Pixel array intelligent control system of MicroLED display

Abstract

The invention belongs to the technical field of semiconductor display, and particularly relates to a pixel array intelligent control system of a MicroLED display. The system comprises a pixel sensing array, a dynamic compensation engine, a self-adaptive driving scheduling module, an environment coupling feedback module and a central cooperative control unit, and realizes single-pixel-level brightness and chrominance correction, power consumption suppression and fault redundancy mapping by collecting pixel-level and environment multi-dimensional data in real time, dynamically generating a compensation coefficient and optimizing driving parameters. And the display uniformity, the color fidelity and the system reliability are obviously improved.

Description

Technical Field

[0001] This invention belongs to the field of semiconductor display technology, and specifically relates to an intelligent control system for pixel arrays of MicroLED displays. Background Technology

[0002] MicroLED display technology, as a next-generation self-emissive display solution, is considered a core development direction in the future display field due to its advantages of high brightness, high contrast, low power consumption, and long lifespan. This technology achieves pixel-level independent control by directly integrating a micrometer-scale LED array onto the driver backplane, resulting in significantly superior image quality and energy efficiency compared to traditional LCD and OLED technologies. Precise driving and real-time control of the pixel array are crucial for ensuring the stable performance and image quality of MicroLED displays.

[0003] The pixel array intelligent control system aims to compensate for display unevenness caused by manufacturing process deviations, temperature drift, and aging effects by dynamically sensing and closed-loop adjusting the current, brightness, and operating status of each MicroLED pixel unit. This system typically integrates sensor feedback, driving circuits, and intelligent algorithms to achieve efficient and precise management of large-scale pixel arrays.

[0004] Existing technologies for controlling MicroLED pixel arrays still face multiple bottlenecks: limited integration of driving circuits leads to high wiring complexity, making it difficult to support independent pixel addressing at ultra-high resolutions; the lack of a real-time monitoring mechanism for pixel-level electro-optical characteristics makes it impossible to dynamically correct brightness decay and color shift; traditional control strategies rely on fixed parameter configurations, making it difficult to adapt to changes in ambient temperature and nonlinear degradation during long-term operation; simultaneously, existing systems suffer from low computational efficiency and significant response latency when processing large-scale pixel data, severely impacting the display effects of high refresh rates and high dynamic ranges. Therefore, there is an urgent need for an intelligent control system for MicroLED display pixel arrays that achieves high integration, real-time sensing, adaptive adjustment, and efficient computation. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent control system for the pixel array of a MicroLED display, so as to solve the systemic technical bottlenecks in existing MicroLED display technology caused by pixel miniaturization, increased complexity of driving circuits, and ambient light interference, such as uneven brightness, color drift, excessive power consumption, and response delay.

[0006] The technical solution of this invention is an intelligent control system for a pixel array of a MicroLED display, comprising a pixel sensing array module, a dynamic compensation engine module, an adaptive drive scheduling module, an environmental coupling feedback module, and a central collaborative control unit. The pixel sensing array module is embedded inside the MicroLED pixel substrate and is used to collect real-time data on the luminous intensity, color coordinates, operating temperature, and current fluctuations of each pixel unit. The dynamic compensation engine module is connected to the pixel sensing array module and is used to construct a pixel performance degradation model based on the collected pixel-level real-time data, and generate independent brightness and color compensation coefficients for each pixel. The adaptive drive scheduling module is connected to... The dynamic compensation engine module is connected to the driving circuit of the MicroLED pixel array, which is used to dynamically adjust the duty cycle of the pulse width modulation signal and the driving current amplitude of each pixel according to the compensation coefficient. The environmental coupling feedback module is deployed on the surface of the display shell and key internal hot nodes, which is used to synchronously collect ambient light intensity, ambient color temperature, ambient temperature and humidity data, and transmit the environmental data to the central collaborative control unit. The central collaborative control unit is communicatively connected to the dynamic compensation engine module, the adaptive drive scheduling module and the environmental coupling feedback module, respectively, and is used to fuse pixel-level state data and multi-dimensional environmental parameters, execute global optimization decisions and output unified driving strategy instructions.

[0007] The pixel sensing array module consists of a miniature photoelectric sensor, a temperature sensor, and a current sampling circuit integrated near the cathode or anode of each MicroLED pixel. The miniature photoelectric sensor is an InGaN-based detector integrated with the same process as the MicroLED epitaxial layer, with a spectral response range covering 400 nm to 700 nm. The temperature sensor is a thermistor based on the temperature coefficient of resistance of polycrystalline silicon, with an accuracy of ±0.5 degrees Celsius. The current sampling circuit is implemented by a low-resistance sampling resistor connected in series in the pixel driving path in conjunction with a differential amplifier, with a sampling frequency of not less than 10 kHz.

[0008] The dynamic compensation engine module includes a pixel performance modeling submodule and a compensation coefficient generation submodule. The pixel performance modeling submodule uses historical data and current real-time data to fit the brightness decay trend of each pixel using an exponential decay function, and constructs a three-dimensional state space model by combining the color coordinate offset vector. The compensation coefficient generation submodule calculates the inverse mapping parameters that make the current pixel output approximate the standard chromaticity coordinates and the target brightness according to the three-dimensional state space model, and outputs them as independent red, green, and blue three-channel gain coefficients.

[0009] The adaptive drive scheduling module includes a pulse width modulation controller and a current amplitude regulator. The pulse width modulation controller receives the gain coefficient from the dynamic compensation engine module and converts it into the PWM duty cycle adjustment amount of the corresponding channel, with an adjustment range of 0% to 100% and a resolution of not less than 12 bits. The current amplitude regulator dynamically limits the maximum drive current based on pixel operating temperature and ambient temperature data to prevent thermal runaway. Its current adjustment step is 0.1 mA, and the adjustment range is 1 mA to 20 mA.

[0010] The environmental coupling feedback module includes an ambient light sensor array, a color temperature sensor, and a temperature and humidity sensor group. The ambient light sensor array consists of at least four silicon photodiodes distributed at the four corners of the display, used to detect the spatial distribution of incident light intensity. The color temperature sensor is based on a dual-channel filter and a photoelectric conversion circuit, outputting an ambient color temperature value with a measurement range of 2000K to 10000K. The temperature and humidity sensor group is deployed on the back heat sink fins, the top of the driver IC package, and the back of the pixel array, with a sampling period of 1 second.

[0011] The central collaborative control unit runs a multi-objective optimization algorithm, which aims to minimize global power consumption, maximize color consistency, and minimize visual latency. The constraints include maximum allowable temperature rise, pixel current safety threshold, and ambient light adaptability requirements. The central collaborative control unit performs a global policy update every 50 milliseconds and synchronizes the updated driving parameters to the adaptive driving scheduling module.

[0012] The system also includes a fault pixel isolation and redundancy mapping module. The fault pixel isolation and redundancy mapping module monitors the abnormal signals output by the pixel sensing array module. When the brightness of a pixel is lower than the threshold or the current fluctuates abnormally by more than 20% for three consecutive sampling periods, it is determined to be a fault pixel. The system automatically maps the display task of the pixel to its four adjacent normal pixels and performs brightness and chromaticity interpolation compensation through a sub-pixel rendering algorithm to ensure image continuity.

[0013] The system supports multi-region independent dimming; the central collaborative control unit divides the entire display area into no less than 64 logical dimming zones, each zone independently performs ambient light adaptation and dynamic compensation, and the zone boundaries are dynamically adjusted according to the edge features of the image content, with an adjustment delay of no more than 2 frames.

[0014] The pixel sensing array module and the dynamic compensation engine module use differential signal transmission with a communication rate of 1 gigabit per second and a bit error rate of less than 10 to the power of -9. The central collaborative control unit adopts a dual-core ARM Cortex-M7 architecture with a main frequency of 480 MHz and has a built-in hardware accelerator for performing color space conversion and Bayesian state estimation.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: This invention integrates a high-precision sensing unit within the MicroLED pixel substrate, enabling real-time closed-loop monitoring of the working status of each pixel. This fundamentally solves the display unevenness problem caused by device aging, process deviations, and temperature drift in traditional open-loop driving modes. The dynamic compensation engine module generates independent compensation coefficients based on a pixel-level performance degradation model, achieving single-pixel level accuracy in brightness and color correction, significantly improving display uniformity and color fidelity. The adaptive drive scheduling module, combined with environmental coupling feedback data, dynamically adjusts drive parameters, automatically enhancing brightness and contrast in strong light environments and reducing power consumption and suppressing light pollution in dark environments, achieving synergistic optimization of display performance and energy efficiency. The central collaborative control unit, through a multi-objective global optimization algorithm, effectively suppresses system power consumption and heat accumulation while ensuring image quality, extending the lifespan of the MicroLED display. The fault pixel isolation and redundant mapping mechanism ensures high-integrity image output even when some pixels fail, significantly improving system reliability and robustness. The multi-region independent dimming function, combined with a content-aware zoning strategy, further enhances HDR display effects and visual comfort. In summary, this invention constructs an integrated intelligent control system that combines perception, decision-making, and execution, providing key technical support for the large-scale application of MicroLED displays in the high-end display field. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall technical solution architecture of the present invention; Figure 2 This is a schematic diagram of the core principle framework of the dynamic compensation engine module in this invention. Detailed Implementation

[0017] Example 1 This invention provides an intelligent control system for the pixel array of a MicroLED display. This system addresses key technical challenges faced by current MicroLED display technology, including uneven brightness, color drift, excessive power consumption, and response delay, due to pixel miniaturization, high complexity of driving circuits, and ambient light interference. By constructing an intelligent closed-loop architecture integrating perception, decision-making, and execution, this control system achieves refined and adaptive optimization of MicroLED display performance, thereby significantly improving display quality, reducing energy consumption, and extending lifespan.

[0018] This system architecture comprises a pixel sensing array module, a dynamic compensation engine module, an adaptive drive scheduling module, an environmental coupling feedback module, and a central collaborative control unit. The overall system operation logic begins with the pixel sensing array module acquiring the real-time status of each MicroLED pixel. The acquired high-precision data is then transmitted to the dynamic compensation engine module for performance degradation modeling and compensation coefficient generation. Simultaneously, the environmental coupling feedback module acquires multi-dimensional environmental parameters in parallel. All this data converges at the central collaborative control unit, where global optimization decisions are made and unified drive strategy instructions are generated. Finally, the adaptive drive scheduling module adjusts the drive parameters of the MicroLED pixel array in real time based on these instructions, thus forming a continuously optimized, highly adaptive intelligent control closed loop.

[0019] The pixel sensing array module is embedded inside the MicroLED pixel substrate. Its core function is to acquire data on the luminous intensity, color coordinates, operating temperature, and driving current fluctuations of each individual pixel unit in real time and with high precision. The module is deployed close to the MicroLED pixels to maximize the accuracy and real-time performance of data acquisition.

[0020] Specifically, the pixel sensing array module consists of a miniature photoelectric sensor, a temperature sensor, and a current sampling circuit integrated near the cathode or anode of each MicroLED pixel. The miniature photoelectric sensor uses an indium gallium nitride (IGaN) detector integrated with the same process as the MicroLED epitaxial layer, offering a wide spectral response range covering 400 nm to 700 nm, comprehensively capturing the luminescence characteristics of red, green, and blue primary colors. Each miniature photoelectric sensor unit contains an integrated filter array corresponding to the red, green, and blue spectral responses, converting the incident light signal into a current signal via a photodiode. These current signals are converted into voltage signals by a transimpedance amplifier and then digitized by a high-resolution analog-to-digital converter. The digitized raw photoelectric data undergoes linearization calibration and noise filtering by an internal signal processing unit, and is finally transmitted to the local data aggregation node within the module via a high-speed serial interface (e.g., SPI or I2C) in the form of a triplet of pixel ID, timestamp, and raw light intensity value. Color coordinates are calculated at each pixel data aggregation point by mapping the three-channel light intensity values ​​to the CIE 1931 chromaticity diagram to obtain real-time color coordinate values.

[0021] The temperature sensor is a thermistor based on the temperature coefficient of resistance of polycrystalline silicon, with an accuracy of ±0.5 degrees Celsius. Each temperature sensor is closely attached to the substrate or heat dissipation structure of the MicroLED pixel, reflecting the pixel's operating temperature by measuring the change in the resistance value of the polycrystalline silicon resistor. The resistance change signal is converted into a differential voltage signal by a precision Wheatstone bridge, then amplified by a differential amplifier with a high common-mode rejection ratio, and finally digitized by a 16-bit analog-to-digital converter. The digitized temperature data is then calibrated to eliminate nonlinear errors, ensuring high accuracy in temperature measurement. The temperature sensor data not only reflects the thermal state of the pixel itself but also serves as an important basis for subsequent thermal management and drive current limiting.

[0022] The current sampling circuit is implemented using a low-resistance sampling resistor connected in series in the pixel drive path, in conjunction with a differential amplifier, with a sampling frequency of at least 10 kHz. A miniature, high-precision sampling resistor, typically in the milliohm range, is connected in series in the drive path of each pixel to minimize its impact on the drive voltage. The current flowing through the MicroLED pixel generates a small voltage drop across the sampling resistor, which is amplified by an ultra-low noise, high-bandwidth instrumentation amplifier. The amplified voltage signal is sampled by a high-speed, high-resolution analog-to-digital converter to capture instantaneous fluctuations in the drive current in real time. The current sampling circuit also integrates an overcurrent protection mechanism; once an abnormal current spike is detected, an alarm is immediately sent to the central coordination control unit, and an emergency shutdown can be performed in extreme cases. All data collected from the miniature photoelectric sensor, temperature sensor, and current sampling circuit are initially synchronized and timestamped within the pixel sensing array module, encapsulated into standard data frames, and transmitted at high speed to the dynamic compensation engine module in differential signal form. The data transmission rate is set to 1 gigabits per second. Low-voltage differential signal transmission technology is used to ensure signal integrity and anti-interference capability at high speeds, and the bit error rate is strictly controlled to below 10 to the power of -9.

[0023] The dynamic compensation engine module connects to the pixel sensing array module. Its core responsibility is to build and continuously update a pixel performance degradation model based on the received pixel-level real-time data, thereby generating independent and accurate brightness and color compensation coefficients for each pixel. This module aims to offset the display performance degradation caused by MicroLED device aging, manufacturing process deviations, and operating temperature variations.

[0024] The dynamic compensation engine module consists of a pixel performance modeling submodule and a compensation coefficient generation submodule. The pixel performance modeling submodule, based on historical data and current real-time data, uses an exponential decay function to fit the brightness decay trend of each pixel and constructs a three-dimensional state space model by combining it with the color coordinate offset vector. The brightness decay trend fitting process utilizes the long-term historical brightness data of each pixel. Specifically, this submodule maintains the historical brightness record of each pixel and applies a weighted least squares method combined with the exponential decay function model: in, Indicates time brightness at that time Initial brightness, The attenuation coefficient is... It is the base of the natural logarithm, approximately equal to 2.718. This model accurately predicts the brightness decay path of pixels. The value is obtained by fitting historical data of the pixel's actual running time and brightness change rate, and is usually updated periodically to adapt to the actual operating environment. When new brightness data is received, the submodule uses Kalman filtering or extended Kalman filtering algorithms to adjust the pixel's brightness attenuation coefficient. Perform online estimation and correction to reflect the latest attenuation status of pixels in real time.

[0025] Meanwhile, the construction of the color coordinate offset vector is achieved by tracking the difference vector between the real-time color coordinates of each pixel on the CIE 1931 or CIE 1976 UCS chromaticity map and its standard target color coordinates. These offset vectors constitute the chromaticity dimension of the three-dimensional state space model, which, together with the luminance decay model, describes the complete performance state of the pixel. The three-dimensional state space model not only includes luminance decay and chromaticity drift, but may also integrate additional dimensions such as pixel response time and contrast performance, forming a multi-dimensional set of performance indicators. This submodule internally maintains a large pixel-level historical performance database for model training, validation, and online updates, ensuring that the model accurately characterizes the device's features. Each model update performs outlier detection and cleaning on the newly acquired pixel data to prevent noisy data from interfering with the model's accuracy.

[0026] The compensation coefficient generation submodule calculates the inverse mapping parameters that approximate the standard chromaticity coordinates and target brightness of the current pixel output based on the 3D state space model output by the pixel performance modeling submodule, and outputs them as independent red, green, and blue channel gain coefficients. This inverse mapping process aims to find the optimal adjustment amount of the driving parameters to offset the modeled performance degradation. The calculation process first determines the current actual brightness and chromaticity output of each pixel and compares it with preset standard brightness and chromaticity targets. Then, based on the pixel performance model, the submodule uses a nonlinear optimization algorithm (such as gradient descent or Newton's method) to iteratively solve for the required red, green, and blue channel driving intensities so that the actual output of the pixel is as close as possible to the target value. These solved driving intensity values ​​are converted into corresponding gain coefficients. For example, if the brightness of a red pixel decreases by 10%, its red channel gain coefficient may be calculated as 1.1 to compensate for this 10% brightness loss. The calculation of the gain coefficients also considers the cross-coupling effect, that is, the adjustment of one channel may affect the output of other channels, so a multivariate control strategy is adopted. The generated gain coefficients are typically expressed as floating-point numbers and then converted to fixed-point form to suit subsequent digital circuit processing. Their values ​​usually range from 0.8 to 1.5, depending on the degree of pixel degradation. The generation frequency of the compensation coefficients is matched to the data acquisition frequency of the pixel sensing array module to ensure real-time compensation.

[0027] The adaptive drive scheduling module connects the dynamic compensation engine module and the drive circuit of the MicroLED pixel array. Its core function is to dynamically adjust the duty cycle of the pulse width modulation signal and the amplitude of the drive current for each MicroLED pixel based on the compensation coefficients generated by the dynamic compensation engine module and the environmental data provided by the environmental coupling feedback module. This module is a key link connecting intelligent decision-making and physical execution.

[0028] The adaptive drive scheduling module includes a pulse width modulation (PWM) controller and a current amplitude regulator. The PWM controller receives the gain coefficient from the dynamic compensation engine module and converts it into the corresponding PWM duty cycle adjustment for the channel. The adjustment range is 0% to 100%, with a resolution of at least 12 bits. Each pixel's PWM controller contains a high-speed counter and a comparator. Based on the received gain coefficient and a preset brightness response curve, the controller calculates the required PWM duty cycle. For example, if a pixel's red channel requires a gain coefficient of 1.1, the PWM controller consults a non-linear mapping table to convert the gain coefficient into a specific duty cycle value, ensuring the final output brightness matches the desired brightness. The 12-bit resolution means the PWM signal is subdivided into 4096 levels, providing the hardware foundation for extremely fine brightness control and effectively eliminating color quantization errors and flickering at low brightness. The PWM controller also features a jitter cancellation mechanism, which effectively reduces visual artifacts and stripe effects by randomizing the start time of the PWM cycle. The frequency of the PWM signal is typically set at several megahertz to ensure that flickering is imperceptible to the human eye.

[0029] The current amplitude regulator dynamically limits the maximum drive current based on pixel operating temperature and ambient temperature data to prevent thermal runaway. Its current adjustment step is 0.1 mA, with an adjustment range of 1 mA to 20 mA. This regulator is a key component for thermal management and extending device lifespan. It receives real-time pixel operating temperature data from the pixel sensing array module and ambient temperature data from the environmental coupling feedback module. Based on this temperature data, the regulator internally runs a thermal model that predicts the pixel's temperature rise at the current drive current. When the predicted temperature rise exceeds a preset safety threshold, the current amplitude regulator reduces the maximum allowable drive current to prevent pixel overheating. For example, if a pixel operates in a high-temperature environment, its maximum drive current might be limited to 15 mA instead of the usual 20 mA. Current regulation is achieved by controlling a constant current source circuit through a high-precision digital-to-analog converter. The 0.1 mA adjustment step size provides sufficient precision to match the current characteristic curve of MicroLED devices, while the 1 mA to 20 mA adjustment range covers the normal operating current range of MicroLEDs and provides ample redundancy to handle brightness demands in extreme conditions. The current amplitude regulator also integrates a fast-response hardware current-limiting circuit, which can cut off or significantly reduce the drive current within milliseconds to handle sudden overcurrent or short-circuit events and ensure system safety.

[0030] The environmental coupling feedback module is deployed on the surface of the display casing and at key internal thermal nodes. Its main function is to synchronously collect data on ambient light intensity, ambient color temperature, ambient temperature, and humidity, and transmit this multi-dimensional environmental data to the central collaborative control unit. This module ensures that the system has a comprehensive understanding of the external environment in which the display is located.

[0031] The environmental coupling feedback module includes an ambient light sensor array, a color temperature sensor, and a temperature and humidity sensor group. The ambient light sensor array consists of at least four silicon photodiodes distributed at the four corners of the display to detect the spatial distribution of incident light intensity. Each silicon photodiode is equipped with a dedicated optical diffusion layer to ensure that the measurement results are direction-independent, thus accurately reflecting the average intensity and spatial gradient of ambient light. The output current of each photodiode is converted to voltage by a transimpedance amplifier and digitized by a high-resolution analog-to-digital converter. After calibration and linearization, the data accurately quantifies the ambient light intensity, for example, in lux units. Through the array layout of at least four sensors, the system constructs a spatial distribution map of ambient light, thereby enabling finer adjustment of display brightness, such as compensation for local shadow areas on the screen.

[0032] The color temperature sensor, based on a dual-channel filter and photoelectric conversion circuit, outputs the ambient color temperature value, with a measurement range of 2000K to 10000K. The sensor measures the intensity of red and blue light in the incident light using a red channel filter and a blue channel filter, respectively, and then calculates the ambient light color temperature by comparing the relative intensities of these two channels. For example, a high color temperature ambient light (such as a clear sky) will result in a significantly higher blue channel reading than the red channel reading, while a low color temperature ambient light (such as candlelight) will show the opposite trend. The measurement range of 2000K to 10000K covers a wide color temperature range from warm-toned incandescent light to cool-toned daylight or fluorescent light, ensuring accurate white point adjustment of the display under various ambient lighting conditions. The color temperature sensor data is transmitted to the module's internal data processor via I2C or SPI interface for further processing and calibration.

[0033] A temperature and humidity sensor array is deployed on the backplane heatsink fins, the top of the driver IC package, and the back of the pixel array, with a sampling period of 1 second. Sensors on the backplane heatsink fins monitor the overall heat dissipation efficiency of the display and the impact of the external environment on heat dissipation. Sensors on the top of the driver IC package monitor the operating temperature of critical driver chips in real time to prevent overheating. Sensors on the back of the pixel array acquire thermal environment data near the pixel substrate, serving as auxiliary input for pixel-level thermal management and current limiting. These sensors typically employ integrated MEMS sensors with digital output, providing both temperature and humidity data. The 1-second sampling period is sufficient to capture slow changes in ambient temperature and humidity, providing a stable and reliable input for the system's thermal management strategy. All environmental data is timestamped and encapsulated within the environmental coupling feedback module and transmitted to the central collaborative control unit via a dedicated communication bus. An error checking mechanism is employed during transmission to ensure data integrity and accuracy.

[0034] The central collaborative control unit communicates with the dynamic compensation engine module, the adaptive drive scheduling module, and the environmental coupling feedback module. Its core function is to integrate pixel-level state data with multi-dimensional environmental parameters, execute global optimization decisions, and output unified drive strategy instructions. This unit is the brain of the entire intelligent control system, responsible for coordinating the work of all sub-modules.

[0035] The central collaborative control unit runs a multi-objective optimization algorithm. This algorithm aims to minimize global power consumption, maximize color consistency, and minimize visual latency, while being constrained by maximum allowable temperature rise, pixel current safety thresholds, and ambient light adaptability requirements. This multi-objective optimization algorithm may employ advanced optimization techniques based on genetic algorithms, particle swarm optimization, or reinforcement learning. For example, in pursuing minimum power consumption, the algorithm attempts to find the lowest average drive current and PWM duty cycle required while satisfying brightness, color, and temperature constraints. To maximize color consistency, the algorithm adjusts the compensation coefficients of each pixel to ensure highly uniform color performance across the entire display area. Minimizing visual latency requires the algorithm to find a balance between processing speed and optimization accuracy, ensuring rapid response to user input or content changes.

[0036] This optimization process requires a complex cost function that integrates multiple optimization objectives, including power consumption, color deviation, and response time, and balances them through weighting coefficients. Constraints are strictly integrated into the optimization algorithm; for example, any generated driving strategy instruction cannot cause any pixel's operating temperature to exceed its maximum allowable temperature rise, nor can it cause the driving current to exceed the pixel current safety threshold. Ambient light adaptability requirements mean that the system needs to dynamically adjust the white point, brightness curve, and contrast based on ambient light intensity and color temperature to provide the best visual experience. The central collaborative control unit performs a global strategy update every 50 milliseconds and synchronizes the updated driving parameters to the adaptive drive scheduling module. This 50-millisecond update cycle ensures the system's rapid response to dynamic changes while avoiding system instability that might result from excessively frequent strategy updates. To achieve such a high update frequency, the central collaborative control unit adopts a dual-core ARM Cortex-M7 architecture with a main frequency of up to 480 MHz and incorporates hardware accelerators for color space conversion and Bayesian state estimation. The dual-core architecture allows for parallel processing of different tasks; for example, one core is responsible for data fusion and optimization algorithm execution, while the other core is responsible for the communication protocol stack and system management tasks. Hardware accelerators are specifically designed to handle computationally intensive tasks, such as large-scale matrix multiplication (for color space conversion) and probabilistic inference (for Bayesian state estimation, which may be used in fault pixel isolation modules), thereby significantly reducing policy generation time.

[0037] This system also includes a fault pixel isolation and redundancy mapping module. This module monitors abnormal signals output by the pixel sensing array module. When a pixel's brightness is below a threshold or its current fluctuates abnormally by more than 20% for three consecutive sampling periods, it is identified as a faulty pixel. The fault pixel determination logic is based on rigorous statistical analysis and threshold settings. For example, "brightness below the threshold" might be defined as the actual brightness output being less than 30% of its design brightness under a given drive current. "Abnormal current fluctuation exceeding 20%" refers to an instantaneous deviation of the current value exceeding the average value by more than 20% under stable drive conditions, which may indicate a short circuit or open circuit inside the pixel. Anomaly detection for three consecutive sampling periods is to avoid misjudgments caused by transient noise or sporadic events, ensuring the robustness of fault determination.

[0038] Once a pixel is identified as faulty, the system automatically maps its display task to its four neighboring normal pixels. A sub-pixel rendering algorithm is used for brightness and chromaticity interpolation compensation to ensure image continuity. The redundancy mapping mechanism works as follows: When a pixel is marked as faulty, its display task data is no longer directly sent to its own driver but is redirected to the faulty pixel isolation and redundancy mapping module. This module first identifies the precise location of the faulty pixel on the display screen and its four nearest neighbor pixels. Then, it uses an advanced sub-pixel rendering algorithm, such as Bayer interpolation or a Lagrange-based interpolation algorithm, to perform brightness and chromaticity interpolation compensation on the faulty pixel area based on the brightness and chromaticity information of these four neighboring pixels. The interpolation algorithm performs weighted calculations based on the average brightness, color, and spatial distribution of the normal pixels surrounding the faulty pixel, generating a virtual brightness and chromaticity value. This virtual value is then assigned to the four neighboring pixels, allowing them to collectively fill the gap left by the faulty pixel without affecting their own display content. In this way, the display effect of the faulty pixel area is made as smooth and continuous as possible from a visual perspective, avoiding obvious black spots or color blocks, thereby greatly improving the reliability and robustness of the system.

[0039] This system supports multi-zone independent dimming. The central coordinating control unit divides the entire display area into no fewer than 64 logical dimming zones, each independently performing ambient light adaptation and dynamic compensation. Zone boundaries are dynamically adjusted based on the edge features of the image content, with an adjustment delay of no more than 2 frames. This feature aims to provide better local contrast and energy efficiency. 64 logical dimming zones mean the display is finely divided into an 8x8 or larger grid. Each zone has its own independent brightness, contrast, and color temperature adjustment parameters. Data within a zone comes from the average perceived data of all pixels within that zone, as well as ambient light sensor data covering that zone. The dynamic adjustment of zone boundaries is an innovative feature. For example, when a bright object appears at the boundary between a bright object and a dark background in an image, the zone manager adjusts the boundaries of the dimming zones based on the edge features identified by the image processing unit, aligning them as closely as possible with the edges of the image content to avoid blooming and loss of detail in dark areas. This dynamic adjustment strategy significantly improves HDR (High Dynamic Range) display performance. Adjusting the latency to no more than 2 frames means that the time interval from changes in image content to the completion of partition boundary adjustments is extremely short, ensuring a smooth visual experience. The central collaborative control unit contains a dedicated partition manager submodule responsible for dynamically maintaining the partition map and distributing the optimization parameters for each partition to the corresponding adaptive driver scheduling module instance.

[0040] Through the collaborative work of the aforementioned units, this system constructs a highly integrated and intelligent closed-loop control system. This system no longer relies on the traditional open-loop driving mode, but instead achieves real-time closed-loop monitoring of the working status of each MicroLED pixel. It fundamentally solves the display unevenness problem caused by device aging, process deviations, and temperature drift in traditional display technologies, achieving ultra-fine control over each pixel. The dynamic compensation engine module generates independent compensation coefficients based on a pixel-level performance degradation model, enabling brightness and color correction accuracy to reach the single-pixel level, significantly improving display uniformity and color fidelity. The adaptive drive scheduling module, combined with environmental coupling feedback data, dynamically adjusts drive parameters, automatically improving brightness and contrast in strong light environments and reducing power consumption and suppressing light pollution in dark environments, achieving synergistic optimization of display performance and energy efficiency. The central collaborative control unit, through a multi-objective global optimization algorithm, effectively suppresses system power consumption and heat accumulation while ensuring image quality, extending the lifespan of the MicroLED display. Fault pixel isolation and redundant mapping mechanisms ensure high-integrity image output even when some pixels fail. Multi-region independent dimming function combined with a content-aware zoning strategy further enhances HDR display effects and visual comfort. This invention constructs an intelligent control system that integrates perception, decision-making, and execution, providing key technical support for the large-scale application of MicroLED displays in the high-end display field.

[0041] The dynamic compensation engine module ensures the accuracy and adaptability of the compensation coefficients through continuous online learning and model updates. In the initial stage of system deployment, the pixel performance modeling submodule uses initial calibration data obtained during the MicroLED display manufacturing phase using high-precision optoelectronic testing equipment as a benchmark. This initial data includes the LI curve (luminance-current characteristic curve), chromaticity coordinates, and initial uniformity data for each pixel. After the display is put into use, real-time operating data continuously reported by the pixel sensing array module, such as actual brightness, chromaticity coordinates, and operating current, is used to continuously correct and improve the performance degradation model.

[0042] The pixel performance modeling submodule internally maintains a predictor based on a Gaussian process regression or long short-term memory network machine learning model. This predictor takes the pixel's ID, cumulative operating time, average operating current, and average operating temperature as input to predict its future brightness decay rate and chromaticity drift direction. Each time new pixel-level real-time data is received from the pixel sensing array module, the predictor uses this new data for online learning or fine-tuning of the model parameters. For example, if the actual brightness decay rate of a pixel is faster than the model predicts, its corresponding decay coefficient... This will be immediately increased, allowing the compensation coefficient generation submodule to compensate for it more quickly. The importance of historical data is reflected in model training; the submodule periodically extracts data from a large-scale historical database for offline retraining to capture more common performance degradation patterns, and pushes these updated general model parameters to the online learning module for fine-tuning.

[0043] Upon receiving updated performance model data, the compensation coefficient generation submodule calculates the required gain coefficients for the red, green, and blue channels in real time. These gain coefficient calculations consider not only the brightness attenuation and chromaticity shift of the current pixel but also the interaction effects between adjacent pixels. For example, to avoid halo effects or unclear color edges caused by improper compensation between adjacent pixels, a local averaging filter is considered in the compensation coefficient calculation. Furthermore, to ensure the stability of compensation under extreme brightness or chromaticity deviations, the output of the gain coefficients is subject to strict upper and lower limits. For example, the gain coefficient of a single channel will not exceed 1.5 to prevent over-driving from damaging the pixels. The compensation coefficient generation algorithm is also fine-tuned based on the displayed content. For example, when displaying static images, the system allows for longer calculation cycles to achieve more accurate compensation; while when displaying high-speed dynamic video, computational speed and response time are prioritized, potentially employing a simpler compensation model.

[0044] After receiving the gain coefficient from the dynamic compensation engine module and the global drive strategy instructions from the central coordination control unit, the adaptive drive scheduling module precisely controls the PWM signal and drive current through a multi-stage modulator. The pulse width modulation controller does more than simply map the gain coefficient to the PWM duty cycle. It also considers the nonlinear brightness response characteristics of the MicroLED device. To convert the gain coefficient into a visually uniform brightness adjustment, the controller typically includes a gamma correction lookup table or a nonlinear function. For example, if the gain coefficient requires a 10% brightness increase, the actual PWM duty cycle adjustment may need to be larger or smaller, depending on the gamma characteristics of the MicroLED device at the current brightness level. The 12-bit PWM resolution is achieved through a high-speed digital counter operating at hundreds of megahertz, ensuring precise division of the PWM cycle. The counter is set to a fixed period (e.g., 256 clock cycles) and then compared with a preset threshold based on the 12-bit duty cycle value to generate the corresponding PWM waveform. At the start of each PWM cycle, a hardware trigger synchronously starts, ensuring high temporal synchronization of the PWM signals for all pixels and avoiding the visual "scan line" effect.

[0045] The current amplitude regulator employs a prediction-based thermal management strategy when dynamically limiting the maximum drive current. In addition to real-time monitoring of pixel operating temperature and ambient temperature, the regulator integrates a simplified MicroLED thermal diffusion model. This model predicts the rising trend of the pixel junction temperature under the current drive current and duration. For example, when the drive current instantaneously increases to display a high-brightness image, the thermal diffusion model predicts the junction temperature change within milliseconds and limits the current in advance, rather than waiting for the junction temperature to actually rise. This predictive control effectively prevents transient overheating, further improving device lifespan and stability. The 0.1 mA current adjustment step relies on a high-precision current source IC, which integrates a digital-to-analog converter and a feedback control loop to accurately convert the digitized current setpoint into an analog drive current output. This constant current source features high output impedance and low noise characteristics, ensuring the stability of the drive current and thus avoiding the impact of current ripple on MicroLED luminous uniformity and colorimetry. When the system determines that the maximum drive current needs to be reduced, the regulator smoothly decreases the current rather than drastically, to avoid sudden changes or flickering in the image.

[0046] The data acquisition process of the environmental coupling feedback module has also been meticulously designed to ensure data accuracy and reliability. Each silicon photodiode in the ambient light sensor array, after acquiring the light signal, sequentially sends the analog signal to a high-performance analog-to-digital converter via a multiplexer. Before conversion, the signal from each channel passes through a high-pass filter and a low-pass filter to filter out high-frequency noise and low-frequency drift from the environment. Since the photodiode response is non-linear, the digitized data is corrected using an embedded linearization algorithm. To resist electromagnetic interference in the environment, shielded twisted-pair cables are used for the connection between the sensor and the data processing unit, and differential transmission is used to further enhance anti-interference capabilities. The sensor self-calibration function automatically adjusts the gain and bias by periodically acquiring data under known lighting conditions and comparing it with a preset reference value to compensate for performance drift caused by sensor aging or environmental factors. When calculating color temperature, the color temperature sensor, in addition to the ratio method of the red and blue channels, also incorporates a green channel and uses a tristimulus value matching method to obtain a more accurate color temperature value. During deployment, the temperature and humidity sensor array's sensor nodes are encapsulated with thermosetting adhesive to prevent dust and moisture intrusion, while ensuring good thermal contact with the measured node, thereby improving temperature measurement accuracy and response speed. All sensor data is packaged into data frames and have a CRC checksum added within the environmental coupling feedback module before being transmitted to the central collaborative control unit, ensuring data transmission integrity.

[0047] The complexity of the central collaborative control unit in executing the multi-objective optimization algorithm lies in effectively balancing the three mutually constraining objectives of power consumption, color consistency, and visual latency. This unit constructs a weighted summation objective function, where each objective (such as global power consumption, color uniformity index, and latency penalty) has a configurable weight coefficient. These weight coefficients are not fixed but dynamically adjusted based on the display's application scenario (e.g., color consistency may be more important in movie mode, while power consumption is more important in battery-powered mode). The output of the optimization algorithm is a set of drive strategy instructions, including the PWM duty cycle correction value for each pixel, the maximum current limit, and specific parameters for multi-region dimming partitions. To meet the 50-millisecond update cycle, the optimization algorithm employs incremental or block optimization techniques. For example, within each 50-millisecond cycle, instead of re-optimizing all pixels globally, it identifies the pixels or regions with the most significant changes and prioritizes local optimization of these regions.

[0048] The dual-core ARM Cortex-M7 architecture plays a crucial role in performing these complex optimization tasks. One core is dedicated to real-time data acquisition, preprocessing, and communication tasks, ensuring smooth data flow and low latency. The other core focuses on executing multi-objective optimization algorithms and managing fault pixel isolation and redundant mapping modules, as well as multi-region independent dimming functions. Hardware accelerators further offload the computational burden from the general-purpose processor cores. For example, the color space conversion hardware accelerator performs RGB to XYZ or Lab color space conversions at extremely high speeds, which is essential for color consistency calculations and compensation. The Bayesian state estimation hardware accelerator provides efficient probabilistic inference capabilities, such as those needed in fault pixel prediction or sensor data fusion. This heterogeneous computing architecture significantly improves the overall processing power and energy efficiency of the system.

[0049] After identifying faulty pixels, the faulty pixel isolation and redundancy mapping module does not simply perform linear interpolation in its redundancy mapping algorithm. To achieve better visual effects, it may employ texture synthesis or neural network-based super-resolution reconstruction techniques. For example, when a pixel malfunctions, the algorithm analyzes the texture information of surrounding normal pixels and attempts to synthesize an image patch that matches the surrounding texture to fill the faulty area, rather than simply smoothing it. This method effectively preserves image details and sharpness. In practical applications, the subpixel rendering algorithm also considers the differences in human eye sensitivity to different color pixels. For example, since the human eye is more sensitive to green light, a higher weight or a more refined algorithm may be used when interpolating the green light channel. The redundancy mapping module also maintains a list of faulty pixels and records their historical behavior. When the display is in low brightness mode, the visibility of faulty pixels decreases, and the system may choose an interpolation algorithm with lower computational complexity to save power.

[0050] The multi-zone independent dimming function's dynamic boundary adjustment strategy leverages the image content analysis capabilities of the central collaborative control unit. The central collaborative control unit analyzes input image frames in real time through an integrated image processing unit (or utilizes a hardware accelerator), extracting brightness, contrast, color, and edge information. For example, it runs an edge detection algorithm (such as the Canny or Sobel operator) to identify significant edges in the image and adjusts the logical boundaries of the dimming zones based on this edge information. The latency of zone adjustment is strictly controlled within two frames, meaning that from a change in image content to the dimming zone being adjusted, it takes a maximum of approximately 33 milliseconds (assuming a monitor refresh rate of 60 Hz). To achieve such low latency, the zone boundary adjustment algorithm is highly optimized and executed in parallel on the hardware accelerator. Each dimming zone independently performs ambient light adaptation and dynamic compensation, meaning each zone has its own independent brightness, color temperature, and contrast parameters, which are calculated based on the image content within that zone, the ambient light intensity and color temperature covered by that zone, and the performance degradation of the pixels within that zone. For example, in a high-contrast scene, the dimming zones corresponding to bright areas of the image will be brightened, while the dimming zones corresponding to dark areas will be dimmed, thus achieving an optimal balance between overall power consumption and visual effect. This fine-tuned dimming control significantly improves the display's local contrast, dynamic range, and energy efficiency.

[0051] Example 2 This embodiment further elaborates on the internal structure and working principle of the micro photoelectric sensor in the pixel sensing array module, as well as its integration with the MicroLED epitaxial layer.

[0052] As a core component of the pixel sensing array module, the miniature photoelectric sensor is designed to achieve precise, real-time monitoring of the luminous characteristics of individual MicroLED pixels. Each miniature photoelectric sensor unit is located adjacent to the MicroLED light-emitting diode and is typically integrated on the same substrate as the MicroLED array. It employs an indium gallium nitride (IGaN) detector with a spectral response range covering 400 nm to 700 nm. This in-material integration technology maximizes manufacturing process compatibility and ensures a high degree of consistency between the sensor and the MicroLED, reducing thermomechanical stress problems caused by material differences.

[0053] Each miniature photodetector internally comprises a photodiode array, a spectral filter layer, a transimpedance amplifier, an analog-to-digital converter, and a local digital processing unit. The photodiode array consists of several indium gallium nitride (IGaN) PIN photodiodes designed to have high quantum efficiency and low dark current characteristics, ensuring high sensitivity under low-light conditions. The selection of the detector material ensures that its spectral response curve perfectly matches the visible light wavelength range.

[0054] The spectral filtering layer is crucial for colorimetric sensing. Micrometer-scale dielectric thin-film filters are deposited above each photodiode array. These filters are precisely designed to allow only specific wavelengths of light—red, green, and blue—to pass through, thus separating the composite spectrum of the incident light into its three primary color components. For example, a red filter allows light with wavelengths concentrated in the 620-750 nm range, a green filter allows light with wavelengths concentrated in the 495-570 nm range, and a blue filter allows light with wavelengths concentrated in the 450-495 nm range. This optical filtering technique allows the sensor to independently measure the intensity of the red, green, and blue light emitted by the MicroLED pixels. To optimize the quantum efficiency of the photodiodes, an anti-reflective layer is typically integrated beneath the spectral filtering layer to reduce light reflection loss as it enters the detector.

[0055] The weak photocurrent signal output from the photodiode array is first fed into a high-precision transimpedance amplifier. The transimpedance amplifier converts the photocurrent into a voltage signal with adjustable gain to accommodate dynamic range under varying brightness and ambient lighting conditions. To ensure measurement accuracy, the transimpedance amplifier employs a low-noise, high-bandwidth operational amplifier design and integrates temperature compensation circuitry to minimize the impact of temperature drift on the gain. The amplified voltage signal is then input to a high-resolution analog-to-digital converter (ADC) to convert it into a digital signal. This ADC is typically 10-bit or 12-bit, ensuring precise quantization of light intensity variations.

[0056] The local digital processing unit (LDU) is responsible for preprocessing the raw digital light intensity data after analog-to-digital conversion. Preprocessing includes: noise filtering, such as using moving average filtering or median filtering to eliminate random noise; linearization calibration to correct the sensor's nonlinear response to a linear response; and ambient light background subtraction, which uses a specially designed light-shielding structure and algorithm to distinguish between the light emitted by the MicroLED itself and ambient stray light. In addition, the LDU is also responsible for converting the three-channel light intensity data into CIE chromaticity coordinates. It takes the digital light intensity values ​​of the red, green, and blue channels as input, and calculates the x and y chromaticity coordinates of the current pixel in real time using a preset color matrix and conversion algorithm. This can be further converted to CIE L*a*b* or other perceptually uniform color spaces. All this processed data, including pixel ID, timestamp, red, green, and blue light intensity values, and chromaticity coordinates, is packaged into fixed-format data frames and transmitted through a high-speed serial interface (such as MIPI-DSI or a custom differential serial interface) to the data aggregation bus inside the pixel sensing array module. Finally, the data is aggregated and sent to the dynamic compensation engine module.

[0057] This micro-photoelectric sensor, integrated into the same manufacturing process, achieves high-precision real-time monitoring of the luminous intensity and color of MicroLED pixels through its precise structure and high-performance signal chain. This provides the most fundamental and crucial data support for subsequent intelligent compensation and control. The sensor's low-power design also ensures stable operation within the pixel substrate for extended periods, without significantly impacting the overall energy efficiency and heat dissipation of the MicroLED display.

[0058] Example 3 This embodiment further elaborates on the specific execution strategy of the multi-objective optimization algorithm in the central collaborative control unit and its collaborative working mechanism with the hardware accelerator.

[0059] The central collaborative control unit, serving as the brain of the intelligent control system for the MicroLED display pixel array, has one of its core functions: running a complex multi-objective optimization algorithm. This algorithm aims to simultaneously achieve three main objectives: minimizing global power consumption, maximizing color consistency, and minimizing visual latency, while strictly adhering to a series of constraints, including maximum allowable temperature rise, pixel current safety threshold, and ambient light adaptability requirements.

[0060] The multi-objective optimization algorithm employs a hierarchical, progressive hybrid optimization approach to achieve a balance between computational efficiency and optimization effectiveness. This strategy includes: global coarse-grained optimization, local fine-grained optimization, and model-prediction-based optimization.

[0061] Global coarse-grained optimization: Whenever the displayed content or key environmental parameters (such as ambient light intensity or color temperature) change significantly, the central coordinating control unit initiates a global optimization phase. In this phase, the system treats the entire MicroLED display area as a whole, or divides it into a few large regions (e.g., 16 or 32 partitions). The optimization algorithm rapidly converges to a near-optimal global driving strategy through a reinforcement learning-based policy network. The policy network's inputs include: environmental parameters collected by all environmental coupling feedback modules (global light intensity, color temperature, ambient temperature, humidity), the average state of the pixel performance model provided by the dynamic compensation engine module, and current display content summary features (e.g., average image brightness, main color distribution, contrast). The policy network's outputs are the average brightness target, contrast enhancement coefficient, and overall color temperature adjustment value for each large partition or the entire screen. The reinforcement learning method learns the optimal driving strategy under different environmental and content inputs through extensive interaction with the display's virtual model and is trained using policy gradient descent. The trained policy network rapidly generates an initial global driving strategy within approximately milliseconds, ensuring the system's rapid response to environmental changes.

[0062] Local Fine-Grained Optimization: Following global coarse-grained optimization, the system immediately initiates local fine-grained optimization. In this stage, the optimization algorithm focuses on each logical dimming zone (no fewer than 64 zones) and abnormal pixels reported by the pixel sensing array module. For each dimming zone, the algorithm invokes an optimizer based on multivariate nonlinear programming. This optimizer targets the average power consumption, color deviation, and local response delay within the zone. Constraints include the maximum temperature rise of pixels within the zone, a current safety threshold, and requirements for smooth transitions in brightness and chromaticity with adjacent zones. Detailed pixel-level performance degradation models provided by the pixel performance modeling submodule and gain coefficients generated by the compensation coefficient generation submodule serve as inputs to the optimization. The optimizer iteratively searches to generate precise PWM duty cycles and maximum drive current values ​​for each pixel or pixel group within the zone. This process utilizes the built-in hardware accelerator for matrix operations and numerical solutions, significantly reducing computation time. For example, when handling color consistency targets, the hardware accelerator performs high-speed color space conversion and color deviation (e.g., Delta E) calculations.

[0063] Model-Predictive Optimization: To minimize visual latency and improve the system's predictive capabilities, the central co-control unit also integrates a Model Predictive Control (MPC) optimization component. This component utilizes degradation models provided by the pixel performance modeling submodule and environmental trend predictions (e.g., predictions of ambient temperature and light intensity based on historical data) provided by the environment coupling feedback module to predict pixel performance changes and potential changes in environmental conditions within the next tens of milliseconds. Based on these predictions, the MPC controller generates a forward-looking driving strategy sequence at the current moment, thereby compensating for potential performance degradation or environmental changes in advance, further reducing visual latency and improving system robustness. For example, if it is predicted that the junction temperature of MicroLED pixels in a certain area will rise significantly within the next 100 milliseconds, the MPC controller will reduce the upper limit of the driving current in that area in advance to avoid actual overheating, while adjusting the driving parameters of surrounding pixels to maintain overall brightness consistency.

[0064] The constraints of the multi-objective optimization algorithm have also been refined: Maximum permissible temperature rise: This is achieved by introducing a penalty term into the cost function. When the predicted pixel junction temperature approaches or exceeds the maximum permissible threshold, the penalty term increases sharply, forcing the optimization algorithm to reduce the drive current.

[0065] Pixel current safety threshold: enforced through hard constraints, any optimized driving current cannot exceed the maximum rated current of the MicroLED pixel.

[0066] Ambient light adaptability requirements: These are met by adjusting the monitor's white point, gamma curve, and contrast in real time. For example, in strong ambient light, the system increases the monitor's overall brightness and contrast; in dark environments, it reduces brightness and adjusts the white point to minimize eye strain.

[0067] The hardware accelerators of the central coordinating control unit play a crucial role throughout the optimization process. The color space conversion hardware accelerator pipelines massive amounts of pixel color data, performing rapid conversions from RGB to XYZ and XYZ to Lab color spaces. This is essential for calculating color consistency, color drift, and white point adjustment, as these calculations must be performed within a perceptibly uniform color space. The Bayesian state estimation hardware accelerator provides efficient probabilistic inference capabilities, such as in sensor data fusion, probabilistic determination of faulty pixels, and uncertainty quantification in model predictive control. These accelerators perform complex mathematical operations through dedicated hardware logic gate arrays, making them tens or even hundreds of times faster than software implementations on general-purpose processor cores, and are key to achieving a 50-millisecond global policy update cycle.

[0068] The dual-core ARM Cortex-M7 architecture achieves parallel processing through task allocation. One core may focus on I / O operations, communication protocol stack management, and real-time operating system scheduling, ensuring smooth data transmission from various modules (pixel sensing array, environmental coupling feedback) to the central collaborative control unit, and sending instructions to the adaptive drive scheduling module. The other core is fully dedicated to the execution of multi-objective optimization algorithms, model prediction, and the logical processing of fault pixel isolation and redundancy mapping modules. This division of labor greatly improves the overall throughput and real-time response capability of the system.

[0069] This multi-objective optimization algorithm and its collaborative working mechanism with hardware accelerators enable the intelligent control system for MicroLED displays to achieve excellent display quality, energy efficiency, and reliability while meeting stringent performance requirements. It not only dynamically and meticulously optimizes MicroLED displays under various complex environments but also anticipates and addresses potential performance degradation and failures, thereby ensuring the display always operates at its best.

Claims

1. A pixel array intelligent control system for a MicroLED display, characterized in that, include: The pixel sensing array module is embedded inside the MicroLED pixel substrate and is used to collect data on the luminous intensity, color coordinates, operating temperature and current fluctuation of each pixel unit in real time to form pixel-level real-time data. The dynamic compensation engine module, connected to the pixel sensing array module, is used to construct a pixel performance degradation model based on the collected pixel-level real-time data and generate independent brightness and chromaticity compensation coefficients for each pixel. An adaptive drive scheduling module, which connects the dynamic compensation engine module to the drive circuit of the MicroLED pixel array, is used to dynamically adjust the duty cycle of the pulse width modulation signal and the amplitude of the drive current of each pixel according to the chromaticity compensation coefficient. An environmental coupling feedback module is deployed on the surface of the display casing and key internal thermal nodes to synchronously collect environmental data such as ambient light intensity, ambient color temperature, ambient temperature and humidity, and transmit the environmental data to the central collaborative control unit. The central collaborative control unit is communicatively connected to the dynamic compensation engine module, the adaptive drive scheduling module, and the environmental coupling feedback module, respectively. It is used to fuse pixel-level real-time data and environmental data, execute global optimization decisions, and output unified drive strategy instructions.

2. The intelligent control system for a pixel array of a MicroLED display according to claim 1, characterized in that, The pixel sensing array module consists of a miniature photoelectric sensor, a temperature sensor, and a current sampling circuit integrated near the cathode or anode of each MicroLED pixel. The miniature photoelectric sensor is an InGaN-based detector integrated with the same process as the MicroLED epitaxial layer, with a spectral response range covering 400 nm to 700 nm. The temperature sensor is a thermistor based on the temperature coefficient of resistance of polycrystalline silicon, with an accuracy of ±0.5 degrees Celsius. The current sampling circuit is implemented by a low-resistance sampling resistor connected in series in the pixel driving path in conjunction with a differential amplifier, with a sampling frequency of not less than 10 kHz.

3. The intelligent control system for a pixel array of a MicroLED display according to claim 1, characterized in that, The dynamic compensation engine module includes a pixel performance modeling submodule and a compensation coefficient generation submodule; The pixel performance modeling submodule uses historical data and current real-time data to fit the brightness decay trend of each pixel using an exponential decay function, and combines it with the color coordinate offset vector to construct a three-dimensional state space model. The compensation coefficient generation submodule calculates the inverse mapping parameters that make the current pixel output approximate the standard chromaticity coordinates and the target brightness based on the three-dimensional state space model, and outputs them as independent red, green and blue three-channel gain coefficients.

4. The intelligent control system for a pixel array of a MicroLED display according to claim 1, characterized in that, The adaptive drive scheduling module includes a pulse width modulation controller and a current amplitude regulator. The pulse width modulation controller receives the gain coefficient from the dynamic compensation engine module and converts it into the PWM duty cycle adjustment amount of the corresponding channel, with an adjustment range of 0% to 100% and a resolution of not less than 12 bits. The current amplitude regulator dynamically limits the maximum drive current based on pixel operating temperature and ambient temperature data to prevent thermal runaway. Its current adjustment step is 0.1 mA, and the adjustment range is 1 mA to 20 mA.

5. The intelligent control system for a pixel array of a MicroLED display according to claim 1, characterized in that, The environmental coupling feedback module includes an ambient light sensor array, a color temperature sensor, and a temperature and humidity sensor group. The ambient light sensor array consists of at least four silicon photodiodes distributed at the four corners of the display, used to detect the spatial distribution of incident light intensity. The color temperature sensor is based on a dual-channel filter and a photoelectric conversion circuit, outputting an ambient color temperature value with a measurement range of 2000K to 10000K. The temperature and humidity sensor group is deployed on the back panel heat sink fins, the top of the driver IC package, and the back of the pixel array, with a sampling period of 1 second.

6. The intelligent control system for a pixel array of a MicroLED display according to claim 1, characterized in that, The central collaborative control unit runs a multi-objective optimization algorithm, which aims to minimize global power consumption, maximize color consistency, and minimize visual latency. The constraints include maximum allowable temperature rise, pixel current safety threshold, and ambient light adaptability requirements. The central collaborative control unit performs a global policy update every 50 milliseconds and synchronizes the updated driving parameters to the adaptive driving scheduling module.

7. The intelligent control system for pixel array of a MicroLED display according to claim 1, characterized in that, The system also includes a fault pixel isolation and redundancy mapping module; the fault pixel isolation and redundancy mapping module monitors the abnormal signals output by the pixel sensing array module. When the brightness of a pixel is lower than the threshold or the current fluctuates abnormally by more than 20% for three consecutive sampling periods, it is determined to be a fault pixel. The system automatically maps the display task of the pixel to its four adjacent normal pixels and performs brightness and chromaticity interpolation compensation through sub-pixel rendering algorithm to ensure image continuity.

8. The intelligent control system for pixel array of a MicroLED display according to claim 1, characterized in that, It supports multi-zone independent dimming function; the central collaborative control unit divides the entire display area into no less than 64 logical dimming zones, each zone independently performs ambient light adaptation and dynamic compensation, and the zone boundary is dynamically adjusted according to the edge features of the image content, with an adjustment delay of no more than 2 frames.

9. The intelligent control system for a pixel array of a MicroLED display according to claim 2, characterized in that, The pixel sensing array module and the dynamic compensation engine module use differential signal transmission with a communication rate of 1 gigabit per second and a bit error rate of less than 10 to the power of negative 9.

10. The intelligent control system for a pixel array of a MicroLED display according to claim 6, characterized in that, The central collaborative control unit adopts a dual-core ARM Cortex-M7 architecture with a main frequency of 480 MHz and has a built-in hardware accelerator for performing color space conversion and Bayesian state estimation.

Images