggering filtering method based on multi-sensor fusion as described in claim 1, characterized in that, When the aforementioned lighting control false trigger filtering method based on multi-sensor fusion is run in parallel in multiple different areas, each area independently performs human presence detection. The lighting in the area where the target is detected leaving is turned off after a delay, while the lighting in the area where the target is detected entering remains on.

10. A lighting control false triggering filtering system based on multi-sensor fusion, used to implement the lighting control false triggering filtering method based on multi-sensor fusion as described in any one of claims 1-9, characterized in that, include: The signal acquisition unit has a built-in analog-to-digital converter and is used to simultaneously acquire the phase value of the millimeter-wave radar echo signal and the output voltage value of the passive infrared sensor. A direction comparator is configured with a change direction comparison algorithm to calculate the change direction of the phase value and the output voltage value at adjacent time points; A statistical analyzer is used to count the proportion of sampling times with the same direction of change within the preset time window to the total number of sampling times within the time window; The detector is used to output a lighting control signal when the ratio is greater than or equal to a preset threshold. A zone controller, connected to the determiner, is used to determine the spatial zone where the human body is located based on the human body azimuth angle and target distance measured by the millimeter-wave radar, and output drive signals to the lighting fixtures of the corresponding zone. The output of the signal acquisition unit is connected to the input of the direction comparator, the output of the direction comparator is connected to the input of the statistician, the output of the statistician is connected to the input of the decision-maker, and the output of the decision-maker is connected to the input of the partition controller.

Images