Landscape lighting intelligent adjusting system based on tone and motion state fusion analysis

Abstract

The invention discloses a landscape lighting intelligent adjusting system based on tone and motion state fusion analysis, and relates to the technical field of digital information processing. The method is used for solving the problems of multi-modal data logic conflict and control signal step. The method comprises the following steps: firstly, executing frequency domain transformation and differential operation on acquired space-time sequence data to generate normalized tone and motion intensity; then, calculating a cooperative gain coefficient and a dispersion index, and quantifying a scene active level and a data deviation; then, calculating a credibility weight based on the dispersion, and executing fusion iterative operation on a preset parameter and a historical state to generate a target control parameter; and finally, generating a pulse width modulation digital sequence in smooth transition by using a linear interpolation algorithm, and outputting a control instruction. According to the system, high-reliability fusion of multi-source data and adaptive smoothing of lighting effect control are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of digital information processing, in particular to a landscape lighting intelligent adjustment system based on tone and motion state fusion analysis. BACKGROUND

[0002] With the diversification of urban public space functions, landscape lighting has evolved from single night functional lighting to creating space atmosphere and enhancing interactive experience. In open areas such as squares, parks and commercial pedestrian streets, lighting systems become an important medium connecting tourists and the environment, which can adjust the light environment in real time according to the activity state of the crowd, and has important significance for improving the vitality of public space and improving the visual experience of tourists.

[0003] However, the existing landscape lighting adjustment method has substantial technical defects in practical application. On the one hand, the existing control logic mostly relies on single modal data triggering, such as only determining the light brightness according to the amplitude of the sound signal, or only switching the lighting mode according to the triggering frequency of the infrared sensor. This processing method cannot analyze the scene in depth, is easily disturbed by environmental background noise, and causes the light environment atmosphere to be unmatched with the actual crowd psychological state when detecting high-decibel non-human noise to trigger the active mode, or when the crowd is stationary and talking to misjudge as an empty scene. On the other hand, the existing system mostly adopts threshold jump mechanism in control signal generation, lacks consistency check and smoothing processing of multi-dimensional data, and causes the lighting device to produce abrupt brightness jump when switching modes, which not only affects the visual comfort, but also reduces the control stability of the system. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a landscape lighting intelligent adjustment system based on tone and motion state fusion analysis, which solves the problems of the above background technology.

[0005] To achieve the above object, the application is implemented by the following technical solutions: a landscape lighting intelligent adjustment system based on tone and motion state fusion analysis, comprising the following modules: a feature extraction and quantization module, configured to perform short-time Fourier transform on the collected audio frame sequence, extract the fundamental frequency mean and frequency fluctuation variance in the frequency domain as tone feature components, and perform normalization processing to generate tone intensity values, and simultaneously perform differential operation on the collected motion target coordinate sequence, extract the displacement speed mean and direction change rate within a unit time as motion feature components, and perform normalization processing to generate motion intensity values; a collaborative interaction calculation module, configured to linearly weight and sum the tone intensity values and the motion intensity values, obtain an interaction collaborative gain coefficient representing sound-motion synchronization, determine the basic activity level of the current scene, and simultaneously calculate the ratio deviation between the tone intensity values and the motion intensity values to generate a data dispersion index; a confidence light efficiency correction module, configured to index the preset initial lighting parameters with the basic activity level, substitute the data dispersion index into an inverse proportion weighting function to calculate the credibility weight of the current data, and perform weighted fusion iterative operation on the initial lighting parameters and the historical lighting parameters to generate target color temperature values and target brightness values; a time sequence interpolation driving module, configured to convert the target color temperature values and the target brightness values into target pulse width modulation values, construct a to-be-executed queue containing the current pulse width modulation values and the target pulse width modulation values, generate a transition step sequence between the current and target values through a linear interpolation algorithm, and output pulse width modulation control instructions in a preset clock period.

[0006] Further, the specific process of performing short-time Fourier transform on the collected audio frame sequence, extracting the fundamental frequency mean and frequency fluctuation variance in the frequency domain as tone feature components, and performing normalization processing to generate tone intensity values is as follows: the audio frame sequence is framed and windowed by a Hanning window function, and fast Fourier transform is performed to obtain the frequency spectrum amplitude spectrum of each frame, the fundamental frequency peak point is located by a frequency spectrum peak value search algorithm, the numerical average value of the fundamental frequency peak points of consecutive multiple frames is analyzed as the fundamental frequency mean, and the second-order central moment of the fundamental frequency peak point on the time axis is quantified as the frequency fluctuation variance; the fundamental frequency mean and the frequency fluctuation variance are projected to a preset emotion feature mapping interval, the projected feature data is compressed to the interval of zero to one by a Sigmoid activation function, and the compressed fundamental frequency data and variance data are combined by weighted aggregation operation, and the tone intensity value representing the sound emotion activity degree is output.

[0007] Further, the specific process of performing differential operation on the collected motion target coordinate sequence, extracting the displacement speed average and direction change rate in unit time as motion feature components, and performing normalization processing to generate motion intensity value is as follows: performing first-order differential operation on adjacent coordinate points in the motion target coordinate sequence, analyzing the Euclidean distance to obtain the instantaneous speed vector module length, and calculating the arithmetic average of all instantaneous speed vector module lengths in the preset time window as the displacement speed average; at the same time, the geometric angle relationship between adjacent instantaneous speed vectors is extracted, and the tangent angle change amount of the motion trajectory is obtained by inverse cosine function analysis, and the cumulative value of the angle change amount in unit time is taken as the direction change rate; the displacement speed average and the direction change rate are mapped to the standard dimensionless interval by maximum minimum standardization, the standardized speed average is taken as the active component, and the standardized direction change rate is taken as the entropy value component, the active component and the entropy value component are weighted and fused, and the motion intensity value representing the intensity of group behavior is output.

[0008] Further, the specific process of performing linear weighting summation on the tone intensity value and the motion intensity value to obtain the interactive synergy gain coefficient representing the sound-motion synchronization is as follows: calling the preset environmental bias weight coefficient, giving the tone intensity value and the motion intensity value priority weight respectively, analyzing the weighted sum value of the two as the initial synergy module, coupling modulation of the tone intensity value and the motion intensity value to generate the coupling gain term, and superimposing the initial synergy module and the coupling gain term to generate the interactive synergy gain coefficient; constructing a multi-level step threshold decision logic, comparing the interactive synergy gain coefficient with the preset quiet threshold, interactive threshold and carnival threshold in turn, and when the coefficient falls into different threshold intervals, output the corresponding discrete state identifier as the basic active level of the current scene.

[0009] Further, the specific process of simultaneously calculating the ratio deviation between the tone intensity value and the motion intensity value to generate the data dispersion index is as follows: analyzing the difference module between the tone intensity value and the motion intensity value as the modal absolute error, calculating the arithmetic sum of the tone intensity value and the motion intensity value as the modal total intensity; performing superposition operation on the modal total intensity and the regularization constant to generate the denominator term, calculating the relative deviation ratio between the modal absolute error and the denominator term by relative ratio quantization to generate the relative deviation ratio; inputting the relative deviation ratio into the preset index amplification function, nonlinear gain amplifying the input deviation signal, and outputting the amplified value as the data dispersion index.

[0010] Further, the specific process of retrieving the preset initial lighting parameters indexed by the basic activity level and substituting the data dispersion index into the inverse weighted function to calculate the credibility weight of the current data is as follows: an association database containing the activity level field and the lighting parameter field is established, the corresponding color temperature value and the brightness value are extracted as the initial lighting parameters by matching and querying in the database through the basic activity level as the key value; the inverse weighted calculation logic is constructed, the preset sensitivity coefficient and the reference constant are set, the product modulation of the sensitivity coefficient and the data dispersion index is performed to generate the decay term, the superposition operation of the reference constant and the decay term is performed to generate the denominator, the credibility weight of the current data is output by mapping the reference constant and the reciprocal of the denominator.

[0011] Further, the specific process of performing weighted fusion iterative operation on the initial lighting parameters and the historical lighting parameters to generate the target color temperature value and the target brightness value is as follows: the output value of the last control period is read from the system register as the historical lighting parameter, the inertia retention weight is obtained by complementary calculation of the unit value and the credibility weight; the initial lighting parameters and the credibility weight are weighted and modulated to generate a new state component, the historical lighting parameters and the inertia retention weight are weighted and modulated to generate a historical state component; the vector aggregation is performed on the new state component and the historical state component, and the fused color temperature data and brightness data are respectively output as the target color temperature value and the target brightness value.

[0012] Further, the specific process of converting the target color temperature value and the target brightness value into the target pulse width modulation value and constructing the to-be-executed queue containing the current pulse width modulation value and the target pulse width modulation value is as follows: the preset color temperature light mixing proportion algorithm is called, the power distribution ratio of the warm color temperature channel and the cold color temperature channel is calculated according to the target color temperature value, and the target duty cycle of each channel is mapped respectively according to the target brightness value to quantize the target pulse width modulation value; the pulse width modulation value being output at the current time is read from the register as the current node, and the target pulse width modulation value is taken as the terminal node, and a bidirectional linked list containing the start time stamp, the current node and the terminal node is constructed as the to-be-executed queue.

[0013] Further, the specific process of generating the transition step sequence between the current and target values by the linear interpolation algorithm and sequentially outputting the pulse width modulation control instruction according to the preset clock period is as follows: the value difference between the terminal node and the current node is calculated, the preset transition total time is obtained, the single-step change slope is obtained by ratio analysis of the value difference and the number of clock periods contained in the transition total time; the accumulator is initialized, and when the rising edge of each preset clock period comes, the output value of the last period and the single-step change slope are executed to generate the intermediate step value at the current time; the intermediate step value is converted into a binary control signal, and the binary control signal is written into the driving register through the general input and output interface to output the pulse width modulation control instruction.

[0014] The present application has the following advantages:

[0015] (1) The landscape lighting intelligent adjustment system based on tone and motion state fusion analysis realizes accurate quantitative identification of complex scene atmosphere through the cooperation of the feature extraction quantization module and the collaborative interaction calculation module. The system first performs frequency domain transformation and differential operation on audio and motion data, extracts tone intensity reflecting emotional color and motion intensity reflecting behavior intensity, and uses linear weighted summation and ratio deviation calculation to establish an interactive collaborative gain coefficient representing tone-motion synchronization from a mathematical dimension. The advantage of this data processing mechanism is that it no longer simply relies on the absolute intensity of the signal, but judges the authenticity of the scene by calculating the dispersion index between tone and motion. When the tone is high but the motion is static, or the motion is intense but the tone is low, the system can identify the high dispersion of the data and determine it as an atypical active scene, effectively shielding single-dimensional environmental interference and solving the misjudgment problem caused by the lack of multi-modal verification in the prior art, greatly improving the accuracy of scene recognition.

[0016] (2) The landscape lighting intelligent adjustment system based on tone and motion state fusion analysis realizes the adaptive stability and visual smoothness of lighting control through the cooperation of the confidence light effect correction module and the time sequence interpolation driving module. The system uses the data dispersion index to construct an inverse weighted function, dynamically calculates the confidence weight of the current data, and introduces historical lighting parameters for weighted fusion iterative operation. This means that when there is conflict or instability in the data, the system will automatically reduce the weight of the current data and increase the inertia of the historical state, avoiding flickering of the light due to sensor data jitter. At the same time, the time sequence interpolation driving module constructs a to-be-executed queue and generates a transition step sequence, converts the target light effect parameter into a continuous change pulse width modulation instruction, ensures linear transition of the lighting device in color temperature and brightness adjustment, solves the visual abruptness caused by step control signal in the prior art, and improves the comfort of human eye perception.

[0017] Of course, implementing any product of the present application does not necessarily require all the advantages described above. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 The landscape lighting intelligent adjustment system based on tone and motion state fusion analysis of the present application is shown in the flowchart. DETAILED DESCRIPTION

[0019] The landscape lighting intelligent adjustment system based on tone and motion state fusion analysis of the present application solves the technical problems of existing landscape lighting systems being easily affected by single environmental signal interference leading to misjudgment and harsh light effect switching process.

[0020] The general idea of the scheme in the embodiments of the present application is as follows: the system first performs feature dimension reduction on the collected original audio and motion trajectory data through digital signal processing technology, extracts normalized tone intensity and motion intensity; then calculates the synergistic gain and ratio deviation of the two in the data fusion layer to quantify the activity level of the scene and the reliability of the data; then weights and iterates the preset lighting parameters and historical states based on the data reliability to generate target control parameters that take into account sensitivity and stability; finally, the linear interpolation algorithm is used to map the target parameters to continuous pulse width modulation timing signals to drive the lighting device to perform smooth light effect adjustment, thereby realizing accurate matching of the light environment and the emotional behavior of the crowd.

[0021] Please refer to Figure 1 The embodiments of the present application provide a technical scheme: a landscape lighting intelligent adjustment system based on tone and motion state fusion analysis, comprising the following modules: a feature extraction and quantification module for performing short-time Fourier transform on the collected audio frame sequence, extracting the fundamental frequency mean and frequency fluctuation variance in the frequency domain as tone feature components, and performing normalization processing to generate tone intensity values, while performing differential operation on the collected motion target coordinate sequence, extracting the displacement speed mean and direction change rate in unit time as motion feature components, and performing normalization processing to generate motion intensity values; a synergistic interaction calculation module for linearly weighting and summing the tone intensity values and the motion intensity values to obtain an interaction synergistic gain coefficient representing tone-motion synchronization, determine the basic activity level of the current scene, and calculate the ratio deviation between the tone intensity values and the motion intensity values to generate a data dispersion index; a confidence light effect correction module for indexing the preset initial lighting parameters with the basic activity level, and substituting the data dispersion index into an inverse proportion weighting function to calculate the reliability weight of the current data, and performing weighted fusion iteration operation on the initial lighting parameters and historical lighting parameters to generate target color temperature values and target brightness values; a timing interpolation driving module for converting the target color temperature values and the target brightness values into target pulse width modulation values, constructing a to-be-executed queue containing the current pulse width modulation values and the target pulse width modulation values, generating a transition step sequence between the current and target values through a linear interpolation algorithm, and outputting pulse width modulation control instructions in a preset clock period.

[0022] In this embodiment, the feature extraction and quantification module is mainly responsible for converting the simulation signals of the physical world into standardized digital feature vectors that can be processed by computers. The module first performs a short-time Fourier transform on the collected audio frame sequence, which is a mathematical transformation method that maps time-domain signals to frequency-domain, used to analyze the frequency composition of sound signals within different time windows, so as to accurately extract the mean frequency and frequency fluctuation variance representing the emotional tone of the sound. At the same time, the module performs differential operation on the motion target coordinate sequence, that is, calculates the derivative of the coordinate point with respect to time, so as to obtain the instantaneous displacement speed and direction change rate of the target. In order to eliminate the difference in physical dimension between audio frequency and motion speed, the module uses a normalization processing algorithm to map the above feature components to the dimensionless interval of zero to one, outputting the tone intensity value and the motion intensity value. The technical role of this module is to provide a unified standard data input for subsequent data fusion, solving the technical problem of multi-modal heterogeneous data that is difficult to directly perform mathematical operations. The collaborative interaction calculation module is the core operation unit of the system, used to evaluate the activity level and logical consistency of the current scene. The module performs linear weighted summation operation on the input tone intensity value and motion intensity value, synthesizes the interaction collaborative gain coefficient representing the sound-motion synchronization through the preset weight coefficient, which directly reflects the comprehensive energy level of the scene, used to determine the basic activity level of the current scene. At the same time, the module calculates the ratio deviation between the two to generate the data dispersion index, which is a statistical index used to quantify the deviation between the sound features and the motion features. For example, in a scene with loud noise but no motion, the index will be significantly increased. The technical role of this module is to distinguish between effective active scenes and invalid noise interference scenes from a mathematical logic point of view, ensuring the semantic understanding ability of the system in complex environments. The confidence light effect correction module is mainly used to generate the final lighting control parameters that take into account sensitivity and stability. The module first uses the basic activity level as the index key to retrieve the corresponding initial lighting parameters in the preloaded database, that is, the color temperature and brightness that theoretically best match the current activity level. Then, the module substitutes the data dispersion index into the inverse weighted function, which is a nonlinear mathematical model whose logic is that the higher the dispersion, the lower the calculated weight value, thereby reducing the trustworthiness of the current conflicting data. The module uses the calculated confidence weight to perform weighted fusion iteration operation on the current initial lighting parameters and the historical lighting parameters at the last time, that is, uses the inertia of historical data to smooth the fluctuations of current data. The technical role of this module is to build an adaptive filtering mechanism based on data reliability, effectively preventing the lighting state from flickering dramatically due to environmental sudden noise or sensor jitter. The time sequence interpolation driving module is mainly responsible for converting the target parameters at the logical level into time sequence control signals that can be executed by the underlying hardware. The module first quantizes and converts the target color temperature value and the target brightness value into target pulse width modulation values, which is a technology that controls the power of analog circuits by adjusting the duty cycle of digital signals.Then, a module constructs a to-be-executed queue containing the current value and the target value, and applies a linear interpolation algorithm to calculate a series of intermediate step values between the current value and the target value according to a preset clock cycle, to generate a smooth transition step sequence. The technical effect of the module is to solve the step problem of the digital control signal in state switching, realize the continuous smooth change of the lighting light effect on the time axis, and avoid the visual mutation that can be perceived by the human eye.

[0023] Specifically, the specific process of performing short-time Fourier transform on the collected audio frame sequence, extracting the fundamental frequency mean and frequency fluctuation variance in the frequency domain as the tonal feature component, and performing normalization processing to generate the tonal intensity value is as follows: the audio frame sequence is framed and windowed by a Hanning window function, and fast Fourier transform is performed to obtain the frequency spectrum amplitude spectrum of each frame, the fundamental frequency peak point is located by a spectrum peak search algorithm, the numerical average value of the fundamental frequency peak points of continuous multiple frames is analyzed as the fundamental frequency mean, and the second-order central moment of the fundamental frequency peak point on the time axis is quantified as the frequency fluctuation variance; the fundamental frequency mean and the frequency fluctuation variance are projected into a preset emotional feature mapping interval, the projected feature data is compressed into the interval of zero to one by a Sigmoid activation function, and the compressed fundamental frequency data and variance data are combined by a weighted aggregation operation, and the tonal intensity value representing the sound emotional activity is output.

[0024] In the embodiment, for the processing process of audio data, first, preprocessing and frequency domain conversion are performed. Since the original audio signal has non-stationarity in the time domain, direct truncation will cause spectral leakage phenomenon, therefore, the audio frame sequence is framed and windowed by a Hanning window function, the smoothness of the window function reduces the mutation at the signal truncation point, and suppresses the sidelobe interference. Subsequently, fast Fourier transform is performed to convert the time domain discrete signal into a frequency domain signal, and the frequency spectrum amplitude spectrum of each frame is obtained, this step converts the invisible sound waveform into frequency distribution data recognizable by a computer. Then, a spectrum peak search algorithm is used to locate the frequency point with the maximum energy in the amplitude spectrum as the fundamental frequency peak point, which represents the main pitch of the human voice. In order to quantify the emotional features of the sound, the system calculates the fundamental frequency mean and the frequency fluctuation variance respectively. The fundamental frequency mean reflects the overall pitch level of the sound, and the calculation formula is as follows: ; wherein, : the fundamental frequency mean; : the total number of audio frames in the statistical time window; : the sequence number index of the current audio frame; : the fundamental frequency peak point value of the i-th frame extracted by the peak search algorithm. : the fundamental frequency peak point value of the i-th frame extracted by the peak search algorithm. Meanwhile, the frequency fluctuation variance is used to represent the jitter degree of the sound, i.e. the stability of the emotion, and high variance usually corresponds to excited or panic emotions, and its calculation is based on the principle of second-order central moment, and the formula is as follows: ; wherein, : Frequency fluctuation variance (standard deviation is used here to unify the dimension series); , , and The definition is consistent with the aforementioned formula. Finally, in order to transform the above physical characteristics into control signals that the system can process, normalization mapping is required. The system uses the Sigmoid activation function to compress the feature data to the interval between zero and one. This function has good nonlinear saturation characteristics and can suppress the influence of extreme outliers. Subsequently, the fundamental frequency and variance data are combined through weighted aggregation operations to generate the tone intensity value, calculated as follows: ;in, The final output is the normalized pitch intensity value; The preset weighting coefficient for the fundamental frequency characteristic is determined by analyzing the correlation between pitch and activity in historical scene data; for example, a value of 0.6 is used. :Preset baseband reference center value; : The preset variance reference center value; The base of the natural logarithm. Through the above calculations, the system successfully transforms abstract vocal emotions into calculable numerical indicators.

[0025] Specifically, the process of performing differential operations on the collected coordinate sequence of the moving target, extracting the mean displacement velocity and the rate of change of direction per unit time as motion feature components, and then normalizing them to generate a motion intensity value is as follows: First-order differential operations are performed on adjacent coordinate points in the coordinate sequence of the moving target to analyze the Euclidean distance and obtain the magnitude of the instantaneous velocity vector. The arithmetic mean of the magnitudes of all instantaneous velocity vectors within a preset time window is calculated as the mean displacement velocity. Simultaneously, the geometric angle relationship between adjacent instantaneous velocity vectors is extracted, and the change in tangential angle of the motion trajectory is obtained analytically through the inverse cosine function. The cumulative value of the angle change per unit time is used as the rate of change of direction. The mean displacement velocity and the rate of change of direction are mapped to a standard dimensionless interval through max-min standardization. The standardized mean velocity is used as the active component, and the standardized rate of change of direction is used as the entropy component. The active component and the entropy component are weighted and fused to output a motion intensity value representing the intensity of group behavior.

[0026] In this implementation scheme, the core of the motion data processing lies in analyzing the dynamic characteristics of the target. First, first-order differential operations are performed on adjacent coordinate points in the target's coordinate sequence, i.e., instantaneous velocity is obtained by calculating the change in position over time. During this process, the magnitude of the instantaneous velocity vector is analyzed using the Euclidean distance formula, a standard mathematical method for calculating the straight-line distance between two points in a two-dimensional plane. The arithmetic mean within a preset time window is then used as the average displacement velocity, which directly reflects the movement rate of the crowd. Next, to capture the degree of disorder in group behavior (i.e., the entropy component), the system extracts the geometric angle relationship between adjacent instantaneous velocity vectors. The cosine value is calculated by the ratio of the vector dot product to the product of the magnitudes, and then the inverse cosine function is used to analyze the change in the tangential angle of the trajectory. The accumulated angle change per unit time is used as the rate of change of direction. This step aims to distinguish between two different high-energy-consuming states: straight-line running and disorderly play. To generate the final motion intensity value, the system uses a minimax normalization method to map the above features with different physical dimensions to a standard dimensionless interval and performs weighted fusion. This calculation process considers not only the magnitude of the velocity but also the complexity of the direction. The specific calculation formula is as follows: ;in, The final output is the normalized motion intensity value; Weighting coefficient for the mean displacement velocity; The weighting coefficient of the rate of change of direction, and The weights are determined based on the scenario's pre-defined preferences for speed sensitivity and order sensitivity; The calculated average displacement velocity; Preset minimum speed threshold; Preset maximum speed threshold; The index of the time sampling point; T: the total number of times within the sampling window; At that moment The instantaneous velocity vector; At that moment The instantaneous velocity vector; Represents the modulo operation of vectors; Represents the dot product operation of vectors; : The preset upper limit constant for the normalization of the rate of change of direction. Using this formula, the system can comprehensively determine whether the target is moving rapidly in a straight line or in violent, disorderly motion, thus outputting an accurate indicator of the intensity of group behavior.

[0027] Specifically, the sound intensity value and the motion intensity value are linearly weighted and summed to obtain an interaction synergy gain coefficient representing sound-motion synchronization, and the specific process of determining the basic activity level of the current scene is as follows: the preset environmental bias weight coefficient is called, the sound intensity value and the motion intensity value are respectively given priority weight, the weighted sum value of the two is analyzed as an initial synergy modulus, the sound intensity value and the motion intensity value are coupled and modulated to generate a coupling gain term, and the initial synergy modulus and the coupling gain term are superimposed to generate an interaction synergy gain coefficient; a multi-level threshold decision logic is constructed, and the interaction synergy gain coefficient is compared with the preset quiet threshold, interaction threshold and carnival threshold in turn, and when the coefficient falls into different threshold intervals, the corresponding discrete state identifier is output as the basic activity level of the current scene.

[0028] In the embodiment, the core of the processing process of the synergy interaction calculation module is to fuse multi-modal data through mathematical means to evaluate the real activity level of the scene. First, the system calls the preset environmental bias weight coefficient, which is a parameter preset according to the physical properties of the application scene, for example, a higher weight is given to the sound tone in the music square scene, and a higher weight is given to the motion in the sports park scene. After giving the sound intensity value and the motion intensity value priority weight, the weighted sum value of the two is analyzed as an initial synergy modulus, which represents the basic energy at the linear superposition level. In order to further capture the characteristics of sound-motion synchronization, the system introduces a coupling modulation mechanism, that is, the sound intensity value and the motion intensity value are multiplied to generate a coupling gain term. The role of this step is to use the nonlinear amplification characteristics of multiplication. Only when sound and motion are both high, the coupling term will increase significantly, thus rewarding synchronous active behavior. The initial synergy modulus and the coupling gain term are superimposed to generate the final interaction synergy gain coefficient, and the calculation formula is as follows: ; wherein, : interaction synergy gain coefficient, used to quantify the comprehensive active energy of the current scene; : environmental bias weight coefficient of sound intensity value; : environmental bias weight coefficient of motion intensity value, and : sound intensity value generated in the previous step; : motion intensity value generated in the previous step; : preset synchronization gain factor, which is used to adjust the degree of reward for the sound-motion synchronization phenomenon, and is usually between 0.3 and 0.5. After obtaining the continuous gain coefficient, the system constructs a multi-level threshold decision logic to discretize it. The interactive synergy gain coefficient is compared with the preset quiet threshold, interactive threshold and carnival threshold in turn. This processing method is similar to the quantization process in digital signal processing, which maps the continuously changing values to state labels (such as state 0, state 1, state 2) that are easy for computers to process. When the coefficient falls into different threshold intervals, the corresponding discrete state identifier is output as the basic activity level of the current scene, thereby providing a definite index basis for subsequent table lookup operations.

[0029] Specifically, the specific process of simultaneously calculating the ratio deviation between the tone intensity value and the motion intensity value to generate the data dispersion index is as follows: the difference modulus between the tone intensity value and the motion intensity value is analyzed as the modal absolute error, and the arithmetic sum of the tone intensity value and the motion intensity value is calculated as the modal total intensity; the modal total intensity is superimposed with a regularization constant to generate a denominator item, and the numerical relationship between the modal absolute error and the denominator item is calculated by relative ratio quantization to generate a relative deviation ratio; the relative deviation ratio is input into a preset exponential amplification function, the input deviation signal is nonlinearly amplified, and the amplified value is output as the data dispersion index.

[0030] In the calculation process of the data dispersion index in the embodiment, the system aims to identify the logical conflicts between multi-modal data through statistical methods. The system first analyzes the difference modulus between the tone intensity value and the motion intensity value, that is, calculates the absolute value of the difference between the two as the modal absolute error, which intuitively reflects the deviation degree of the two-dimensional data. At the same time, in order to eliminate the influence of signal intensity itself on error judgment (for example, a small difference at low intensity should not be considered as a conflict), the system calculates the arithmetic sum of the tone intensity value and the motion intensity value as the modal total intensity. Subsequently, the relative ratio quantization is performed for normalization, the modal total intensity is superimposed with a regularization constant to generate a denominator item, and the numerical relationship between the modal absolute error and the denominator item is calculated by division operation to generate a relative deviation ratio. Finally, in order to highlight the abnormal state of serious misadjustment, the relative deviation ratio is input into a preset exponential amplification function for nonlinear gain amplification, and the calculation formula is as follows: ; wherein, : the final output data dispersion index, the higher the value, the more unreliable the data; : preset exponential amplification ratio coefficient, used to adjust the output dynamic range of the dispersion index; : natural logarithm base, which uses the growth characteristics of the exponential function to enlarge the numerical gap between low conflicts and high conflicts; : modal absolute error, representing the original gap of sound-motion data; : Total intensity of the modality, as a normalized reference; : Regularization constant, usually takes a very small positive number (e.g. 0.001), mainly to prevent overflow when the tone and motion intensity are both zero, and to smooth the noise in low energy range. Through this formula, the system can quantify the degree of sound-painting mismatch as a high-sensitivity numerical indicator, so that in abnormal situations such as someone shouting loudly without body movement (e.g. noise), a high dispersion index is output to suppress the false triggering of the light.

[0031] Specifically, the specific process of retrieving the preset initial lighting parameters indexed by the basic activity level and substituting the data dispersion index into the inverse weighting function to calculate the credibility weight of the current data is as follows: an association database containing the activity level field and the lighting parameter field is established, the corresponding color temperature value and brightness value are extracted as the initial lighting parameters through matching query in the database by taking the basic activity level as the key value; inverse weighting calculation logic is constructed, a preset sensitivity coefficient and a reference constant are set, the sensitivity coefficient and the data dispersion index are multiplied to generate an attenuation term, the reference constant and the attenuation term are superimposed to generate a denominator, and the credibility weight of the current data is output through the reciprocal mapping of the reference constant and the denominator.

[0032] In the parameter retrieval and weight calculation process of the confidence light effect correction module in the embodiment, the parameter matching based on the lookup table is first performed. The system establishes an association database containing the activity level field and the lighting parameter field. The database is essentially a key-value mapping table in which expert strategies are pre-stored. Through matching query in the database by taking the basic activity level as the key value, the color temperature value and the brightness value that are theoretically most matched with the current scene atmosphere can be quickly extracted as the initial lighting parameters. This step converts the discrete scene label into a specific physical control target, but the parameters at this time have not been verified for environmental stability. In order to prevent control errors caused by data conflicts, the system constructs inverse weighting calculation logic to quantify the credibility of the current data. The mathematical meaning of inverse weighting is that when the dispersion (i.e. the conflict degree) of the input signal increases, the corresponding output weight should show a nonlinear downward trend. The system sets a preset sensitivity coefficient and a reference constant, multiplies the sensitivity coefficient and the data dispersion index generated in the previous step to generate an attenuation term. The greater the value of the attenuation term, the stronger the inconsistency of the current sound-motion data. Then, the reference constant and the attenuation term are superimposed to generate a denominator, and the credibility weight of the current data is output through the reciprocal mapping of the reference constant and the denominator. This calculation process realizes soft suppression of abnormal data, and the formula is as follows: ; wherein, : Credibility weight of the current data, which is limited to a range of zero to one, and the closer the value is to one, the more credible the data is; : preset reference constant, used to control the convergence base of the weight function, usually taking the value of 1.0; : preset sensitivity coefficient, used to adjust the sensitivity of the system to data conflicts, which is determined by experiments, i.e. the reciprocal of the mean of the dispersion measured under standard noise environment; : data dispersion index calculated in the previous step. Through the formula, when the dispersion index is extremely high, the denominator increases rapidly, resulting in the credibility weight tending to zero, thus realizing the automatic shielding of low-quality instructions at the algorithm level.

[0033] Specifically, the specific process of generating the target color temperature value and the target brightness value through the weighted fusion iterative operation of the initial lighting parameter and the historical lighting parameter is as follows: the output value of the previous control period is read from the system register as the historical lighting parameter, the inertia retention weight is obtained through the complementary calculation of the unit value and the credibility weight; the initial lighting parameter and the credibility weight are executed to generate a new state component, and the historical lighting parameter and the inertia retention weight are executed to generate a historical state component; the vector aggregation is executed on the new state component and the historical state component, and the fused color temperature data and brightness data are respectively output as the target color temperature value and the target brightness value.

[0034] In the embodiment, in the generation process of the target light efficiency parameter, the system adopts the weighted fusion iterative operation to realize the time domain smoothing of the control signal. This process is essentially a digital signal processing link based on the principle of first-order lag filter. First, the output value of the previous control period is read from the system register as the historical lighting parameter, which represents the current actual physical state of the device. Then, the inertia retention weight is calculated by subtracting the credibility weight from the unit value (i.e. the value 1). The inertia retention weight here reflects the tendency of the system to maintain the current state: the lower the credibility, the greater the inertia. Subsequently, vector aggregation is performed on the color temperature and brightness data respectively. The system executes weighted modulation on the initial lighting parameter (the current target to be changed) and the credibility weight to generate a new state component, and executes weighted modulation on the historical lighting parameter (the past state) and the inertia retention weight to generate a historical state component. Finally, vector addition operation is performed on the new state component and the historical state component, and the fused color temperature data and brightness data are respectively output as the target color temperature value and the target brightness value. The calculation formula is as follows: ; wherein, : target color temperature value or target brightness value output in the i-th control period; : credibility weight calculated in the current period; : initial lighting parameter retrieved in the current period, i.e. the theoretical target value; : target color temperature value or target brightness value output in the i-th control period; : credibility weight calculated in the current period; the historical lighting parameter output in the control cycle, i.e. the system state value at the last moment; 1: represents the unit constant of the complete weight. Through the formula, the system realizes a dynamic soft switching mechanism: when the scene recognition is accurate (high reliability), the system quickly responds and approaches the initial lighting parameter; when the scene recognition is doubtful (low reliability), the system mainly follows the historical parameter, maintains the stability of the light environment, and thus effectively solves the light effect flicker problem caused by sensor jitter.

[0035] Specifically, the specific process of converting the target color temperature value and the target brightness value into target pulse width modulation values and constructing a to-be-executed queue containing the current pulse width modulation value and the target pulse width modulation value is as follows: the preset color temperature mixed light proportion algorithm is called, the power distribution ratio of the warm color temperature channel and the cold color temperature channel is calculated according to the target color temperature value, the target duty cycle of each channel is mapped in combination with the target brightness value, and the target pulse width modulation value is quantized; the pulse width modulation value being output in the register at the current moment is read as the current node, the target pulse width modulation value is taken as the terminal node, and a bidirectional linked list containing the start time stamp, the current node and the terminal node is constructed as the to-be-executed queue.

[0036] In the conversion and queue construction process of the pulse width modulation value in the timing interpolation driving module in the embodiment, the core is to map the abstract light effect parameter into a specific digital circuit control signal. The system first calls the preset color temperature mixed light proportion algorithm, which is based on the double color temperature mixed light principle, i.e. synthesizing the target color temperature by adjusting the light power proportion of the warm color temperature LED lamp bead and the cold color temperature LED lamp bead. The power distribution ratio of the warm color temperature channel and the cold color temperature channel is calculated according to the input target color temperature value, which ensures that the synthesized light color is consistent with the expected emotional atmosphere. Then, the target duty cycle of each channel is mapped in combination with the target brightness value, and the target pulse width modulation value is quantized. Pulse width modulation (PWM) is a technology for controlling the output power of an analog circuit by changing the high level duration (i.e. duty cycle) of a digital pulse signal. The calculation formula of the conversion process is as follows: ; ; wherein, : the target pulse width modulation value of the cold color temperature channel; : the target pulse width modulation value of the warm color temperature channel; : the target brightness value generated in the previous step; : the target color temperature value generated in the previous step; : the lowest warm color temperature parameter supported by the lamp hardware, for example, 2700; : the highest cold color temperature parameter supported by the lamp hardware, for example, 6500; : preset hardware calibration coefficient, used to compensate the difference of luminous efficiency of different batches of LED lamp beads, which is determined by factory light color calibration test. After numerical quantization, the system reads the pulse width modulation value being output in the current time register as the current node, and takes the calculated target pulse width modulation value as the termination node. The system constructs a bidirectional linked list containing the start time stamp, the current node and the termination node as the to-be-executed queue. The bidirectional linked list is a data structure that allows the system to quickly insert new high-priority instructions or delete expired instructions during the light efficiency change process, ensuring the real-time and flexibility of the control logic.

[0037] Specifically, the specific process of generating a transition step sequence between the current and target values by the linear interpolation algorithm and sequentially outputting the pulse width modulation control instructions according to the preset clock period is as follows: calculating the value difference between the termination node and the current node, obtaining the preset total transition time, and obtaining the single-step change slope by ratio analysis of the value difference and the number of clock periods contained in the total transition time; initializing the accumulator, and at the rising edge of each preset clock period, performing cumulative iteration of the output value of the last period and the single-step change slope to generate the intermediate step value at the current time; converting the intermediate step value into a binary control signal and writing it into the driving register through the general input and output interface to output the pulse width modulation control instruction.

[0038] In the linear interpolation and instruction output process in the embodiment, the purpose is to generate a continuous and smooth intermediate state through an algorithm to solve the problem of light efficiency mutation. The system first calculates the value difference between the termination node and the current node, and obtains the preset total transition time, which is usually set according to the adaptation characteristics of the human eye to light changes. Then, the single-step change slope is obtained by ratio analysis of the value difference and the number of clock periods contained in the total transition time. This step is essentially to calculate the small step size by which the PWM value should be increased or decreased in each clock period, and the calculation formula is as follows: ; wherein, : the calculated single-step change slope, which can be a floating-point number; : the termination node value (i.e. the target PWM value) in the to-be-executed queue; : the current node value (i.e. the current PWM value); : the preset total transition time, for example, 3 seconds, which determines the speed of light efficiency change; : the preset clock period frequency of the system, i.e. the number of updates performed per second. Then, the accumulator is initialized, which is a memory variable used to store intermediate calculation results. At the rising edge of each preset clock period, the system performs cumulative iteration operation to add the output value of the last period and the single-step change slope to generate the intermediate step value at the current time. This iteration process realizes the linear gradient of the value, and the formula is as follows: ; wherein, :No. The intermediate step value calculated over one clock cycle; :No. The value over one clock cycle (initial value) ; The previously calculated single-step change slope is used. Finally, the system converts the calculated intermediate step values ​​into binary control signals and writes them to the driver register via a general-purpose input / output interface (GPIO). The driver register is the physical unit that directly controls the on / off state of the hardware circuit. As the register value is periodically refreshed, the pulse width output by the LED driver circuit gradually changes, thereby achieving a seamless and smooth transition of lighting effect in physical space.

[0039] In summary, this application has at least the following effects:

[0040] The intelligent landscape lighting adjustment system based on the fusion analysis of tone and motion states achieves accurate quantification and authenticity verification of scene atmosphere by normalizing and extracting tone emotion and motion intensity and performing collaborative interactive calculations. It utilizes the interactive collaborative gain coefficient and data dispersion index to effectively shield the interference of single-modal environmental noise. At the same time, it combines confidence-based light effect correction and time-series interpolation driving technology based on inverse weighting to dynamically balance the current response and historical inertia according to data credibility, generating continuous and smooth pulse width modulation control commands. While ensuring that the lighting logic is highly matched with the actual psychological state of the crowd, it completely solves the problems of false triggering and sudden changes in light effect caused by signal jitter or logic conflict in traditional control systems, significantly improving the intelligent interactive stability and visual comfort of landscape lighting.

[0041] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0042] This invention is described with reference to flowchart illustrations and / or block diagrams of systems, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1one or more processes and / or blocks Figure 1 apparatus for performing the functions specified in the block or blocks.

[0043] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the Figure 1 one or more processes and / or blocks Figure 1 the functions specified in the block or blocks.

[0044] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions that are executed on the computer or other programmable apparatus provide steps for implementing the Figure 1 one or more processes and / or blocks Figure 1 the steps of the functions specified in the block or blocks.

[0045] While the preferred embodiments of the application have been described, additional variations and modifications can be made to the embodiments by those skilled in the art once they learn of the basic inventive concepts. Therefore, the appended claims are intended to encompass within their scope all such variations and modifications as are included within the scope of the invention.

[0046] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A landscape lighting intelligent adjustment system based on tone and motion state fusion analysis, characterized in that, The method comprises the following modules: A feature extraction and quantification module is configured to perform short-time Fourier transform on the collected audio frame sequence, extract the mean of the fundamental frequency and the variance of the frequency fluctuation in the frequency domain as the tonal feature component, and perform normalization processing to generate the tonal intensity value, and simultaneously perform differential operation on the collected motion target coordinate sequence, extract the mean of the displacement speed and the direction change rate in unit time as the motion feature component, and perform normalization processing to generate the motion intensity value; A collaborative interaction calculation module is configured to perform linear weighted summation on the tonal intensity value and the motion intensity value, obtain the interaction collaborative gain coefficient representing the sound motion synchronization, determine the basic activity level of the current scene, and simultaneously calculate the ratio deviation between the tonal intensity value and the motion intensity value to generate the data dispersion index; A confidence light effect correction module is configured to use the basic activity level as an index to retrieve the preset initial lighting parameter, calculate the confidence weight of the current data by substituting the data dispersion index into the inverse proportion weighted function, and perform weighted fusion iterative operation on the initial lighting parameter and the historical lighting parameter to generate the target color temperature value and the target brightness value; A time sequence interpolation driving module is configured to convert the target color temperature value and the target brightness value into target pulse width modulation values, construct a to-be-executed queue containing the current pulse width modulation value and the target pulse width modulation value, generate a transition step sequence between the current value and the target value by a linear interpolation algorithm, and output the pulse width modulation control instruction in a preset clock period. 2.The landscape lighting intelligent adjustment system based on the fusion analysis of tonal and motion state according to claim 1, characterized in that: The specific process of performing short-time Fourier transform on the collected audio frame sequence, extracting the mean of the fundamental frequency and the variance of the frequency fluctuation in the frequency domain as the tonal feature component, and performing normalization processing to generate the tonal intensity value is as follows: The audio frame sequence is processed by framing and windowing through the Hanning window function, and the fast Fourier transform is performed to obtain the frequency spectrum amplitude spectrum of each frame. The fundamental frequency peak value point is located by the frequency spectrum peak value search algorithm, the numerical average value of the fundamental frequency peak value points of continuous multiple frames is analyzed as the mean of the fundamental frequency, and the second-order central moment of the fundamental frequency peak value point on the time axis is quantified as the frequency fluctuation variance; The mean of the fundamental frequency and the variance of the frequency fluctuation are projected into the preset emotional feature mapping interval, the projected feature data is compressed into the interval of zero to one through the Sigmoid activation function, and the compressed fundamental frequency data and variance data are combined through weighted aggregation operation, and the tonal intensity value representing the sound emotional activity is output. 3.The landscape lighting intelligent adjustment system based on the fusion analysis of tonal and motion state according to claim 1, characterized in that: The specific process of performing differential operation on the collected motion target coordinate sequence, extracting the mean of the displacement speed and the direction change rate in unit time as the motion feature component, and performing normalization processing to generate the motion intensity value is as follows: First-order differential operation is performed on adjacent coordinate points in the motion target coordinate sequence, the Euclidean distance is analyzed to obtain the instantaneous speed vector module length, and the arithmetic average value of all instantaneous speed vector module lengths in a preset time window is calculated as the mean of the displacement speed; Meanwhile, the geometric angle relationship between adjacent instantaneous speed vectors is extracted, the tangent angle change amount of the motion trajectory is analyzed through the inverse cosine function, and the cumulative value of the angle change amount in unit time is taken as the direction change rate; The displacement velocity average and the direction change rate are mapped to a standard dimensionless interval by maximum minimum standardization, the standardized velocity average is taken as an active component, the standardized direction change rate is taken as an entropy value component, the active component and the entropy value component are weighted and fused, and a motion intensity value representing the intensity of group behavior is output. 4.The landscape lighting intelligent adjustment system based on the fusion analysis of tonal and motion state according to claim 1, characterized in that: The specific process of determining the basic active level of the current scene by linearly weighting and summing the tone intensity value and the motion intensity value is as follows: The preset environmental bias weight coefficient is called to give the tone intensity value and the motion intensity value priority weight respectively, the weighted sum value of the two is analyzed as an initial cooperative modulus, the tone intensity value and the motion intensity value are coupled and modulated to generate a coupling gain term, and the initial cooperative modulus and the coupling gain term are superimposed to generate an interactive cooperative gain coefficient; A multi-level threshold value decision logic is constructed, and the interactive cooperative gain coefficient is compared with the preset quiet threshold value, the interactive threshold value and the carnival threshold value in turn, and when the coefficient falls into different threshold value intervals, the corresponding discrete state identifier is output as the basic active level of the current scene. 5.The landscape lighting intelligent adjustment system based on the fusion analysis of tonal and motion state according to claim 4, characterized in that: The specific process of simultaneously calculating the ratio deviation between the tone intensity value and the motion intensity value to generate a data dispersion index is as follows: The difference modulus between the tone intensity value and the motion intensity value is analyzed as a modal absolute error, and the arithmetic sum of the tone intensity value and the motion intensity value is calculated as a modal total intensity; The modal total intensity and the regularization constant are superimposed to generate a denominator term, and the relative deviation ratio is generated by quantifying the numerical relationship between the modal absolute error and the denominator term; The relative deviation ratio is input into the preset exponential amplification function, the input deviation signal is nonlinearly amplified, and the amplified value is output as the data dispersion index. 6.The landscape lighting intelligent adjustment system based on the fusion analysis of tonal and motion state according to claim 1, characterized in that: The specific process of taking the basic active level as an index to retrieve the preset initial lighting parameter and substituting the data dispersion index into an inverse proportion weighting function to calculate the credibility weight of the current data is as follows: An associated database containing an active level field and a lighting parameter field is established, the basic active level is used as a key value for matching query in the database, and the corresponding color temperature value and brightness value are extracted as the initial lighting parameter; An inverse proportion weighting calculation logic is constructed, a preset sensitivity coefficient and a reference constant are set, the sensitivity coefficient and the data dispersion index are multiplied to generate an attenuation term, the reference constant and the attenuation term are superimposed to generate a denominator, and the credibility weight of the current data is output by mapping the reference constant and the reciprocal of the denominator. 7.The landscape lighting intelligent adjustment system based on the fusion analysis of tonal and motion state according to claim 6, characterized in that: The specific process of performing weighted fusion iterative operation on the initial lighting parameter and the historical lighting parameter to generate a target color temperature value and a target brightness value is as follows: The output value of the last control period is read from the system register as the historical lighting parameter, and the complementary calculation of the unit value and the credibility weight is performed to obtain an inertia retention weight; The initial lighting parameter and the credibility weight are weighted and modulated to generate a new state component, and the historical lighting parameter and the inertia retention weight are weighted and modulated to generate a historical state component; The vector aggregation is performed on the new state component and the historical state component, and the fused color temperature data and the brightness data are respectively output as a target color temperature value and a target brightness value. 8.The landscape lighting intelligent adjustment system based on the fusion analysis of tonal and motion state according to claim 1, characterized in that: The target color temperature value and the target brightness value are converted into a target pulse width modulation value, and a specific process of constructing a to-be-executed queue including the current pulse width modulation value and the target pulse width modulation value is as follows: A preset color temperature mixed light proportion algorithm is called, a power distribution ratio of a warm color temperature channel and a cold color temperature channel is calculated according to the target color temperature value, and a target duty cycle of each channel is mapped in combination with the target brightness value to quantize the target pulse width modulation value; A pulse width modulation value being output in a register at a current time is read as a current node, and the target pulse width modulation value is taken as a terminal node, and a bidirectional linked list including a start time stamp, the current node and the terminal node is constructed as the to-be-executed queue. 9.The landscape lighting intelligent adjustment system based on the fusion analysis of tonal and motion state according to claim 8, characterized in that: A specific process of generating a transition step sequence between the current value and the target value through a linear interpolation algorithm and sequentially outputting pulse width modulation control instructions according to a preset clock period is as follows: A value difference between the terminal node and the current node is calculated, a preset transition total time length is obtained, a single-step change slope is derived through a ratio resolution of the value difference and a clock period number included in the transition total time length; An accumulator is initialized, and at the arrival of a rising edge of each preset clock period, an output value of a last period and the single-step change slope are executed to generate an intermediate step value at a current time. The intermediate step value is converted into a binary control signal, and is written into a driving register through a general input and output interface to output the pulse width modulation control instruction.

Images