System and method for controlling an array of mems micromirrors

Abstract

The application relates to the technical field of micro-electro-mechanical system control, and discloses a system and method for controlling a MEMS micromirror array. The method comprises the following steps: establishing a space-time characteristic evaluation model comprising an energy flow analysis domain and a mechanical response domain, synchronously collecting the input and load energy waveform characteristics of the system, and acquiring the stress distribution data of the driving mechanism and the real-time attitude data of the reflecting surface of each micromirror in parallel. The data are fused to generate a multi-dimensional feature vector, which is compared with preset stability boundary conditions, so that a targeted regulation strategy is output. The method realizes multi-dimensional state perception from energy transmission efficiency to internal stress of the mechanical structure, overcomes the hysteresis of the traditional method which only depends on motion feedback, can identify the performance degradation trend and potential mechanical damage in advance, makes the control system change from passive error correction to active early warning and health management, and improves the working stability, adaptability and service life of the micromirror array.

Description

Technical Field

[0001] This invention relates to the field of microelectromechanical systems (MEMS) control technology, specifically to a system and method for controlling MEMS micromirror arrays. Background Technology

[0002] MEMS (Micromirror Arrays) are key components in systems such as optical communication and lidar, and their dynamic performance and long-term reliability directly affect the overall system efficiency. Currently, the control of these devices is mainly based on a simple "electrical signal input - mirror angle feedback" control mode. This method uses a preset drive voltage or current waveform, combined with optical sensors to monitor the actual deflection angle of the mirror, to form a closed-loop control. Its core technology lies in how to quickly eliminate angle deviation by adjusting electrical parameters.

[0003] This traditional control method has shortcomings. The system treats the micromirror array as a black box, focusing only on external electrical characteristics and the final motion result, failing to observe the internal dynamic process. The control system lacks monitoring of energy transfer efficiency. Multiple loss paths exist in the process of converting input electrical energy into mirror mechanical energy; when abnormal responses or mechanical characteristic drift occur in the array, the energy loss pattern changes. Existing technology, unable to simultaneously compare the input energy and the effective load energy, struggles to identify these early signs of performance degradation and can only passively compensate after motion errors appear.

[0004] Current methods completely neglect real-time assessment of the mechanical load on the drive mechanism. During high-frequency movement of the micromirrors, their cantilever beams, torsion bars, and other structures bear cyclic stresses, making them prone to fatigue accumulation. Traditional control only focuses on whether the mirror is in position, remaining completely unaware of the stress distribution within the drive mechanism. This lack of critical mechanical state information prevents the system from implementing protective controls in the early stages of structural damage, limiting the array's lifespan and increasing the risk of sudden failures under high load conditions. Existing control technologies urgently need breakthroughs in achieving a deeper level of state perception. Summary of the Invention

[0005] The purpose of this invention is to provide a system and method for controlling MEMS micromirror arrays, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides a method for controlling a MEMS micromirror array, the method comprising:

[0007] A spatiotemporal characteristic evaluation model for the MEMS micromirror array is established, which includes an energy flow analysis domain and a mechanical response domain;

[0008] In the energy flow analysis domain, the feed energy waveform characteristics and load energy waveform characteristics of the MEMS micromirror array are simultaneously acquired;

[0009] In the mechanical response domain, stress distribution data of each micromirror driving mechanism and real-time attitude data of each micromirror reflecting surface in the MEMS micromirror array are acquired in parallel.

[0010] The feed energy waveform characteristics, the load energy waveform characteristics, the stress distribution data, and the real-time attitude data are input into the spatiotemporal characteristic evaluation model to generate a multidimensional feature vector;

[0011] Based on the comparison between the multidimensional feature vector and the preset stability boundary conditions, the control strategy of the MEMS micromirror array is output.

[0012] Preferably, the step of establishing the spatiotemporal characteristic evaluation model specifically includes:

[0013] Define the data sampling period of the energy flow analysis domain and the data sampling period of the mechanical response domain, and keep the two sampling periods synchronized;

[0014] An input sensor network for monitoring the waveform characteristics of the feed energy and an output sensor network for monitoring the waveform characteristics of the load energy are set in the energy flow analysis domain;

[0015] An array of micro-strain gauges for measuring the stress distribution data and an optical ranging unit for capturing the real-time attitude data are deployed in the mechanical response domain.

[0016] The data fusion core of the spatiotemporal characteristic evaluation model is configured to receive and time-align asynchronous data streams from the input sensor network, the output sensor network, the micro-strain gauge array, and the optical ranging unit.

[0017] Preferably, the step of synchronously acquiring the waveform characteristics of the input energy and the waveform characteristics of the load energy includes:

[0018] The voltage transient and current transient values ​​in the power supply circuit of the MEMS micromirror array are continuously recorded through the input sensor network.

[0019] The output sensor network continuously records the transient voltage and current values ​​in the load circuit of the MEMS micromirror array;

[0020] The recorded transient voltage and current values ​​of the power supply circuit are convolved to extract the waveform features of the fed-in energy.

[0021] The recorded transient values ​​of load circuit voltage and current are convolved to extract the load energy waveform characteristics.

[0022] Preferably, the step of acquiring stress distribution data and real-time attitude data in parallel includes:

[0023] The surface strain of each micromirror drive mechanism under different drive phases is measured by spatial interpolation using the microstrain gauge array.

[0024] The surface strain data is converted into the stress distribution data;

[0025] The optical ranging unit emits a measurement beam to each micromirror reflecting surface and receives the phase shift of the reflected beam.

[0026] The tilt angle and offset of each micromirror reflector relative to the reference position are calculated based on the phase offset to form the real-time attitude data.

[0027] Preferably, the step of generating multidimensional feature vectors includes:

[0028] The input energy waveform characteristics are decomposed in the frequency domain to obtain the energy input spectrum components;

[0029] The load energy waveform characteristics are decomposed in the frequency domain to obtain the energy output spectrum components;

[0030] Principal component analysis was performed on the stress distribution data to extract key stress modes;

[0031] Cluster analysis is performed on the real-time attitude data to identify the cooperative motion patterns of the micromirror group;

[0032] The energy input spectral component, the energy output spectral component, the key stress mode, and the cooperative motion mode are combined into the multidimensional feature vector.

[0033] Preferably, the step of comparing the multidimensional feature vector with the stability boundary condition includes:

[0034] The multidimensional feature vectors are mapped onto a pre-calibrated stability map;

[0035] Calculate the Mahalanobis distance between the multidimensional feature vector and the center of each stable region in the stability map;

[0036] Determine the target stable region that has the smallest Mahalanobis distance to the multidimensional feature vector;

[0037] Query the historical control parameter records associated with the target stable region;

[0038] The current control strategy is generated based on the historical control parameters recorded.

[0039] The step of mapping the multidimensional feature vector to a pre-calibrated stability map includes:

[0040] Load the pre-calibrated stability map, which is constructed from historical experimental data and contains multiple stable regions, each defined by a central eigenvector and a covariance matrix;

[0041] The multidimensional feature vector is standardized to have the same scale as the feature vector in the stability map.

[0042] Calculate the Mahalanobis distance between the standardized multidimensional feature vector and the central feature vector of each stable region, using the covariance matrix of the corresponding stable region as the weights in the Mahalanobis distance calculation.

[0043] Compare all the Mahalanobis distance values ​​and identify the stable region corresponding to the Mahalanobis distance with the smallest value as the target stable region.

[0044] Preferably, the output control strategy step further includes:

[0045] The control strategy is executed to adjust the driving signal parameters applied to the MEMS micromirror array;

[0046] After the adjustment is completed, a new round of data acquisition is performed on the input energy waveform characteristics, the load energy waveform characteristics, the stress distribution data, and the real-time attitude data.

[0047] The newly acquired data is input into the spatiotemporal characteristic evaluation model to generate a multidimensional feature vector for verification.

[0048] Verify whether the verification multidimensional feature vector falls into the target stable region;

[0049] If it does not fall into the target stable region, the control strategy is iteratively modified until the verification multidimensional feature vector falls into the target stable region.

[0050] Preferably, the step of generating the current control strategy based on the historical control parameter records includes:

[0051] Retrieve all historical control parameter records associated with the target stable region from the storage system. Each record contains historical control parameters and their corresponding system performance indicators.

[0052] Calculate the weighted score of the performance index for each historical control parameter record, the weighted score being determined based on the importance weight of the performance index and the historical usage frequency;

[0053] Select the historical control parameter record with the highest weighted score as the preferred record;

[0054] Historical control parameters are extracted from the preferred records and fine-tuned in combination with the current operating status of the MEMS micromirror array to generate the current control strategy, which includes amplitude, frequency and phase adjustment commands for the driving signal.

[0055] Preferably, the control strategy specifically includes a set of instructions for collaboratively correcting the driving waveform shape, driving frequency spectrum, and driving phase difference between the micromirrors of the MEMS micromirror array.

[0056] Preferably, the present invention also includes a system for controlling a MEMS micromirror array, the system including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor, when executing the computer program, implements the steps of the method for controlling a MEMS micromirror array as described above.

[0057] Compared with the prior art, the beneficial effects of the present invention are:

[0058] By synchronously acquiring the waveform characteristics of the input and load energy, dynamic monitoring of the system's energy transfer efficiency is achieved. Through precise comparison of the dynamic characteristics of input and output energy, abnormal energy consumption caused by microscopic performance changes can be keenly detected. This technology can identify early signals of characterizing unit failure or efficiency degradation before visible deviations occur in the mirror motion, transforming control behavior from delayed correction to proactive intervention based on energy efficiency assessment, thereby improving system stability.

[0059] Parallel acquisition of stress distribution and real-time attitude data of the drive mechanism and mirror surface establishes a direct correlation between the internal state of the machine and its motion performance. This technology enables real-time quantification of the micro-strain of the drive structure, allowing for precise assessment of the mechanical load state by analyzing the stress response under specific motion commands. When abnormal drift in the stress required to maintain the attitude is detected, changes in structural stiffness or potential damage can be promptly identified, enabling protective adjustments based on the machine's health status and extending device lifespan.

[0060] By fusing energy flow characteristics with mechanical response data to generate multi-dimensional feature vectors, a unified state assessment framework is constructed. This technology enables control strategies to become multi-objective integrated decisions for energy efficiency optimization and mechanical protection. Based on specific patterns of feature vector deviations from stability boundaries, the system can intelligently diagnose whether the root cause of a problem is abnormal energy efficiency or mechanical fatigue, and output targeted commands to achieve a synergistic improvement in system performance and reliability under complex operating conditions. Attached Figure Description

[0061] Figure 1 This is a schematic diagram illustrating the working principle of the method for controlling a MEMS micromirror array as described in this invention.

[0062] Figure 2 A flowchart for establishing a spatiotemporal characteristic evaluation model;

[0063] Figure 3 A flowchart for synchronously acquiring the waveform characteristics of the input energy and the load energy;

[0064] Figure 4 A three-dimensional visualization of stress distribution;

[0065] Figure 5 This is a diagram showing the changes in the orientation of the micromirror. Detailed Implementation

[0066] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0067] Please see Figure 1 This invention provides a method for controlling a MEMS micromirror array. The method includes: an energy flow analysis domain focusing on monitoring the energy input and output dynamics of the MEMS micromirror array, capturing in real-time feed energy waveform characteristics and load energy waveform characteristics using a high-precision sensor network. These characteristics reflect the energy transmission efficiency of the drive signal and the load matching state. A mechanical response domain addresses the mechanical structural behavior of the micromirrors, utilizing distributed sensing units to collect stress distribution data of each micromirror drive mechanism and real-time attitude data of each micromirror reflective surface in parallel, thereby comprehensively evaluating the mechanical stability and motion consistency of the micromirror array. A spatiotemporal characteristic evaluation model integrates the energy flow and mechanical response data through data fusion technology to generate a multi-dimensional feature vector that integrates key parameters of energy dynamics and mechanical state. Subsequently, the system compares this multi-dimensional feature vector with preset stability boundary conditions. These stability boundary conditions are pre-calibrated based on historical experimental data, defining the parameter range for the safe and stable operation of the micromirror array. The comparison process involves calculating the similarity between the feature vector and the stability region, ultimately outputting a control strategy. This strategy includes adjusting parameters such as the waveform, frequency, or phase of the drive signal to optimize the performance of the micromirror array.

[0068] Example 1: See Figure 2In specific implementation, defining the data sampling period for the energy flow analysis domain and the data sampling period for the mechanical response domain is the initial step. The data sampling period for the energy flow analysis domain is set to the microsecond level, for example, 1 microsecond, to accommodate high-frequency energy fluctuations, while the data sampling period for the mechanical response domain is set to the millisecond level, for example, 1 millisecond, to match mechanical inertia. The two sampling periods are synchronized through a global clock signal and a phase-locked loop. The phase-locked loop generates a synchronization pulse to ensure that the sampling events are aligned in time. In some embodiments, the data sampling period for the energy flow analysis domain can be dynamically adjusted according to the array's operating frequency, for example, automatically shortening the sampling period when the drive signal frequency increases. The data sampling period for the mechanical response domain is fixed at the millisecond level based on the micromirror response delay. The synchronization mechanism uses a digital phase-locked loop circuit, with a crystal oscillator as the reference clock source for the phase-locked loop, and the output synchronization signal is distributed to each sampling module. Optionally, the sampling period definition process includes a calibration phase. The calibration phase verifies the sampling synchronization accuracy by injecting a test signal to ensure the time consistency of the energy flow and mechanical response data.

[0069] In specific implementations, an input sensor network and an output sensor network are set up in the energy flow analysis domain. The input sensor network is deployed in the power supply loop of the MEMS micromirror array and includes a high-bandwidth differential voltage probe and a Hall effect current sensor. The voltage probe is connected to the positive and negative terminals of the power supply loop, and the current sensor is mounted around the power supply wire to continuously monitor the transient voltage and current values ​​in parallel. Similarly, the output sensor network is installed in the load loop, located at the output of the micromirror drive circuit. An opto-isolator is added to the output sensor network to prevent load feedback interference. Sensor data is transmitted to the processing unit via a high-speed serial interface such as SPI or I2C. In some embodiments, the layout of the input sensor network considers circuit impedance matching. The voltage probe uses a high input impedance design to reduce load effects, and the current sensor uses a Rogowski coil to improve frequency response characteristics. All sensor nodes are connected via a bus topology to achieve centralized data acquisition.

[0070] In practical implementation, a micro-strain gauge array and an optical ranging unit are deployed in the mechanical response domain. The micro-strain gauge array is attached to the surface of each micromirror drive mechanism in a grid pattern. The micro-strain gauge array uses foil resistance strain gauges, and the spatial arrangement of the strain gauges is based on the geometry of the drive mechanism. For example, a uniform grid is used for square mechanisms, with a grid spacing on the micrometer scale to achieve spatial interpolation measurements. The optical ranging unit is integrated around the array and includes a laser emitter and an interferometer detector. The laser emitter generates a coherent beam, which is guided to the micromirror reflecting surface by a beam splitter. The reflected beam is received by the detector, and the phase shift is resolved. Optionally, the micro-strain gauge array is mounted using an epoxy resin adhesive with temperature stability to reduce thermal drift. The optical path calibration of the optical ranging unit is completed through an autofocus mechanism to ensure that the beam is perpendicular to the reflecting surface.

[0071] In practical implementation, a data fusion core for the spatiotemporal characteristic evaluation model is configured. This core is implemented using a field-programmable gate array (FPGA) hardware. It receives asynchronous data streams from the input sensor network, output sensor network, micro-strain gauge array, and optical ranging unit. These asynchronous data streams are timestamped, generated by a global clock. The core executes a time alignment algorithm, which uses interpolation to compensate for sampling delays. For example, for the time difference between energy flow data and mechanical response data, a linear interpolation formula is applied to calculate the intermediate value. The interpolation formula is expressed as:

[0072]

[0073] in: Indicates time Aligned data values, and Adjacent sampling points and The original value, Align the target time point. The data fusion core has a built-in buffer to store temporary data, the logic unit performs interpolation operations, and outputs a synchronous dataset for subsequent feature generation.

[0074] Example 2: See Figure 3 In specific implementations, voltage and current transient values ​​in the power supply circuit of the MEMS micromirror array are continuously recorded via an input sensor network. The input sensor network includes a high-speed differential amplifier and an analog-to-digital converter (ADC). The differential amplifier is connected to the positive and negative terminals of the power supply circuit, and the ADC sampling rate is set to 10 MHz / s to capture microsecond-level transients. The voltage transient value is digitized by the ADC after amplifier gain adjustment, and the current transient value is converted into a voltage signal by a Rogowski coil and then sampled synchronously. In some embodiments, the recording process of the input sensor network employs a ping-pong buffer structure. The ping-pong buffer consists of two storage areas that work alternately. While one buffer stores data, the other buffer transmits data to the processor, ensuring uninterrupted continuous recording. It is understood that the recording of voltage and current transient values ​​in the power supply circuit needs to consider common-mode noise suppression. The differential amplifier is configured with a common-mode rejection ratio greater than 80 dB, and the ADC resolution is not less than 16 bits to ensure signal accuracy.

[0075] In practical implementation, the voltage and current transient values ​​in the load circuit of the MEMS micromirror array are continuously recorded through an output sensor network. The output sensor network is installed on the load circuit between the drive circuit and the micromirror array. A high-impedance probe is used to measure the voltage transient value of the load circuit to reduce load effects, while a current transformer is used to isolate the high-voltage side for the current transient value measurement. The sampling clock of the output sensor network is synchronized with that of the input sensor network, and the sampling clocks are distributed from the same clock source. The load circuit data recording undergoes digital filtering simultaneously, using a finite impulse response (FIR) filter to eliminate high-frequency noise. Optionally, oversampling technology is added to the load circuit recording, with the oversampling rate set to at least 4 times, followed by downsampling to improve the signal-to-noise ratio. It is understood that the recording of the load circuit voltage and current transient values ​​needs to be strictly synchronized with the power supply circuit data, with the synchronization error controlled at the nanosecond level.

[0076] In practice, convolution operations are performed on the recorded transient voltage and current values ​​of the load circuit to extract load energy waveform features. The load circuit convolution operation uses the same kernel function as the power supply circuit but is processed independently. The convolution result includes the time-frequency characteristics of the load energy waveform. A calibration coefficient compensation is added to the load energy waveform feature extraction process. The calibration coefficient is determined through prior calibration experiments and is used to eliminate sensor system errors. Optionally, normalization is performed after the load circuit convolution operation. Normalization scales the energy values ​​to a standard range for easier subsequent comparison. It can be understood that the load energy waveform features reflect the actual energy consumption state of the micromirror array, and together with the feed energy waveform features, constitute a complete dataset for energy flow analysis.

[0077] Example 3: In a specific implementation, the surface strain of each micromirror drive mechanism under different drive phases is measured using a micro-strain gauge array via spatial interpolation. The micro-strain gauge array employs foil resistance strain gauges attached to the surface of the drive mechanism in a grid pattern with a grid spacing of 50 micrometers to achieve spatial interpolation measurement. The spatial interpolation algorithm uses a bilinear interpolation method, calculating the strain value at any point within the region based on the readings of four adjacent strain gauges. The surface strain data is converted into a voltage signal via a Wheatstone bridge circuit, and the voltage signal is amplified and sampled by an analog-to-digital converter. In some embodiments, the surface strain measurement under different drive phases is synchronized with the drive signal, which is provided by the controller. The measurement data of the micro-strain gauge array is grouped and stored according to phase angle. It is understood that temperature compensation needs to be considered in the surface strain data measurement; the micro-strain gauge array integrates a temperature sensor, and the measurement data is corrected for temperature drift in real time.

[0078] In specific implementations, surface strain data is converted into stress distribution data. The conversion process is based on Hooke's Law, and the stress distribution data is calculated using the Young's modulus parameter of the material. For silicon-based MEMS materials, the Young's modulus is taken as 169 GPa. The mapping relationship between surface strain data and stress distribution data is a linear transformation, and the stress value at each measurement point is obtained by multiplying the strain value by the Young's modulus. Optionally, a smoothing filtering algorithm is used for post-processing of the stress distribution data. The smoothing filtering uses Gaussian kernel convolution to reduce the influence of noise. A measurement beam is emitted to each micromirror reflecting surface through an optical ranging unit, and the phase shift of the reflected beam is received. The optical ranging unit adopts a laser interferometer system, and the measurement beam is a helium-neon laser with a wavelength of 632.8 nm and a beam diameter focused to 5 μm. The phase shift of the reflected beam is received by a photodetector, and phase demodulation uses heterodyne interferometry. The relationship between the phase shift and the micromirror displacement is calculated using the optical path difference. In some embodiments, the measurement frequency of the optical ranging unit is 10 kHz, each micromirror reflecting surface is configured with an independent measurement optical path, and the phase shift data is transmitted to the processing unit in real time. It is understandable that real-time attitude data includes tilt angle and offset information. The tilt angle is calculated by measuring the phase offset difference at different points on the reflector.

[0079] In practice, real-time attitude data is generated by calculating the tilt angle and offset of each micromirror's reflecting surface relative to a reference position based on the phase shift. The reference position is determined through a calibration process, which records the initial phase value in a zero-drive state. The tilt angle is calculated using geometric optics formulas, and the offset is derived from the relationship between phase shift and wavelength. The real-time attitude data is stored in array form, containing the pitch and yaw angles of each micromirror. Optionally, a redundant measurement mechanism is used for real-time attitude data verification, where two measurement optical paths are arranged on the same micromirror reflecting surface to compare the results. It can be understood that accurate real-time attitude data reflects the motion state of the micromirrors and is the foundation for cooperative motion pattern recognition.

[0080] In specific implementation, the input energy waveform features are decomposed in the frequency domain to obtain the energy input spectral components. The frequency domain decomposition employs a Fast Fourier Transform (FFT) algorithm with 1024 FFT points, and the sampling frequency matches the sampling rate of the energy waveform features. The energy input spectral components contain amplitude and phase spectrum information, and the amplitude and phase of the main frequency components are extracted as feature vectors. In some embodiments, the input energy waveform features are windowed before frequency domain decomposition, and a Hanning window is used to reduce spectral leakage. It can be understood that the energy input spectral components characterize the frequency distribution characteristics of the driving energy. The load energy waveform features are then decomposed in the frequency domain to obtain the energy output spectral components. The extraction process for the energy output spectral components is the same as that for the energy input spectral components but processed independently. The focus of the energy output spectral components is on the frequency components that differ from the energy input spectral components. The frequency domain decomposition results are logarithmically scaled to enhance the dynamic range, and the amplitude ratio of the main harmonic components is used as a feature quantity. Optionally, the energy output spectral component analysis includes impedance matching index calculation, which reflects energy transmission efficiency. Principal component analysis (PCA) is performed on the stress distribution data to extract key stress modes. The PCA algorithm calculates the covariance matrix of the stress distribution data and solves for the eigenvalues ​​and eigenvectors of the covariance matrix. The eigenvectors corresponding to the largest eigenvalues ​​of the key stress modes are then used to retain principal components with a cumulative contribution rate exceeding 85%. The stress distribution data is preprocessed by standardization to eliminate the influence of dimensions. Standardization involves subtracting the mean from each stress value and then dividing by the standard deviation.

[0081] In practice, cluster analysis is performed on real-time attitude data to identify the cooperative motion patterns of the micromirror groups. The K-means algorithm is used for cluster analysis, and the number of clusters is set to 3-5 groups depending on the size of the micromirror array. The cooperative motion pattern is described by the cluster centers, with the center vector of each cluster representing a typical motion pattern. Real-time attitude data is preprocessed and normalized, mapping angle values ​​to the [0,1] interval. It can be understood that the cooperative motion pattern reflects the motion correlation between the micromirrors.

[0082] In practical implementation, the energy input spectral components, energy output spectral components, key stress modes, and cooperative motion modes are combined into a multi-dimensional feature vector. The combination process employs a feature concatenation method, and the dimension of the multi-dimensional feature vector is the sum of the dimensions of each feature. The feature vector is then normalized to ensure consistency in the magnitude of each feature; the normalization method uses min-max scaling. The mathematical representation of the multi-dimensional feature vector is as follows:

[0083]

[0084] in: For multidimensional feature vectors, The energy input spectral component vector. This is the energy output spectral component vector. For the key stress mode vector, This is the vector for the cooperative motion mode.

[0085] See Figure 4 This figure illustrates the stress distribution on the surface of a MEMS micromirror driving mechanism. The X and Y axes represent the spatial coordinates of the driving mechanism surface, while the Z axis corresponds to the stress value calculated according to Hooke's Law. The entire surface uses a heatmap color gradient from dark blue to bright yellow, clearly showing the spatial distribution differences and trends of stress intensity. The figure shows a distinct concentrated stress distribution pattern in certain areas, reflecting the complex mechanical response characteristics of the micromirror under different driving phases. This three-dimensional visualization method can intuitively reveal the detailed features and overall patterns of stress distribution, providing an important visual analysis tool for assessing the mechanical stability of micromirror mechanisms and predicting potential failure risks.

[0086] Example 4: In a specific implementation, multidimensional feature vectors are mapped onto a pre-calibrated stability map. This pre-calibrated stability map is constructed using historical experimental data, which covers performance records of MEMS micromirror arrays under various operating conditions. The stability map contains multiple stable regions, each defined by a central feature vector and a covariance matrix. The central feature vector represents the set of eigenvalues ​​of the ideal stable state, and the covariance matrix describes the distribution characteristics of data points in that region. The pre-calibrated stability map is loaded by reading data from non-volatile memory. The map data is stored in matrix form, including region identifiers, an array of central feature vectors, and an array of covariance matrices. The loading process uses a high-speed data interface to ensure real-time access. In some embodiments, the stability map is constructed based on clustering algorithms such as K-means clustering, grouping historical feature vectors into stable regions. The central feature vector of each region is the mean within the group, and the covariance matrix calculates the dispersion of data within the group. It can be understood that the mapping process matches the current multidimensional feature vector with the reference state in the map.

[0087] In implementation, the multidimensional feature vectors are standardized using the Z-score method. Each element of the multidimensional feature vector is subtracted from its historical mean and then divided by its historical standard deviation. The historical mean and standard deviation are derived from the global data of the stability map. Standardization ensures that the multidimensional feature vectors have the same scale as the feature vectors in the stability map, avoiding the impact of dimensional differences on distance calculation. The value range of the standardized multidimensional feature vector elements is adjusted to zero mean and unit variance. The processing is completed in real time on the embedded processor. Optionally, an outlier detection step is added to the standardization process; if a feature value exceeds a preset range, it is truncated or scaled. It can be understood that standardization is a prerequisite for distance calculation.

[0088] In practical implementation, the Mahalanobis distance between the standardized multidimensional feature vector and the central feature vector of each stable region is calculated. The formula for calculating the Mahalanobis distance is:

[0089]

[0090] in: Represents Mahalanobis distance, It is a standardized multidimensional feature vector. It is the central feature vector of a certain stable region. It is a transpose operation. It is the covariance matrix of the stable region. It is the inverse of the covariance matrix. Mahalanobis distance calculation uses matrix operations, first calculating the vector difference. Then, multiply by the inverse covariance matrix, and finally take the square root to obtain the distance value. The calculation process optimizes numerical stability and avoids matrix singularity issues by processing the covariance matrix through regularization methods.

[0091] In practical implementation, the target stable region with the minimum Mahalanobis distance to the multidimensional eigenvector is determined. All calculated Mahalanobis distance values ​​are compared, and a sorting algorithm is used to find the minimum value. The stable region corresponding to the minimum value is identified as the target stable region. The selection of the target stable region is based on the minimum distance principle, ensuring that the current state is closest to a certain stable state. Historical control parameter records associated with the target stable region are queried. These historical control parameter records are stored in a database, and each record contains historical control parameters such as the amplitude, frequency, and phase of the driving signal, as well as the corresponding system performance indicators. Refer to Table 1, which shows the stability map structure, including the region identifier, the dimension of the central eigenvector, and the size of the covariance matrix.

[0092] Table 1: Stability Spectrum Region Table

[0093]

[0094] In practical implementation, the current control strategy is generated based on historical control parameter records. All historical records of the target stable region are extracted from the database, with each record containing parameter values ​​and performance indicators. When generating the strategy, a weighted score of the historical performance indicators is calculated. The weighted score is based on the indicator's importance weight and historical usage frequency. The record with the highest score is selected as the preferred record, and its parameters are extracted, fine-tuned, and then applied. In some embodiments, the control strategy includes adjustment instructions for the amplitude, frequency, and phase of the drive signal, which are sent to the drive circuit via a digital interface. Optionally, the effectiveness of the strategy is verified after generation, confirming performance improvement through simulation or real-time testing.

[0095] See Figure 5This graph displays the real-time trends of the tilt angles of multiple micromirrors. The horizontal axis represents time, the vertical axis represents the tilt angle value of the micromirrors, and multiple curves of different colors represent the attitude change trajectories of different micromirrors. These curves exhibit certain periodic and regular variation characteristics, reflecting the dynamic response process of the micromirrors under the action of driving signals. By comparing the variation patterns of different curves, the motion coordination and phase relationship between the micromirror groups can be observed. This time-series analysis graph can intuitively present the motion state and dynamic characteristics of the micromirror system, providing time-dimensional data support for identifying coordinated motion patterns and assessing system stability.

[0096] Example 5: In a specific implementation, all historical control parameter records associated with the target stable region are retrieved from the storage system. The storage system employs a relational database management system, and the database table structure includes a region identifier, historical control parameter fields, and system performance indicator fields. The retrieval operation is performed using a structured query language, with query conditions based on the unique identifier of the target stable region. The returned result set contains multiple historical records. Each historical control parameter record specifically records the amplitude value, frequency value, and phase value of the drive signal, as well as the corresponding system performance indicators such as positioning error value and power consumption value. In some embodiments, the database query uses pre-compiled statements to optimize query efficiency, and the query results are sorted by timestamp to return the latest record. It is understood that the retrieval of historical control parameter records needs to ensure data integrity and consistency.

[0097] In practice, the weighted score of the performance index for each historical control parameter record is calculated. The formula for calculating the weighted score of the performance index is as follows:

[0098]

[0099] in: This represents the weighted score of the j-th historical record. It is the importance weight of the i-th performance metric. It is the normalized value of the i-th performance metric in the j-th record. It is the total number of performance indicators. It is the historical usage frequency of the j-th record. This is an adjustment coefficient for usage frequency. The importance weights of performance indicators are predetermined using the analytic hierarchy process (AHP), and historical usage frequency is obtained from database access logs. The calculation process iterates through all retrieved records, generating a quantitative score for each record. Optionally, the performance indicator normalization process uses the min-max normalization method, mapping indicators of different dimensions to the [0,1] interval. It can be understood that the weighted score calculation comprehensively considers both performance quality and usage history.

[0100] In specific implementation, the historical control parameter record with the highest weighted score is selected as the preferred record. The weighted score values ​​of all records are compared, and the record corresponding to the maximum value is determined using bubble sort or quicksort algorithms. The historical control parameters in the preferred record are extracted, including the specific drive signal amplitude value, drive signal frequency value, and drive signal phase value. The extracted parameters are transmitted to the parameter fine-tuning module, which adjusts them in conjunction with the current operating status of the MEMS micromirror array, such as the ambient temperature value and the device aging coefficient. In some embodiments, parameter fine-tuning is implemented using a fuzzy logic controller, whose input is the operating status parameter and output is the parameter adjustment amount. It can be understood that the selection of the preferred record is based on quantitative evaluation to ensure reliability. The control strategy specifically includes a set of drive signal amplitude adjustment instructions, drive signal frequency adjustment instructions, and drive signal phase adjustment instructions. The correction of the drive signal waveform shape is achieved by changing the lookup table data of the digital waveform synthesizer, the adjustment of the drive frequency spectrum is achieved by modifying the output spectrum using a digital filter, and the coordinated correction of the drive phase difference between micromirrors is accomplished by adjusting the delay parameters of each channel. The control strategy is encapsulated as a control command frame and sent to the drive circuit through the communication interface. Optionally, a safety check can be added during the generation of the control strategy to ensure that the parameter values ​​are within the allowable range. It can be understood that the generation of the control strategy is a core step in achieving closed-loop control.

[0101] In specific implementation, the driving signal parameters applied to the MEMS micromirror array are adjusted using a control strategy. This adjustment involves changing the output voltage amplitude via a digital-to-analog converter, modifying the output frequency via a frequency synthesizer, and adjusting the phase difference between each channel via a phase delay circuit. The adjustment process employs a gradual approach to avoid abrupt changes, with the driving signal parameters progressively transitioning to the target value in specified steps. Immediately after adjustment, data re-acquisition is initiated, re-acquiring feed energy waveform characteristics, load energy waveform characteristics, stress distribution data, and real-time attitude data. In some embodiments, driving signal parameter adjustment and data acquisition are performed synchronously, with the acquisition time window covering the entire adjustment process. The re-acquired data is input into a spatiotemporal characteristic evaluation model to generate a verification multidimensional feature vector. The generation process for the verification multidimensional feature vector is completely consistent with the original multidimensional feature vector, including the same feature extraction and combination steps. The generated verification multidimensional feature vector is compared with the target stable region. The verification process calculates the Mahalanobis distance between the verification multidimensional feature vector and the center feature vector of the target stable region. If the Mahalanobis distance is less than the region radius threshold, the verification multidimensional feature vector is determined to fall within the target stable region; otherwise, an iterative correction process is initiated. Optionally, a maximum number of iterations can be set in the verification process to prevent infinite loops. It's understandable that the verification phase ensures the effectiveness of the control strategy.

[0102] In practice, the control strategy is iteratively revised until the verification multidimensional feature vector falls into the target stable region. This iterative revision is based on gradient descent, adjusting the control parameters according to the deviation direction between the verification result and the target region. After each iteration, the verification multidimensional feature vector is regenerated and re-verified until the convergence condition is met or the maximum number of iterations is reached. The adjustment amount of the control strategy during the revision process is gradually reduced, achieving fine-tuning. The parameters and results of each adjustment are recorded during iterative revision for optimizing subsequent control.

[0103] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method of regulating a MEMS micromirror array, characterized by, The method comprises: establishing a space-time characteristic evaluation model for the MEMS micromirror array, which contains an energy flow analysis domain and a mechanical response domain; in the energy flow analysis domain, synchronously collecting the feed-in energy waveform features and the load energy waveform features of the MEMS micromirror array; in the mechanical response domain, acquiring the stress distribution data of each micromirror driving mechanism and the real-time attitude data of each micromirror reflecting surface of the MEMS micromirror array in parallel; inputting the feed-in energy waveform features, the load energy waveform features, the stress distribution data and the real-time attitude data into the space-time characteristic evaluation model to generate a multi-dimensional feature vector; comparing the multi-dimensional feature vector with a preset stability boundary condition to output a regulation strategy of the MEMS micromirror array.

2. The method of claim 1, wherein, The step of establishing the space-time characteristic evaluation model specifically comprises: defining the data sampling period of the energy flow analysis domain and the data sampling period of the mechanical response domain, and keeping the two sampling periods synchronized; in the energy flow analysis domain, setting an input sensor network for monitoring the feed-in energy waveform features and an output sensor network for monitoring the load energy waveform features; in the mechanical response domain, deploying a micro-strain gauge array for measuring the stress distribution data and an optical distance measuring unit for capturing the real-time attitude data; configuring a data fusion core of the space-time characteristic evaluation model for receiving and time-aligning asynchronous data streams from the input sensor network, the output sensor network, the micro-strain gauge array and the optical distance measuring unit.

3. The method of claim 2, wherein, The step of synchronously collecting the feed-in energy waveform features and the load energy waveform features comprises: continuously recording the voltage transient value and the current transient value in the power supply circuit of the MEMS micromirror array through the input sensor network; continuously recording the voltage transient value and the current transient value in the load circuit of the MEMS micromirror array through the output sensor network; performing convolution operation on the recorded voltage transient value and current transient value of the power supply circuit to extract the feed-in energy waveform features; performing convolution operation on the recorded voltage transient value and current transient value of the load circuit to extract the load energy waveform features.

4. The method of claim 3, wherein, The step of acquiring the stress distribution data and the real-time attitude data in parallel comprises: measuring the surface strain of each micromirror driving mechanism at different driving phases in a spatial interpolation manner through the micro-strain gauge array; converting the surface strain data into the stress distribution data; emitting a measuring light beam to each micromirror reflecting surface through the optical distance measuring unit, and receiving the phase shift of the reflected light beam; calculating the tilt angle and offset of each micromirror reflecting surface relative to the reference position according to the phase shift to form the real-time attitude data.

5. The method of claim 1, wherein, The step of generating a multi-dimensional feature vector comprises: performing frequency domain decomposition on the feed-in energy waveform features to obtain energy input frequency spectrum components; performing frequency domain decomposition on the load energy waveform features to obtain energy output frequency spectrum components; performing principal component analysis on the stress distribution data to extract key stress modes; performing cluster analysis on the real-time attitude data to identify the cooperative motion mode of the micromirror group; combining the energy input spectrum component, the energy output spectrum component, the key stress mode and the cooperative motion mode into the multi-dimensional feature vector.

6. The method of claim 5, wherein, The step of comparing the multi-dimensional feature vector with the preset stability boundary condition comprises: mapping the multi-dimensional feature vector into a pre-labeled stability atlas; calculating Mahalanobis distances between the multi-dimensional feature vector and the center of each stability region in the stability atlas; determining a target stability region with the minimum Mahalanobis distance to the multi-dimensional feature vector; querying a historical control parameter record associated with the target stability region; generating the current control strategy according to the historical control parameter record; The step of mapping the multi-dimensional feature vector into a pre-labeled stability atlas comprises: loading the pre-labeled stability atlas, which is constructed from historical experimental data and contains multiple stability regions, each of which is defined by a center feature vector and a covariance matrix; normalizing the multi-dimensional feature vector so that it has the same scale as the feature vectors in the stability atlas; calculating Mahalanobis distances between the normalized multi-dimensional feature vector and the center feature vector of each stability region, using the covariance matrix of the corresponding stability region as the weight; comparing the values of all Mahalanobis distances to identify the stability region corresponding to the smallest Mahalanobis distance as the target stability region.

7. The method of claim 6, wherein the MEMS micromirror array is a 2D MEMS micromirror array. The step of outputting the control strategy further comprises: executing the control strategy to adjust the driving signal parameters applied to the MEMS micromirror array; after the adjustment is completed, re-collecting a new round of the feed-in energy waveform features, the load energy waveform features, the stress distribution data and the real-time attitude data; inputting the re-collected data into the spatiotemporal characteristic evaluation model to generate a verification multi-dimensional feature vector; verifying whether the verification multi-dimensional feature vector falls within the target stability region; if not, iteratively correcting the control strategy until the verification multi-dimensional feature vector falls within the target stability region.

8. The method of claim 6, wherein the MEMS micromirror array is controlled by a plurality of driving electrodes, and the method further comprises: applying a voltage to the plurality of driving electrodes to control the MEMS micromirror array. The step of generating the current control strategy according to the historical control parameter record comprises: retrieving all historical control parameter records associated with the target stability region from the storage system, each record containing a historical control parameter and its corresponding system performance indicator; calculating a performance indicator weighted score for each historical control parameter record, the weighted score being determined based on the importance weight of the performance indicator and the historical usage frequency; selecting the historical control parameter record with the highest weighted score as the preferred record; extracting the historical control parameter in the preferred record and making parameter fine-tuning in combination with the current running state of the MEMS micromirror array to generate the current control strategy, which includes amplitude, frequency and phase adjustment instructions of the driving signal.

9. The method of claim 1, wherein, The control strategy specifically includes a set of instructions for cooperatively correcting the driving waveform shape, driving frequency spectrum and driving phase difference between micromirrors of the MEMS micromirror array.

10. A system for controlling an array of MEMS micromirrors, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, The processor, when executing the computer program, implements the steps of the method for controlling the MEMS micromirror array according to any one of claims 1 to 9.

Images