Automatically-controlled gradient LED lamp energy-saving illumination method

Abstract

The invention discloses an automatic control gradient LED lamp energy-saving illumination method, which comprises the following steps: S1, asynchronous sampling is carried out through an illumination sensor A and an illumination sensor B, ambient light change data is collected, a sampling timestamp is recorded, a time difference weight is obtained based on the timestamp, data fusion is carried out through the time difference weight, and ambient light intensity data is obtained; s2, performing data calibration on the ambient light intensity data through a state space model, and performing optimization through a weighted moving average method to obtain an ambient light intensity change rate; and S3, acquiring a light change condition based on an environment illumination intensity change rate, adjusting a brightness duty ratio through a probability model based on Bayesian reasoning to output an initial control variable, performing optimization based on the initial control variable in combination with a particle filtering algorithm to obtain a final control scheme, and performing optimization through the probability model and the algorithm to obtain a final control scheme. A relatively accurate brightness adjustment scheme can still be given, and the intelligent level of lamp control is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of gradient lamp control, and in particular to an energy-saving lighting method of an automatically controlled gradient LED lamp. Background Art

[0002] At present, with the increasing demand for indoor and outdoor lighting, good lighting conditions are essential in commercial places, office areas and home environments. However, lighting consumes a lot of energy. According to statistics, lighting electricity accounts for a considerable proportion of the world's total electricity consumption. While pursuing a comfortable lighting environment, how to reduce energy consumption has become the main problem to be solved.

[0003] Traditional automatically controlled gradient LED lamps usually adopt a single-rate brightness gradient method. The single-rate brightness gradient method cannot reasonably adjust the brightness according to the actual situation of the ambient light intensity. For example, when the ambient light intensity changes slowly (such as in the evening when the natural light gradually weakens), the fixed-rate brightness gradient method may cause the brightness adjustment to be too rapid. For example, the single rate is to adjust the brightness duty cycle by 5% every 10 seconds. In such a slowly changing environment, such an adjustment speed will make the lamp brightness change significantly faster than the natural change of the ambient light intensity, causing the human eye to perceive the light change as unnatural and even causing visual fatigue. Conversely, when the ambient light intensity drops rapidly (such as a sudden change in the weather), if the brightness adjustment speed is too slow, the brightness of the lamp will not have time to adapt to the ambient brightness, resulting in too dim a brightness and affecting the lighting effect. Therefore, an energy-saving lighting method for automatically controlled gradient LED lamps is proposed. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention proposes the following technical solution: an automatically controlled gradient LED lamp energy-saving lighting method, comprising the following steps:

[0005] S1: Asynchronously sample light sensors A and B to collect ambient light change data, record the sampling timestamps, obtain time difference weights based on the timestamps, and fuse the data using the time difference weights to obtain ambient light intensity data.

[0006] S2: Calibrate the ambient light intensity data through a state space model and obtain the optimized ambient light change rate through weighted moving average optimization;

[0007] S3: Obtain light changes based on the ambient light intensity change rate, adjust the brightness duty cycle output initial control variables through a probability model based on Bayesian inference, and optimize based on the initial control variables combined with the particle filter algorithm to obtain the final control solution;

[0008] S4: Convert the output final control scheme into control instructions to control the light brightness.

[0009] Two light sensors are used for asynchronous sampling, denoted as sensor A and sensor B.

[0010] Asynchronous sampling is performed by setting a sampling frequency. For example, the sampling frequency of sensor A is set to once every 2 seconds, and the sampling frequency of sensor B is set to once every 3 seconds.

[0011] When sensor A collects data, the recorded ambient light intensity data is At the same time, the corresponding timestamp is recorded as Similarly, the data collected by sensor B is recorded as The corresponding timestamp is

[0012] Set a target time frame;

[0013] Specifically, every 5 minutes is a target time period for data fusion. For each target time point t (evenly selected within the target time period), find the data point closest to this time point in the data collected by sensor A and sensor B, and set it as

[0014] To calculate the time difference weight between sensors A and 2, first obtain the time difference of sensor A: Time difference with sensor B: Where i is the index of sensor A, represents the time difference of the i-th sampling of sensor A, j is the index of sensor B, represents the time difference of the j-th sampling of sensor B;

[0015] Based on time difference and Get time difference weight and

[0016] Calculate the time difference weight of sensor A The calculation formula is:

[0017] Calculate the time difference weight of sensor B The calculation formula is:

[0018] Specifically, unlike the traditional fixed weight allocation method, weights are assigned based on the proximity of each data point to the target time point. This means that for different time points, the weights will be adaptively adjusted to better match the characteristics of the actual data.

[0019] The data collected by sensor A and sensor B are fused based on the time difference weight to obtain the ambient light intensity data, which is expressed as:

[0020]

[0021] Specifically, this time difference-based fusion method fully considers the time factor of data collection and ensures that data at similar time points are fused, which can reduce data deviation caused by time differences. Because in real environments, the ambient light intensity will change rapidly over time;

[0022] For example, when the ambient light intensity changes rapidly, data closer to that moment will be given a higher weight, so that the fusion result can more accurately reflect the current ambient light intensity.

[0023] The process of obtaining the environmental change rate is:

[0024] First, the ambient light intensity data is calibrated through a state space model;

[0025] The state space model consists of a state equation and a calibration equation;

[0026] The true value of the ambient light intensity is defined as state x, and the measured value is z;

[0027] Construct a state equation based on the state x: k =Ax k-1 +Bμ k-1 +w k-1 ;

[0028] Among them, x k represents the true value of the ambient light intensity at time k, x k-1 is the actual value of the ambient light intensity at the previous moment k-1, A and B are state equation matrices, representing the evolution of the state in the absence of external interference. For example, A represents the coefficient matrix of the ambient light intensity during natural attenuation or enhancement, and B is related to external controllable factors (usually set to 0 in ambient light intensity measurement);

[0029] μ k-1 Represents the control vector at the previous moment, which refers to the external control factor applied in the ambient light intensity measurement scene, and is set to 0 here, w k-1 represents the process noise at the previous moment, which represents the unpredictable random interference factors in the process of system state evolution;

[0030] The calibration equation is constructed based on the measurement value z defined by the ambient light intensity data. The calibration equation is expressed as: k =Hx k +v k , where zk Indicates the calibration value of the ambient light intensity at time k, Hx k is the measurement matrix of the calibration value of the ambient light intensity at time k, v k is the measurement noise at time k;

[0031] Based on the calibration value output by the state space model, the light change rate is optimized by the weighted moving average method, and the calibration value z of the ambient light intensity is converted to k As input, a smoothing coefficient that reflects the degree of data fluctuation is introduced to obtain the weighted average value. The implementation formula is: S k =az k +(1-a)S k-1 , where S k is the weighted average value at time k, a is the smoothing coefficient reflecting the degree of data fluctuation, S k-1 is the weighted average of the previous moment;

[0032] The process of obtaining the smoothing coefficient that reflects the degree of data fluctuation is as follows:

[0033] Calculate the standard deviation of the ambient light intensity data to measure the degree of data fluctuation. If the data fluctuates greatly, it means that the ambient light intensity changes drastically. In this case, you can increase the smoothing coefficient appropriately so that the weighted average value can keep up with the changes in ambient light intensity more quickly.

[0034] For example, a threshold σ is set. When the standard deviation of the data is greater than the threshold σ, the value of the smoothing coefficient a is increased. On the contrary, if the data fluctuation is small, it means that the ambient light intensity is relatively stable. In this case, the smoothing coefficient a is reduced.

[0035] After obtaining the weighted average value at time k, calculate the ambient light change rate. The calculation formula is: ΔI * =S k -S k-1 .

[0036] Light change conditions include slow changes in ambient light intensity and rapid changes in ambient light intensity;

[0037] Specifically, the slow change of ambient light intensity is |ΔI * |≤50lux / h, the ambient light intensity changes rapidly as |ΔI * |≥150lux / h;

[0038] The process of obtaining the initial control variables is:

[0039] Based on the changes in ambient light, a probability model based on Bayesian inference is constructed to adjust the brightness duty cycle output initial control variable;

[0040] Define the ambient light change rate |ΔI * |≤50lux / h or |ΔI * |Ambient light intensity data of ≥150 lux / h is used as the input of the probability model based on Bayesian inference;

[0041] Among them, the rate of change of ambient light intensity follows the normal distribution ΔI * ~N(u,v 2 ), u is the mean value of the ambient light intensity data, v 2 is the variance of the ambient light intensity data;

[0042] The prior probability distribution of the brightness duty cycle adjustment amount is defined as a uniform distribution U(f, g), where f and g are ranges set based on experience;

[0043] Specifically, f and g take into account the brightness adjustment capability of the lamp itself. Based on the minimum brightness and maximum brightness, the corresponding brightness duty cycle is determined. The specific situation is:

[0044] Assume that in the case of the slowest change in ambient light intensity (i.e. close to 0), the brightness duty cycle does not need to be adjusted. At this time, in the most extreme case of slow change, the adjustment amount of the brightness duty cycle will not exceed the adjustment range of the lamp itself. For example, the brightness duty cycle adjustment range of the lamp is from 0.1 to 0.9, then the sum can be set within the value range corresponding to this range;

[0045] For example: Assume that the initial brightness duty cycle D0 = 0.5;

[0046] When D takes the minimum value, the brightness duty cycle D=0.1, so ΔD min =0.1-0.5=-0.4;

[0047] When D takes the maximum value, the brightness duty cycle D=0.9, so ΔD max =0.9-0.5=0.4;

[0048] In this case, f = -0.4 and g = 0.4;

[0049] Based on Bayes’ theorem, define a likelihood function P(ΔI * |ΔD), where ΔD is the brightness duty cycle adjustment;

[0050] Specifically, the likelihood function P(ΔI * |ΔD) represents the observed ambient light intensity change rate ΔI under a given brightness duty cycle adjustment ΔD * probability;

[0051] By Bayes' formula: Among them, P(ΔD) is the prior probability, that is, the probability distribution of the brightness duty cycle adjustment amount, P(ΔI * ) is the evidence factor, which serves as a normalization constant to ensure that the sum of the posterior probabilities is 1, P(ΔI * |ΔD) is the posterior probability distribution;

[0052] The brightness duty cycle adjustment is determined based on the expectation of the posterior probability distribution, expressed as ΔD = E{ΔD|ΔI *};

[0053] This brightness duty cycle adjustment ΔD is the initial control variable;

[0054] The process of obtaining the final control solution is:

[0055] The brightness duty cycle adjustment value ΔD is expressed by a set of weighted particles through the particle filter algorithm to represent the brightness duty cycle adjustment value of the brightness gradient model based on Bayesian inference. N particles are randomly generated. And initialize the weights in represents the brightness duty cycle adjustment of the i-th particle at the initial moment;

[0056] Resample according to the particle weights, remove particles with weights lower than the weighted average standard, copy particles with weights higher than the weighted average standard, and obtain a new particle set. Calculate the estimated value of the brightness duty cycle adjustment through all particles in the particle set and their weights. The formula is expressed as: Among them, m represents the time step, represents the weight of the i-th particle at time step m, represents the brightness duty cycle adjustment of the i-th particle at time step m;

[0057] The brightness duty cycle adjustment value That is, the final control plan;

[0058] Specifically, this formula represents the process of calculating the estimated value of the brightness duty cycle adjustment through all particles and their weights. The brightness duty cycle adjustment obtained based on the Bayesian method can be optimized through the particle filtering algorithm. Because the ambient light intensity changes dynamically, Bayesian reasoning combined with particle filtering can adapt well to this dynamic change. Particle filtering can track the changes in ambient light intensity in real time by continuously updating the weights and positions of particles, and then adjust the brightness duty cycle to obtain a more accurate and stable duty cycle adjustment, thereby better controlling the brightness of the lamp.

[0059] Adjust the brightness duty cycle Convert it into lighting control instructions and control the light brightness based on the lighting communication protocol and control interface;

[0060] Specifically, based on the PWM (pulse width modulation) control method, the brightness duty cycle adjustment amount is converted into a corresponding duty cycle parameter of the PWM signal;

[0061] For example: the duty cycle range of the PWM signal received by the lighting control interface is 0% to 100%, and the calculated brightness duty cycle adjustment range is -0.4, which is 0.4 to (according to the example mentioned above, the actual brightness duty cycle adjustment range is -0.4). Within this range), it is converted into a PWM duty cycle through a linear mapping. The converted PWM duty cycle is F. The converted control instruction (PWM duty cycle F) is sent to the lamp through communication to control the brightness.

[0062] The present invention has the following beneficial effects:

[0063] In this invention, the asynchronous sampling of two light sensors and the data fusion method combined with time difference weighting can more accurately obtain ambient light intensity data. The ambient light change rate is optimized based on state space model calibration and weighted moving average method. The brightness duty cycle is dynamically adjusted according to the speed of light change. This avoids the traditional single-rate brightness gradient method that adjusts too quickly when the ambient light intensity changes slowly, resulting in energy waste.

[0064] Secondly, intelligent decision-making is achieved based on probability models and advanced algorithms. A probability model based on Bayesian reasoning is combined with a particle filter algorithm to determine the brightness duty cycle adjustment amount. This method fully utilizes probability statistics and advanced calculation methods. This method can comprehensively consider the uncertainty of changes in ambient light intensity and make more intelligent brightness control decisions.

[0065] Finally, when there is a certain error or uncertainty in the measurement of the ambient light intensity change rate, an accurate brightness adjustment solution can still be given through the optimization of the probability model and algorithm, thereby improving the intelligence level of lighting control. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 This is a method step diagram of an automatically controlled gradient LED lamp energy-saving lighting method proposed by the present invention. DETAILED DESCRIPTION

[0067] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0068] Example 1

[0069] like Figure 1 As shown, the present invention proposes an automatic control gradient LED lamp energy-saving lighting method, comprising the following steps:

[0070] S1: Asynchronously sample light sensors A and B to collect ambient light change data, record the sampling timestamps, obtain time difference weights based on the timestamps, and fuse the data using the time difference weights to obtain ambient light intensity data.

[0071] Two light sensors are used for asynchronous sampling, denoted as sensor A and sensor B.

[0072] Asynchronous sampling is performed by setting a sampling frequency. For example, the sampling frequency of sensor A is set to once every 2 seconds, and the sampling frequency of sensor B is set to once every 3 seconds.

[0073] When sensor A collects data, the recorded ambient light intensity data is At the same time, the corresponding timestamp is recorded as Similarly, the data collected by sensor B is recorded as The corresponding timestamp is

[0074] Set a target time frame;

[0075] Specifically, every 5 minutes is a target time period for data fusion. For each target time point t (evenly selected within the target time period), find the data point closest to this time point in the data collected by sensor A and sensor B, and set it as

[0076] To calculate the time difference weight between sensor A and sensor B, first obtain the time difference of sensor A: Time difference with sensor B: Where i is the index of sensor A, represents the time difference of the i-th sampling of sensor A, j is the index of sensor B, represents the time difference of the j-th sampling of sensor B, Here 1 refers to sensor A. Here, 2 refers to sensor B;

[0077] Based on time difference and Get time difference weight and

[0078] Calculate the time difference weight of sensor A The calculation formula is:

[0079] Calculate the time difference weight of sensor B The calculation formula is:

[0080] Specifically, unlike the traditional fixed weight allocation method, weights are assigned based on the proximity of each data point to the target time point. This means that for different time points, the weights will be adaptively adjusted to better match the characteristics of the actual data.

[0081] The data collected by sensor A and sensor B are fused based on the time difference weight to obtain the ambient light intensity data, which is expressed as:

[0082]

[0083] Specifically, this time difference-based fusion method fully considers the time factor of data collection and ensures that data at similar time points are fused, which can reduce data deviation caused by time differences. Because in real environments, the ambient light intensity will change rapidly over time;

[0084] For example, when the ambient light intensity changes rapidly, data closer to that moment will be given a higher weight, so that the fusion result can more accurately reflect the current ambient light intensity.

[0085] S2: Calibrate the ambient light intensity data through a state space model and obtain the optimized ambient light change rate through weighted moving average optimization;

[0086] The process of obtaining the environmental change rate is:

[0087] First, the ambient light intensity data is calibrated through a state space model;

[0088] The state space model consists of a state equation and a calibration equation;

[0089] The true value of the ambient light intensity is defined as state x, and the measured value is z;

[0090] Construct a state equation based on the state x: k =Ax k-1 +Bμ k-1 +w k-1 ;

[0091] Among them, x k represents the true value of the ambient light intensity at time k, x k-1is the actual value of the ambient light intensity at the previous moment k-1, A and B are state equation matrices, representing the evolution of the state in the absence of external interference. For example, A represents the coefficient matrix of the ambient light intensity during natural attenuation or enhancement, and B is related to external controllable factors (usually set to 0 in ambient light intensity measurement);

[0092] μ k-1 Represents the control vector at the previous moment, which refers to the external control factor applied in the ambient light intensity measurement scene, and is set to 0 here, w k-1 represents the process noise at the previous moment, which represents the unpredictable random interference factors in the process of system state evolution;

[0093] The calibration equation is constructed based on the measurement value z defined by the ambient light intensity data. The calibration equation is expressed as: k =Hx k +v k , where z k Indicates the calibration value of the ambient light intensity at time k, Hx k is the measurement matrix of the calibration value of the ambient light intensity at time k, v k is the measurement noise at time k;

[0094] Based on the calibration value output by the state space model, the light change rate is optimized by the weighted moving average method, and the calibration value z of the ambient light intensity is converted to k As input, a smoothing coefficient that reflects the degree of data fluctuation is introduced to obtain the weighted average value. The implementation formula is: S k =az k +(1-a)S k-1 , where S k is the weighted average value at time k, a is the smoothing coefficient reflecting the degree of data fluctuation, S k-1 is the weighted average of the previous moment;

[0095] The process of obtaining the smoothing coefficient that reflects the degree of data fluctuation is as follows:

[0096] Calculate the standard deviation of the ambient light intensity data to measure the degree of data fluctuation. If the data fluctuates greatly, it means that the ambient light intensity changes drastically. In this case, you can increase the smoothing coefficient appropriately so that the weighted average value can keep up with the changes in ambient light intensity more quickly.

[0097] For example, a threshold σ is set. When the standard deviation of the data is greater than the threshold σ, the value of the smoothing coefficient a is increased. On the contrary, if the data fluctuation is small, it means that the ambient light intensity is relatively stable. In this case, the smoothing coefficient a is reduced.

[0098] After obtaining the weighted average value at time k, calculate the ambient light change rate. The calculation formula is: ΔI *=S k -S k-1 .

[0099] S3: Obtain light changes based on the ambient light intensity change rate, adjust the brightness duty cycle output initial control variables through a probability model based on Bayesian inference, and optimize based on the initial control variables combined with the particle filter algorithm to obtain the final control solution;

[0100] Light change conditions include slow changes in ambient light intensity and rapid changes in ambient light intensity;

[0101] Specifically, the slow change of ambient light intensity is |ΔI * |≤50lux / h, the ambient light intensity changes rapidly as |ΔI * |≥150lux / h;

[0102] The process of obtaining the initial control variables is:

[0103] Based on the changes in ambient light, a probability model based on Bayesian inference is constructed to adjust the brightness duty cycle output initial control variable;

[0104] Define the ambient light change rate |ΔI * |≤50lux / h or |ΔI * |Ambient light intensity data of ≥150 lux / h is used as the input of the probability model based on Bayesian inference;

[0105] Among them, the rate of change of ambient light intensity follows the normal distribution ΔI * ~N(u,v 2 ), u is the mean value of the ambient light intensity data, v 2 is the variance of the ambient light intensity data;

[0106] The prior probability distribution of the brightness duty cycle adjustment amount is defined as a uniform distribution U(f, g), where f and g are ranges set based on experience;

[0107] Specifically, f and g take into account the brightness adjustment capability of the lamp itself. Based on the minimum brightness and maximum brightness, the corresponding brightness duty cycle is determined. The specific situation is:

[0108] Assume that in the case of the slowest change in ambient light intensity (i.e. close to 0), the brightness duty cycle does not need to be adjusted. At this time, in the most extreme case of slow change, the adjustment amount of the brightness duty cycle will not exceed the adjustment range of the lamp itself. For example, the brightness duty cycle adjustment range of the lamp is from 0.1 to 0.9, then the sum can be set within the value range corresponding to this range;

[0109] For example: Assume that the initial brightness duty cycle D0 = 0.5;

[0110] When D takes the minimum value, the brightness duty cycle D=0.1, so ΔD min =0.1-0.5=-0.4;

[0111] When D takes the maximum value, the brightness duty cycle D=0.9, so ΔD max =0.9-0.5=0.4;

[0112] In this case, f = -0.4 and g = 0.4;

[0113] Based on Bayes’ theorem, define a likelihood function P(ΔI * |ΔD), where ΔD is the brightness duty cycle adjustment;

[0114] Specifically, the likelihood function P(ΔI * |ΔD) represents the observed ambient light intensity change rate ΔI under a given brightness duty cycle adjustment ΔD * probability;

[0115] By Bayes' formula: Among them, P(ΔD) is the prior probability, that is, the probability distribution of the brightness duty cycle adjustment amount, P(ΔI * ) is the evidence factor, which serves as a normalization constant to ensure that the sum of the posterior probabilities is 1, P(ΔI * |ΔD) is the posterior probability distribution;

[0116] The brightness duty cycle adjustment is determined based on the expectation of the posterior probability distribution, expressed as ΔD = E{ΔD|ΔI *};

[0117] This brightness duty cycle adjustment ΔD is the initial control variable;

[0118] The process of obtaining the final control solution is:

[0119] The brightness duty cycle adjustment value ΔD is expressed by a set of weighted particles through the particle filter algorithm to represent the brightness duty cycle adjustment value of the brightness gradient model based on Bayesian inference. N particles are randomly generated. And initialize the weights in represents the brightness duty cycle adjustment of the i-th particle at the initial moment;

[0120] Resample according to the particle weights, remove particles with weights lower than the weighted average standard, copy particles with weights higher than the weighted average standard, and obtain a new particle set. Calculate the estimated value of the brightness duty cycle adjustment through all particles in the particle set and their weights. The formula is expressed as: Among them, m represents the time step, represents the weight of the i-th particle at time step m, represents the brightness duty cycle adjustment of the i-th particle at time step m;

[0121] The brightness duty cycle adjustment value That is, the final control plan;

[0122] Specifically, this formula represents the process of calculating the estimated value of the brightness duty cycle adjustment through all particles and their weights. The brightness duty cycle adjustment obtained based on the Bayesian method can be optimized through the particle filtering algorithm. Because the ambient light intensity changes dynamically, Bayesian reasoning combined with particle filtering can adapt well to this dynamic change. Particle filtering can track the changes in ambient light intensity in real time by continuously updating the weights and positions of particles, and then adjust the brightness duty cycle to obtain a more accurate and stable duty cycle adjustment, thereby better controlling the brightness of the lamp.

[0123] S4: Convert the output final control scheme into control instructions to control the light brightness.

[0124] Adjust the brightness duty cycle Convert it into lighting control instructions and control the light brightness based on the lighting communication protocol and control interface;

[0125] Specifically, based on the PWM (pulse width modulation) control method, the brightness duty cycle adjustment amount is converted into a corresponding duty cycle parameter of the PWM signal;

[0126] For example: the duty cycle range of the PWM signal received by the lighting control interface is 0% to 100%, and the calculated brightness duty cycle adjustment range is -0.4, which is 0.4 to (according to the example mentioned above, the actual brightness duty cycle adjustment range is -0.4). Within this range), it is converted into a PWM duty cycle through a linear mapping. The converted PWM duty cycle is F. The converted control instruction (PWM duty cycle F) is sent to the lamp through communication to control the brightness.

[0127] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An automatic controlled gradient LED lamp energy-saving lighting method, characterized in that: The steps include: S1: Asynchronously sample sensors A and B to collect ambient light change data, record the sampling timestamps, obtain time difference weights based on the timestamps, and perform data fusion using the time difference weights to obtain ambient light intensity data. S2: The ambient light intensity data is calibrated using a state-space model, and the ambient light intensity change rate is optimized using the weighted moving average method. S3: Obtain light changes based on the ambient light intensity change rate, adjust the brightness duty cycle output initial control variables through a probability model based on Bayesian inference, and optimize based on the initial control variables combined with the particle filter algorithm to obtain the final control solution; S4: Convert the output final control scheme into control instructions to control the light brightness.

2. The automatic control gradient LED lamp energy-saving lighting method according to claim 1 is characterized in that: The data collected by sensor A is The corresponding timestamp is The data collected by sensor B is The corresponding timestamp is 3. The automatic control gradient LED lamp energy-saving lighting method according to claim 2, characterized in that: The process of obtaining the ambient light intensity data is as follows: To calculate the time difference weight between sensor A and sensor B, first obtain the time difference of sensor A: Time difference with sensor B: Where i is the index of sensor A, represents the time difference of the i-th sampling of sensor A, j is the index of sensor B, represents the time difference of the j-th sampling of sensor B; Based on time difference and Get time difference weight and Calculate the time difference weight of sensor A The calculation formula is: Calculate the time difference weight of sensor B The calculation formula is: The data collected by sensor A and sensor B are fused based on the time difference weight to obtain the ambient light intensity data, which is expressed as:

4. The automatic control gradient LED lamp energy-saving lighting method according to claim 1, characterized in that: The process of obtaining the ambient light change rate is as follows: First, the ambient light intensity data is calibrated through a state space model; The state space model includes a state equation and a calibration equation; The true value of the ambient light intensity is defined as state x, and the measured value is z; Construct a state equation based on the state x: k =Ax k-1 +Bμ k-1 +w k-1 ; Among them, x k represents the true value of the ambient light intensity at time k, x k-1 is the true value of the ambient light intensity at the previous moment k-1, A and B are the state equation matrices, representing the evolution law of the state when there is no external interference, μ k-1 represents the control vector at the previous moment, w k-1 represents the process noise at the previous moment; The calibration equation is constructed based on the measurement value z defined by the ambient light intensity data. The calibration equation is expressed as: k =Hx k +v k , where z k Indicates the calibration value of the ambient light intensity at time k, Hx k is the measurement matrix of the calibration value of the ambient light intensity at time k, v k is the measurement noise at time k; Based on the calibration value output by the state space model, the light change rate is optimized by the weighted moving average method, and the calibration value z of the ambient light intensity is converted to k As input, a smoothing coefficient that reflects the degree of data fluctuation is introduced to obtain the weighted average value. The implementation formula is: S k =az k +(1-a)S k-1 , where S k is the weighted average value at time k, a is the smoothing coefficient reflecting the degree of data fluctuation, S k-1 is the weighted average of the previous moment; After obtaining the weighted average value at time k, calculate the ambient light change rate. The calculation formula is: ΔI * =S k -S k-1 .

5. The automatic control gradient LED lamp energy-saving lighting method according to claim 1, characterized in that: The light change conditions include a slow change in ambient light intensity and a rapid change in ambient light intensity; The slow change of ambient light intensity is |ΔI * |≤50lux / h, the ambient light intensity changes rapidly as |ΔI * |≥150lux / h.

6. The automatic control gradient LED lamp energy-saving lighting method according to claim 5, characterized in that: The process of obtaining the initial control variables is as follows: Based on the changes in ambient light, a probability model based on Bayesian inference is constructed to adjust the brightness duty cycle output initial control variable; Define the ambient light change rate |ΔI * |≤50lux / h or |ΔI * |Ambient light intensity data of ≥150 lux / h is used as the input of the probability model based on Bayesian inference; Among them, the rate of change of ambient light intensity follows the normal distribution ΔI * ~N(u,v 2 ), u is the mean value of the ambient light intensity data, v 2 is the variance of the ambient light intensity data; The prior probability distribution of the brightness duty cycle adjustment amount is defined as a uniform distribution U(f, g), where f and g are set ranges; Based on Bayes’ theorem, define a likelihood function P(ΔI * |ΔD), where ΔD is the brightness duty cycle adjustment; By Bayes' formula: Among them, P(ΔD) is the prior probability, that is, the probability distribution of the brightness duty cycle adjustment amount, P(ΔI * ) is the evidence factor, which serves as a normalization constant to ensure that the sum of the posterior probabilities is 1, P(ΔI * |ΔD) is the posterior probability distribution; The brightness duty cycle adjustment is determined based on the expectation of the posterior probability distribution, expressed as ΔD = E{ΔD|ΔI * }; The brightness duty cycle adjustment amount ΔD is the initial control variable.

7. The automatic control gradient LED lamp energy-saving lighting method according to claim 6, characterized in that: The process of obtaining the final control solution is as follows: The brightness duty cycle adjustment value ΔD is calculated by the particle filter algorithm to form a set of weighted particles to represent the brightness duty cycle adjustment value of the brightness gradient model based on Bayesian inference. N particles are randomly generated. And initialize the weights in represents the brightness duty cycle adjustment of the i-th particle at the initial moment; Resample according to the particle weights, remove particles with weights lower than the weighted average standard, copy particles with weights higher than the weighted average standard, and obtain a new particle set. Calculate the estimated value of the brightness duty cycle adjustment through all particles in the particle set and their weights. The formula is expressed as: Among them, m represents the time step, represents the weight of the i-th particle at time step m, represents the brightness duty cycle adjustment of the i-th particle at time step m; The brightness duty cycle adjustment value That is the final control solution.

8. The automatic controlled gradient LED lamp energy-saving lighting method according to claim 1, characterized in that: The light brightness is controlled by adjusting the brightness duty cycle Convert it into lighting control instructions and control the light brightness based on the lighting communication protocol and control interface.

Images