CN112423434A - Power supply ripple based multi-path LED lamp brightness bypass detection device and method - Google Patents

Abstract

The invention discloses a device and a method for detecting a brightness bypass of a multi-channel LED lamp based on power supply ripple. The invention includes an LED lamp, an LED lamp drive circuit, a power supply ripple signal acquisition circuit and a signal processor; the LED lamp drive circuit is connected with the LED lamp and controls the LED lamp, and several LED lamp drive circuits connected with the LED lamp are connected in parallel with the power supply A multi-channel LED light is formed on the top of the circuit, and the power supply ripple signal acquisition circuit is connected to the power supply and collects the power supply ripple generated by several LED lamp drive circuits to the power supply. The power supply ripple signal acquisition circuit transmits the collected power supply ripple through the USB interface. To the signal processor, the signal processor can obtain the brightness of each LED lamp after processing the mixed ripple signal. The invention is a bypass access type device, which realizes the brightness detection of multi-channel LED lamps under the original circuit, and has simple structure, low cost and certain universality.

Description

Device and method for brightness bypass detection of multi-channel LED lamps based on power supply ripple

technical field

The invention relates to a multi-channel LED lamp brightness detection method, in particular to a multi-channel PWM-modulated LED lamp brightness bypass detection device and method based on power supply ripple.

Background technique

LEDs are solid-state electric light sources and semiconductor lighting devices. Its electrical characteristics are highly discrete. It has the characteristics of small size, high mechanical strength, low power consumption, long life, easy adjustment and control, and no pollution. PWM is a pulse width modulation signal, where the pulse width represents the time of the high level of the pulse, PWM dimming is turned on and off by the PWM wave LED to change the conduction time of the forward current to achieve the effect of brightness adjustment, because the human eye Insensitive to brightness flicker, when the PWM wave frequency is greater than 100Hz, the average brightness of the LED observed by the human eye.

At present, most of the brightness detection for LED lights is based on sensors, and the disadvantages of using sensors are becoming more and more obvious. It is more common that sensors are easily affected by the environment, their performance is unstable in harsh environments, and they are easily affected by external dust. At the same time, when designing the overall LED circuit, the sensor needs to be directly installed in the driving circuit, which is not conducive to the replacement of equipment. Therefore, it is necessary to develop a more intelligent and convenient bypass detection that can realize brightness detection without changing the structure of the original circuit. method and system, but no relevant reports have been found so far.

SUMMARY OF THE INVENTION

In view of the deficiencies in the background technology, the design purpose of the present invention is to provide a brightness bypass detection method of a multi-channel PWM modulated LED lamp based on power supply ripple. The different power supply ripples generated by the disturbance of different states of the lamp to the LED lamp drive circuit regularly change to indirectly detect the brightness corresponding to each LED lamp.

In order to achieve the above object, the technical scheme adopted by the present invention is:

1. A multi-channel LED lamp brightness bypass detection device based on power supply ripple

The device includes an LED lamp, an LED lamp drive circuit, a power supply ripple signal acquisition circuit and a signal processor; the LED lamp drive circuit is connected with the LED lamp and controls the LED lamp, and a plurality of LED lamp drive circuits connected with the LED lamp are connected in parallel to the power supply A multi-channel LED lamp is formed. The power ripple signal acquisition circuit is connected to the power supply and collects the power ripple generated by several LED lamp drive circuits on the power supply. The power ripple signal acquisition circuit transmits the collected power ripple through the CH340USB transfer interface. to the signal processor, which is a computer terminal.

The LED lamp drive circuit includes a drive circuit power supply port, a drive circuit output port, an LED lamp positive and negative interface, a PWM signal duty cycle control circuit and an LED lamp working circuit, and the PWM signal duty cycle control circuit is connected with the LED lamp working circuit. , the LED lamp working circuit is connected to the LED lamp through the drive circuit output port and the LED lamp positive and negative interfaces in turn, and the LED lamp drive circuit is connected to the power supply through the drive circuit power port; the external knob or infrared remote control controls the PWM signal duty cycle control circuit. The duty ratio of the high level and the low level in the PWM wave makes the working time of the LED lamp working circuit be controlled to achieve the effect of adjusting the brightness of the LED lamp.

The power supply ripple signal acquisition circuit includes a filter circuit, a load resistor, a sampling resistor, an AD sampling module and an amplifying circuit; a series-connected load resistor and a sampling resistor are connected between the power supplies, and both ends of the sampling resistor are connected to the amplifying circuit, and the amplifying circuit The weak differential signal at both ends of the sampling resistor is converted and amplified and output. The amplification circuit is connected to the filter circuit, capacitor, and AD sampling module in turn. The AD sampling module transmits the collected power ripple to the signal processor through the CH340 serial port to USB chip.

2. A brightness bypass detection method for multi-channel LED lamps based on power supply ripple

A mixed ripple signal is input into the signal processor, and the method for processing the mixed ripple signal includes the following steps:

1) Perform j-layer decomposition on the mixed ripple signal, and obtain the wavelet coefficients and scale coefficients of the mixed ripple signal after the decomposition and reconstruction;

Described step 1) is specifically:

Using the complex wavelet function basis of Hilbert transform, the mixed ripple signal is decomposed in j layers, and the decomposed signal Y is obtained after the decomposition and reconstruction. From the decomposed signal Y, the wavelet coefficients and scale coefficients of the mixed ripple signal can be obtained:

where d _i is the i-th wavelet coefficient in the decomposed signal Y, and c _j is the j-th scale coefficient in the decomposed signal Y.

2) Select wavelet coefficients and scale coefficients, establish a signal decomposition matrix, and perform dimension reduction and reconstruction on the signal decomposition matrix to obtain a multi-channel signal

Described step 2) is specifically:

Perform dimensionality reduction and reconstruction on the signal decomposition matrix Q to obtain multi-channel signals

Select wavelet coefficients d ₁ , d ₂ ,...,d _j and scale coefficient c _j , where d ₁ represents the first wavelet coefficient, d ₂ represents the second wavelet coefficient, d _j represents the jth wavelet coefficient, Establish the signal decomposition matrix Q=[c _j ,d ₁ ,d ₂ ,...,d _j ] ^T , calculate the eigenvalues of the covariance matrix S Λ=[λ ₁ ,λ ₂ ,...,λ _j+1 ] and the corresponding feature vector V=[ω ₁ ,ω ₂ ,...,ω _j+1 ]:

u₁表示尺度系数c_j的均值，u₂表示小波系数d₁的均值，u_j+1表示小波系数d_j的均值，T表示转置操作，λ₁表示协方差矩阵S中第1个特征值，λ_j+1表示协方差矩阵S中第j+1个特征值，ω₁表示协方差矩阵S中第1个特征向量，ω_j+1表示协方差矩阵S中第j+1个特征向量；in,

u ₁ represents the mean value of the scale coefficient c _j , u ₂ represents the mean value of the wavelet coefficient d ₁ , u _j+1 represents the mean value of the wavelet coefficient d _j , T represents the transpose operation, λ ₁ represents the first feature in the covariance matrix S value, λ _j+1 represents the j+1 eigenvalue in the covariance matrix S, ω ₁ represents the first eigenvector in the covariance matrix S, and ω _j+1 represents the j+1 eigenvalue in the covariance matrix S vector;

Sort the eigenvalues Λ=[λ ₁ ,λ ₂ ,...,λ _j+1 ] from large to small, and select N-1 eigenvalues in turn, where N represents the number of LED lights, from the signal decomposition matrix Select the N-1 signal components of the eigenvalue serial number corresponding to the N-1 eigenvalues from Q and the mixed ripple signal to form a multi-channel signal

进行白化处理获得白化多通道信号

对分离矩阵M随机赋初始值，利用白化多通道信号

将分离矩阵M不断迭代至前一个分离矩阵M_k-1与当前的分离矩阵M_k的误差满足∑∑|M_k-M_k-1|≤σ，获得最终的分离矩阵M；其中，σ为误差，满足0<σ<1，k为分离矩阵M迭代的次数，∑∑|M_k-M_k-1|表示分离矩阵M_k与分离矩阵M_k-1作差后获得中间矩阵，对中间矩阵中所有元素进行求和后获得结果；计算获得原始多通道信号

其中O(t)＝[O₁(t),O₂(t),···,O_N(t)]，O₁(t)表示第1个原始通道信号，N表示总共有N路LED灯；3) Convert the multi-channel signal

Perform whitening processing to obtain whitened multi-channel signals

Randomly assign initial values to the separation matrix M, and use whitening multi-channel signals

The separation matrix M is continuously iterated until the error between the previous separation matrix M _k-1 and the current separation matrix M _k satisfies ∑∑|M _k -M _k-1 |≤σ, and the final separation matrix M is obtained; where σ is Error, satisfying 0<σ<1, k is the number of iterations of the separation matrix M, ∑∑|M _k -M _k-1 | means that the separation matrix M _k and the separation matrix M _k-1 are different to obtain the intermediate matrix, The result is obtained after summing all elements in the matrix; the calculation obtains the original multi-channel signal

Where O(t)=[O ₁ (t),O ₂ (t),...,ON (t)], O ₁ (t) represents the first original channel signal, and _N represents a total of N LEDs lamp;

Described step 3) is specifically:

采用以下公式进行白化处理获得白化多通道信号

multi-channel signal

Use the following formula to perform whitening processing to obtain a whitened multi-channel signal

Randomly assign an initial value to the separation matrix M, and use the following formula to iterate the separation matrix M continuously until the error between the previous separation matrix M _k-1 and the current separation matrix M _k satisfies |M _k -M _k-1 |≤σ, obtain The final separation matrix M;

表示累积分布函数，

表示概率分布函数，m_k-1表示第k-1次迭代获得的分离矩阵M中某一行，y表示白化多通道信号

中与m_k-1相同的行序号的一行，m_k表示第k次迭代获得的分离矩阵M中与m_k-1相同的行序号的一行，M_k-1表示第k-1次迭代计算所得的分离矩阵M，M_k表示第k次迭代计算所得的分离矩阵M，E{}表示期望操作，T表示转置操作；in,

represents the cumulative distribution function,

represents the probability distribution function, m _k-1 represents a row in the separation matrix M obtained by the k-1 iteration, and y represents the whitened multi-channel signal

A row with the same row number as m _k-1 , m _k represents a row with the same row number as m k _- 1 in the separation matrix M obtained by the kth iteration, M _k-1 represents the k-1th iteration calculation The obtained separation matrix M, M _k represents the separation matrix M obtained by the k-th iteration calculation, E{} represents the desired operation, and T represents the transpose operation;

Finally, calculate the original multi-channel signal

4) Convert each sub-original channel signal in the original multi-channel signal O(t) into a square wave signal, calculate the duty cycle of each square wave signal, and form a total duty cycle q=[q ₁ ,q ₂ , ...,q _N ], from the total duty cycle q, the brightness of N LED lights L=[L ₁ , L ₂ , ..., L _N ], where the duty cycle corresponds to the brightness of the LED lights in equal proportions, When the duty cycle is 1, the LED light is the brightest working state; when the duty cycle is 0, the LED light is completely off.

The effective benefit that the present invention has is:

The device and method for detecting the brightness of a multi-channel PWM modulated LED lamp brightness based on the power supply ripple designed by the present invention is different from the traditional sensor-based method for detecting the brightness of an LED lamp. The different power ripples generated by the disturbance of the drive circuit change regularly to indirectly detect the brightness corresponding to each LED lamp.

A multi-channel PWM modulation LED lamp brightness bypass detection device based on power supply ripple designed by the present invention, as a bypass access system, realizes multi-channel PWM modulation without changing the original circuit. LED brightness detection, and the system structure is simple, the implementation cost is low, the use range is wide, and it has a certain universality. Combined with the current ubiquitous power Internet of things, sensorless detection based on power supply ripple provides another idea.

Description of drawings

Fig. 1 is the module schematic diagram of the embodiment of the present invention;

2 is a structural block diagram of an embodiment of the present invention;

Fig. 3 is a schematic diagram of a PWM signal duty cycle control circuit interface;

Figure 4 is a circuit diagram of a power supply ripple signal acquisition circuit;

Fig. 5 is the signal processing algorithm flow chart of the signal processor;

In the figure: 1. LED light, 2. LED light drive circuit, 3. Power supply ripple signal acquisition circuit, 4. Signal processor, 5. Power supply, 6. Drive circuit power supply port, 7. PWM signal duty cycle control circuit , 8, drive circuit output port, 9, LED light positive and negative interface, 10, LED light working circuit, 11, filter circuit, 12, load resistance, 13, sampling resistance, 14, AD sampling module, 15, amplifier circuit.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings.

As shown in FIG. 1 and FIG. 2 , the present invention includes an LED lamp 1 , an LED lamp driving circuit 2 , a power supply ripple signal acquisition circuit 3 and a signal processor 4 ; the LED lamp driving circuit 2 is connected to the LED lamp 1 and controls the LED lamp 1 , a number of LED lamp drive circuits 2 connected to the LED lamp 1 are connected to the power supply 5 to form a multi-channel LED lamp, and the power supply ripple signal acquisition circuit 3 is connected to the power supply 5 and collects several LED lamp drive circuits 2 to the power supply 5 The generated power ripple, the power ripple signal acquisition circuit 3 transmits the collected power ripple to the signal processor 4 through the CH340USB transfer interface, and the signal processor 4 is a computer terminal.

The LED lamp drive circuit 2 includes a drive circuit power supply port 6, a drive circuit output port 8, an LED lamp positive and negative interface 9, a PWM signal duty cycle control circuit 7 and an LED lamp work circuit 10, and the PWM signal duty cycle control circuit 7 and LED The lamp working circuit 10 is connected, the LED lamp working circuit 10 is connected to the LED lamp 1 through the drive circuit output port 8 and the LED lamp positive and negative interfaces 9 in turn, and the LED lamp drive circuit 2 is connected to the power source 5 through the drive circuit power port 6; the external knob Or infrared remote control the duty ratio of high level and low level in the PWM signal duty ratio control circuit 7 , so that the working time of the LED lamp working circuit 10 can be controlled to achieve the effect of adjusting the brightness of the LED lamp 1 .

The duty ratio control circuit of the PWM signal selects a square wave rectangular wave signal generating circuit with an adjustable duty ratio that is readily available on the market. In the implementation of the present invention, a PWM pulse module with liquid crystal display from telesky company is selected, and three PWM signals The generator outputs equal amplitude PWM waves of 50Hz, 200Hz and 600Hz to the working circuit of the LED light.

As shown in FIG. 4 , the power supply ripple signal acquisition circuit 3 includes a filter circuit 11 , a load resistor 12 , a sampling resistor 13 , an AD sampling module 14 and an amplifying circuit 15 ; the power supply 5 is connected with a series-connected load resistor 12 and a sampling resistor 13 . , the two ends of the sampling resistor 13 are connected to the amplifying circuit 15, and the amplifying circuit 15 converts and amplifies the weak differential signal at both ends of the sampling resistor 13 and outputs it. The module 14 transmits the collected power ripple to the signal processor 4 through the CH340 serial port to USB chip.

In the specific implementation, the load resistance 12 is selected as 120Ω, the sampling resistance 13 is selected as 1Ω, the non-inverting input amplifiers A1 and A2 based on OP27 are selected, and the non-inverting input of the non-inverting input amplifiers A1 and A2 realizes the amplification of the input resistance. The differential amplifier A3 based on INA105 converts the differential input signal into a single-ended output signal, which improves the common mode rejection ratio of the amplifier circuit 15, that is, suppresses the common mode, amplifies the differential mode gain, and simultaneously suppresses noise. According to the circuit principle, calculate the differential mode gain of the amplifier:

A _vd =1+2R ₃ /R _X

Where Avd represents the voltage magnification. In the example of the present invention, a sliding varistor with a resistance R3 of 5kΩ and a resistance R4 of 20kΩ are selected, so the minimum magnification is 9 times.

The filter circuit 11 uses a second-order low-pass filter based on OP27 to filter out high-frequency noise and make the waveform smoother. The direct relationship between the cutoff frequency f of the second-order RC low-pass filter and the second-order RC low-pass filter is as follows:

The resistor R5=R6=5kΩ and the capacitor C1=C2=0.01uF are selected in this example, and the cut-off frequency f of the second-order RC low-pass filter is calculated as 1190kHz.

The AD sampling module 14 selects the internal ADC resources of ST's STM32F103 series single-chip microcomputer. The AD sampling module 14 transfers the acquired sampling data to the memory through the internal DMA of the STM32, reducing the burden on the CPU and enhancing the stability and accuracy of AD sampling. The sampled data, that is, the mixed ripple signal, is transmitted to the signal processor 4 through the UART serial communication through the onboard CH340 serial port to the USB chip. Frequency Theorem.

As shown in FIG. 5 , the mixed ripple signal is input to the signal processor 4, and the method for processing the mixed ripple signal includes the following steps:

1) Perform j-layer decomposition on the mixed ripple signal, and obtain the wavelet coefficients and scale coefficients of the mixed ripple signal after the decomposition and reconstruction;

Step 1) is specifically:

Using the complex wavelet function basis of Hilbert transform, the mixed ripple signal is decomposed in j layers, and the decomposed signal Y is obtained after the decomposition and reconstruction. From the decomposed signal Y, the wavelet coefficients and scale coefficients of the mixed ripple signal can be obtained:

where d _i is the i-th wavelet coefficient in the decomposed signal Y, and c _j is the j-th scale coefficient in the decomposed signal Y.

In the specific implementation, j=4, the corresponding signal components obtained are:

2) Select wavelet coefficients and scale coefficients, establish a signal decomposition matrix, and perform dimension reduction and reconstruction on the signal decomposition matrix to obtain a multi-channel signal

Step 2) is specifically:

Perform dimensionality reduction and reconstruction on the signal decomposition matrix Q to obtain multi-channel signals

Select wavelet coefficients d ₁ , d ₂ ,...,d _j and scale coefficient c _j , where d ₁ represents the first wavelet coefficient, d ₂ represents the second wavelet coefficient, d _j represents the jth wavelet coefficient, Establish the signal decomposition matrix Q=[c _j ,d ₁ ,d ₂ ,...,d _j ] ^T , calculate the eigenvalues of the covariance matrix S Λ=[λ ₁ ,λ ₂ ,...,λ _j+1 ] and the corresponding feature vector V=[ω ₁ ,ω ₂ ,...,ω _j+1 ]:

u₁表示尺度系数c_j的均值，u₂表示小波系数d₁的均值，u_j+1表示小波系数d_j的均值，T表示转置操作，λ₁表示协方差矩阵S中第1个特征值，λ_j+1表示协方差矩阵S中第j+1个特征值，ω₁表示协方差矩阵S中第1个特征向量，ω_j+1表示协方差矩阵S中第j+1个特征向量；in,

u ₁ represents the mean value of the scale coefficient c _j , u ₂ represents the mean value of the wavelet coefficient d ₁ , u _j+1 represents the mean value of the wavelet coefficient d _j , T represents the transpose operation, λ ₁ represents the first feature in the covariance matrix S value, λ _j+1 represents the j+1 eigenvalue in the covariance matrix S, ω ₁ represents the first eigenvector in the covariance matrix S, and ω _j+1 represents the j+1 eigenvalue in the covariance matrix S vector;

Sort the eigenvalues Λ=[λ ₁ ,λ ₂ ,...,λ _j+1 ] from large to small, and select N-1 eigenvalues in turn, where N represents the number of LED lights, from the signal decomposition matrix Select the N-1 signal components of the eigenvalue serial number corresponding to the N-1 eigenvalues from Q and the mixed ripple signal to form a multi-channel signal

In the specific implementation, the signal decomposition matrix Q=[c ₄ , d ₁ , d ₂ , d ₃ , d ₄ ] ^T , first calculate the mean vector of each signal component of the signal decomposition matrix Q:

where L is the length of the signal component.

The eigenvalues Λ=[λ ₁ ,λ ₂ ,...,λ ₅ ] of the covariance matrix S and the corresponding eigenvectors V=[ω ₁ ,ω ₂ ,...,ω ₅ ]; the eigenvalues Λ= [λ ₁ ,λ ₂ ,...,λ ₅ ] are sorted from large to small, two eigenvalues are selected in turn, and two eigenvalue numbers corresponding to the two eigenvalues are selected from the signal decomposition matrix Q The signal components and the mixed ripple signal together form a multi-channel signal

进行白化处理获得白化多通道信号

对分离矩阵M随机赋初始值，利用白化多通道信号

将分离矩阵M不断迭代至前一个分离矩阵M_k-1与当前的分离矩阵M_k的误差满足∑∑|M_k-M_k-1|≤σ，获得最终的分离矩阵M；其中，σ为误差，满足0<σ<1，k为分离矩阵M迭代的次数，∑∑|M_k-M_k-1|表示分离矩阵M_k与分离矩阵M_k-1作差后获得中间矩阵，对中间矩阵中所有元素进行求和后获得结果；计算获得原始多通道信号

其中O(t)＝[O₁(t),O₂(t),···,O_N(t)]，O₁(t)表示第1个原始通道信号，即第1路LED灯所产生的原始通道信号，N表示总共有N路LED灯；3) Convert the multi-channel signal

Perform whitening processing to obtain whitened multi-channel signals

Randomly assign initial values to the separation matrix M, and use whitening multi-channel signals

The separation matrix M is continuously iterated until the error between the previous separation matrix M _k-1 and the current separation matrix M _k satisfies ∑∑|M _k -M _k-1 |≤σ, and the final separation matrix M is obtained; where σ is Error, satisfying 0<σ<1, k is the number of iterations of the separation matrix M, ∑∑|M _k -M _k-1 | means that the separation matrix M _k and the separation matrix M _k-1 are different to obtain the intermediate matrix, The result is obtained after summing all elements in the matrix; the calculation obtains the original multi-channel signal

Where O(t) ₌ [O ₁ (t),O ₂ (t),...,ON (t)], O ₁ (t) represents the first original channel signal, that is, the first LED lamp The original channel signal generated, N means there are N LED lights in total;

Step 3) is specifically:

采用以下公式进行白化处理获得白化多通道信号

multi-channel signal

Use the following formula to perform whitening processing to obtain a whitened multi-channel signal

Randomly assign an initial value to the separation matrix M, and use the following formula to iterate the separation matrix M continuously until the error between the previous separation matrix M _k-1 and the current separation matrix M _k satisfies ∑∑|M _k -M _k-1 |≤σ , to obtain the final separation matrix M;

表示累积分布函数，

表示概率分布函数，m_k-1表示第k-1次迭代获得的分离矩阵M中某一行，y表示白化多通道信号

中与m_k-1相同的行序号的一行，m_k表示第k次迭代获得的分离矩阵M中与m_k-1相同的行序号的一行，M_k-1表示第k-1次迭代计算所得的分离矩阵M，M_k表示第k次迭代计算所得的分离矩阵M，E{}表示期望操作，T表示转置操作；in,

represents the cumulative distribution function,

represents the probability distribution function, m _k-1 represents a row in the separation matrix M obtained by the k-1 iteration, and y represents the whitened multi-channel signal

A row with the same row number as m _k-1 , m _k represents a row with the same row number as m k _- 1 in the separation matrix M obtained by the kth iteration, M _k-1 represents the k-1th iteration calculation The obtained separation matrix M, M _k represents the separation matrix M obtained by the k-th iteration calculation, E{} represents the desired operation, and T represents the transpose operation;

Finally, calculate the original multi-channel signal

4) Convert each sub-original channel signal in the original multi-channel signal O(t) into a square wave signal, calculate the duty cycle of each square wave signal, and form a total duty cycle q=[q ₁ ,q ₂ , ...,q _N ], from the total duty cycle q, the brightness of N LED lights L=[L ₁ , L ₂ , ..., L _N ], where the duty cycle corresponds to the brightness of the LED lights in equal proportions, When the duty cycle is 1, the LED light is the brightest working state; when the duty cycle is 0, the LED light is completely off.

In a specific implementation, the total duty ratio q=[q ₁ , q ₂ , q ₃ ], and the corresponding luminances of the three LED lamps L=[L ₁ , L ₂ , L ₃ ].

Claims (7)

1. The utility model provides a multichannel LED lamp luminance bypass detection device based on power ripple which characterized in that: the LED power supply comprises an LED lamp (1), an LED lamp driving circuit (2), a power supply ripple signal acquisition circuit (3) and a signal processor (4); LED lamp drive circuit (2) link to each other and control LED lamp (1) with LED lamp (1), a plurality of even has LED lamp drive circuit (2) of LED lamp (1) and forms multichannel LED lamp on hookup power (5), power ripple signal acquisition circuit (3) are received on power (5) and are gathered power ripple that a plurality of LED lamp drive circuit (2) produced power (5), power ripple signal acquisition circuit (3) transmit the power ripple of gathering to signal processor (4) through the USB switching mouth.

2. The power supply ripple-based multi-path LED lamp brightness bypass detection device according to claim 1, wherein: the LED lamp driving circuit (2) comprises a PWM signal duty ratio control circuit (7) and an LED lamp working circuit (10), the PWM signal duty ratio control circuit (7) is connected with the LED lamp working circuit (10), and the LED lamp working circuit (10) is connected with the LED lamp (1); the duty ratio of the high level and the low level in the PWM wave in the PWM signal duty ratio control circuit (7) is controlled by an external knob or infrared remote control, so that the working time of the LED lamp working circuit (10) is controlled, and the effect of adjusting the brightness of the LED lamp (1) is achieved.

3. The power supply ripple-based multi-path LED lamp brightness bypass detection device according to claim 1, wherein: the power supply ripple signal acquisition circuit (3) comprises a filter circuit (11), a load resistor (12), a sampling resistor (13), an AD sampling module (14) and an amplifying circuit (15); load resistance (12) and sampling resistor (13) that are connected with the series between power (5), the both ends and amplifier circuit (15) of sampling resistor (13) are connected, and amplifier circuit (15) link to each other with filter circuit (11), electric capacity, AD sampling module (14) in proper order, and AD sampling module (14) change the power ripple transmission to signal processor (4) that the USB chip will gather through the CH340 serial ports.

4. A power supply ripple-based multi-path LED lamp brightness bypass detection method applied to the multi-path LED lamp brightness bypass detection device of any one of claims 1 to 3 is characterized in that: the mixed ripple signal is input into the signal processor (4), and the method for processing the mixed ripple signal comprises the following steps:

1) j-layer decomposition is carried out on the mixed ripple signal, and a wavelet coefficient and a scale coefficient of the mixed ripple signal are obtained after decomposition and reconstruction;

2) selecting wavelet coefficient and scale coefficient, establishing signal decomposition matrix, and performing dimensionality reduction reconstruction on the signal decomposition matrix to obtain multi-channel signal

3) Combining multiple channel signals

Whitening processing is carried out to obtain whitened multichannel signal

To the separation matrix M randomAssigning initial values using whitened multichannel signals

Continuously iterating the separation matrix M to the previous separation matrix M_k-1From the current separation matrix M_kSatisfies sigma M_k-M_k-1Obtaining a final separation matrix M, wherein | < sigma; wherein σ is an error, and satisfies 0<σ<1, k is the number of iterations of the separation matrix M, Σ | M_k-M_k-1I denotes a separation matrix M_kAnd a separation matrix M_k-1Obtaining an intermediate matrix after difference is made, and summing all elements in the intermediate matrix to obtain a result; calculating to obtain original multi-channel signal

Wherein O (t) ═ O₁(t),O₂(t),···,O_N(t)]，O₁(t) represents the 1 st original channel signal, and N represents the total N LED lamps;

4) converting each sub-original channel signal in the original multi-channel signal O (t) into a square wave signal, calculating the duty ratio of each square wave signal, and forming a total duty ratio q ═ q₁,q₂,···,q_N]And obtaining the brightness L of the N LED lamps [ L ] from the total duty ratio q₁,L₂,···,L_N]The duty ratio corresponds to the brightness of the LED lamp in equal proportion, and when the duty ratio is 1, the LED lamp is in the brightest working state; when the duty ratio is 0, the LED lamp is completely extinguished.

5. The method of claim 4, wherein the method comprises the following steps: the step 1) is specifically as follows:

j-layer decomposition is carried out on the mixed ripple signal by using a complex wavelet function basis of Hilbert transform, a decomposed signal Y is obtained after decomposition and reconstruction, and a wavelet coefficient and a scale coefficient of the mixed ripple signal can be obtained from the decomposed signal Y:

wherein d is_iFor decomposing the i-th wavelet coefficient, c, in the signal Y_jTo decompose the jth scale factor in the signal Y.

6. The method of claim 5, wherein the method comprises the following steps: the step 2) is specifically as follows:

performing dimensionality reduction reconstruction on the signal decomposition matrix Q to obtain a multi-channel signal

Selecting wavelet coefficient d₁,d₂,···,d_jAnd a scale factor c_jWherein d is₁Represents the 1 st wavelet coefficient, d₂Representing the 2 nd wavelet coefficient, d_jExpressing the jth wavelet coefficient, and establishing a signal decomposition matrix Q ═ c_j,d₁,d₂,···,d_j]^TAnd calculating the eigenvalue Λ ═ λ of the covariance matrix S₁,λ₂,...,λ_j+1]And the corresponding feature vector V ═ ω₁,ω₂,…,ω_j+1]：

Wherein,

u₁represents the scale factor c_jMean value of u₂Representing wavelet coefficients d₁Mean value of u_j+1Representing wavelet coefficients d_jT denotes the transposition operation, λ₁Represents the 1 st eigenvalue, λ, in the covariance matrix S_j+1Representing covarianceThe j +1 th eigenvalue, ω, in the matrix S₁Represents the 1 st eigenvector, ω, in the covariance matrix S_j+1Representing the j +1 th eigenvector in the covariance matrix S;

let the eigenvalue Λ ═ λ₁,λ₂,...,λ_j+1]Sorting from big to small, sequentially selecting N-1 eigenvalues, wherein N represents the number of LED lamps, and selecting N-1 signal components of eigenvalue serial numbers corresponding to the N-1 eigenvalues from the signal decomposition matrix Q to form a multi-channel signal together with the mixed ripple signal

7. The method of claim 6, wherein the method comprises the following steps: the step 3) is specifically as follows:

combining multiple channel signals

Whitening processing is carried out by adopting the following formula to obtain a whitened multichannel signal

Randomly assigning an initial value to the separation matrix M, and continuously iterating the separation matrix M to the previous separation matrix M by using the following formula_k-1From the current separation matrix M_kSatisfies sigma M_k-M_k-1Obtaining a final separation matrix M, wherein | < sigma;

wherein,

the function of the cumulative distribution is represented,

representing a probability distribution function, m_k-1Represents a row in the separation matrix M obtained in the (k-1) th iteration, and y represents the whitened multi-channel signal

M and_k-1one line of the same line sequence number, m_kDenotes M and M in the separation matrix M obtained in the k-th iteration_k-1One line of the same line sequence number, M_k-1Representing the separation matrix M, M obtained by the k-1 th iteration calculation_kRepresenting a separation matrix M obtained by the k-th iterative computation, wherein E { } represents an expected operation, and T represents a transposition operation;

finally, the original multi-channel signal is calculated

Images