Dynamic self-repairing interconnection system and method for mini-led display module

Abstract

The application relates to the technical field of micro light emitting diode display and chip testing, and particularly discloses a dynamic self-repairing interconnection system and method for a MiniLED display module, which comprises a honeycomb-shaped redundant interconnection network embedded in a packaging substrate, a programmable switch matrix at a cross node, a failure detection unit, a distributed game algorithm module and a pixel driving control unit; the electrical connectivity and brightness consistency are maintained under single-point or multi-point failure by real-time detection of faults and dynamic reconstruction of a current path based on a non-cooperative game model and a Nash equilibrium principle. The application realizes logical level soft repair through the above scheme, avoids additional physical repair materials, improves the reliability and service life of a high-density MiniLED module, and supports low-delay reconstruction under a high refresh rate.

Description

Technical Field

[0001] This invention belongs to the field of micro-light-emitting diode display and chip testing technology, specifically relating to a dynamic self-healing interconnection system and method for MiniLED display modules. Background Technology

[0002] With the continuous advancement of semiconductor display technology, MiniLED display modules, with their high brightness, high contrast, and excellent light control performance, have become a core solution in the field of high-performance displays. In the process of achieving ultra-high pixel density, the interconnection network inside the packaging substrate carries the power transmission and signal driving tasks for tens of thousands of pixels. Its structural stability and operational reliability determine the visual quality and overall service life of the display terminal.

[0003] Interconnect topology design for MiniLED modules typically requires the introduction of redundant architectures to mitigate the impact of manufacturing defects and aging during service on pixel driving capabilities. By pre-setting backup paths and logic control units in complex circuits, the aim is to achieve current redirection in the event of a fault, ensuring that the display array can maintain basic display functions and image integrity even when facing partial electrical faults.

[0004] Existing interconnect redundancy solutions mostly rely on fixed physical backup paths, which not only leads to low utilization of redundant resources and occupies valuable substrate wiring space, but also makes it difficult to cope with topology instability and local connection islanding caused by multi-point random failures. Due to the lack of dynamic scheduling capabilities for complex network loads, uneven current distribution is easily generated during the repair of failed nodes, which can induce brightness inaccuracies in surrounding pixels and serious current crosstalk risks. Furthermore, traditional repair mechanisms often ignore the coupling game relationship between pixel nodes, failing to achieve globally optimal drive path reconstruction under high dynamic display conditions, resulting in reduced reliability and display consistency of the repaired module. Therefore, a dynamic self-healing interconnect system and method for MiniLED display modules is desired. Summary of the Invention

[0005] The purpose of this invention is to provide a dynamic self-healing interconnection system for MiniLED display modules, which can solve the problems of islanding effect, current crosstalk and low utilization of redundant resources in the above-mentioned background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A dynamic self-healing interconnect system for MiniLED display modules includes a cellular redundant interconnect network, a programmable switch matrix, a distributed game theory algorithm module, a failure detection unit, and a pixel drive control unit, wherein: The honeycomb redundant interconnection network is embedded inside the packaging substrate of the MiniLED display module. It is configured to provide a main path and multiple backup interconnection paths for each MiniLED chip. Each path is interconnected in a hexagonal topology to form an interconnection architecture with local connectivity and global reachability. The programmable switch matrix is ​​integrated at the cross nodes of the cellular redundant interconnection network. It is composed of micro electronic switches and is configured to dynamically switch the on or off state of each interconnection path according to control commands to realize the reconstruction of the current path. The failure detection unit is configured to monitor the electrical status of each interconnection line and solder joint in real time. When an open circuit or short circuit fault is detected, a failure event signal is generated and the location of the failure node is located. The distributed game algorithm module is configured to, upon receiving the failure event signal, treat the failed node and its multiple neighboring MiniLED beads and their associated redundant nodes as independent intelligent agents, initiate path reconstruction negotiation based on a non-cooperative game model, and calculate the interconnection topology reconstruction scheme that makes the global current distribution most uniform and the brightness consistency optimal based on the Nash equilibrium principle. The pixel drive control unit is configured to receive the reconstruction scheme and send corresponding switch control commands to the programmable switch matrix to drive each microelectronic switch to perform on / off operations and complete the dynamic rerouting of the current path.

[0007] Preferably, in the cellular redundant interconnection network, each MiniLED bead is connected to its three adjacent redundant nodes through at least three independent interconnection paths, ensuring that there are still at least two feasible current paths in the event of failure at any single point.

[0008] Preferably, the microelectronic switches in the programmable switch matrix are transistor switches or microelectromechanical system switches, and their response time is less than a preset threshold to meet the real-time reconstruction requirements in high refresh rate display scenarios.

[0009] Preferably, the failure detection unit uses a combination of differential voltage sampling and current loop integrity verification to periodically scan the interconnection lines, which can identify impedance abnormalities or signal interruption faults within a predetermined time.

[0010] Preferably, the distributed game algorithm module embeds a local routing protocol based on the Ad-hoc self-organizing network concept, which enables each agent to complete local game decisions by exchanging state information with its neighboring agents within one or two hops, thus avoiding excessive global communication overhead.

[0011] Preferably, when calculating the reconstruction scheme, the distributed game algorithm module uses the expected driving current of each pixel, the current load state, and the brightness tolerance of neighboring nodes as input parameters of the game utility function to ensure that the current distribution after reconstruction will not cause brightness deviations in surrounding pixels that are perceptible to the human eye.

[0012] Preferably, the pixel driving control unit works synchronously with the display driving timing, and performs the reconstruction operation of the switching matrix during the blanking period of each frame image refresh cycle to avoid introducing current disturbances during the effective display period.

[0013] Preferably, the dynamic self-healing interconnection system further includes a state feedback verification unit, configured to collect actual current distribution data after reconstruction and compare it with the expected reconstruction target. If the deviation exceeds the preset tolerance, a secondary game optimization process is triggered.

[0014] Preferably, the wiring density of the honeycomb redundant interconnect network matches the arrangement density of the MiniLED beads, and all redundant paths are located in the same metal layer or adjacent metal layers of the packaging substrate, without occupying additional external pins or increasing the module thickness.

[0015] Preferably, the distributed game algorithm module supports multi-point concurrent failure handling. When the system detects multiple non-adjacent failed nodes at the same time, it can start multiple local game processes in parallel and coordinate resource competition between each reconstruction area through a conflict resolution mechanism to prevent new failures caused by path overlap.

[0016] Preferably, for failure regions with non-overlapping impact areas, the distributed game algorithm module starts multiple parallel computing threads to simultaneously perform local game decisions; for failure groups with overlapping impact areas, the distributed game algorithm module introduces a coordinating agent role and allocates common redundant path resources among multiple local game processes through a conflict resolution mechanism. The state feedback verification unit is also connected to a miniature photodiode array, which is used to capture the intensity of scattered light near the reconstruction region. If the deviation value of the scattered light intensity reflects a brightness difference that is visible to the naked eye, the global coordination unit will forcibly intervene and fine-tune the brightness weight parameter in the game utility function.

[0017] Preferably, the packaging substrate adopts a glass-based packaging process, and the honeycomb redundant interconnect network is formed by photolithography on the glass substrate, and the roughness of its wire edges is controlled within the nanometer range. Each microelectronic switch in the programmable switch matrix has a built-in ferroelectric storage unit, which is used to physically save the repaired topology state information when the system is powered off. The pixel drive control unit also integrates a load prediction engine, which is configured to parse the image data stream to be displayed and predict the distribution of high brightness areas in the next frame. If the high brightness area overlaps with the path being reconstructed, the pixel drive control unit performs path reinforcement operation before the image is displayed. A cross-module physical interconnection interface is provided at the edge of the display module, configured to request a current compensation loop from the redundant network of adjacent modules via the cross-module communication link when the resources of a single module are insufficient. Compared with the prior art, the present invention has the following beneficial effects: 1. The dynamic self-healing interconnect system for MiniLED display modules provided by this invention achieves logic-level "soft repair" capability without additional physical repair materials by constructing a honeycomb redundant interconnect network and a programmable switch matrix inside the packaging substrate, combined with a distributed reconstruction algorithm based on multi-agent game theory. This system eliminates the resource waste problem caused by traditional fixed redundancy design and can dynamically avoid islanding effects in multi-point random failure scenarios, ensuring the electrical connectivity of the display array.

[0018] 2. By introducing a game-theoretic optimization mechanism with Nash equilibrium constraints, the system actively regulates current distribution to maintain brightness consistency among pixels while repairing failed paths, suppressing potential current crosstalk and visual defects after repair. This system employs localized communication and parallel game-theoretic strategies, reducing algorithm complexity and response latency. It is suitable for high-density, high-refresh-rate MiniLED display applications, improving the long-term reliability and service life of the module. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the overall technical solution architecture according to the present invention; Figure 2 This is a schematic diagram of the core principle framework of path reconstruction based on the non-cooperative game model and the Nash equilibrium principle in this invention. Figure 3 This is a flowchart illustrating the main stages of logic flow from failure detection and location to dynamic current path detour in this invention. Figure 4 This is a schematic diagram illustrating the multi-level interaction relationship and data flow between the cellular redundant interconnection network and the programmable switch matrix according to the present invention. Figure 5 This is a schematic diagram comparing the quadratic game optimization principle based on state feedback verification and expected reconstruction target in this invention. Detailed Implementation

[0020] Example 1: Please refer to the appendix Figure 1 To be continued Figure 5To make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to specific embodiments.

[0021] A dynamic self-healing interconnection system for MiniLED display modules includes a cellular redundant interconnection network, a programmable switch matrix, a distributed game algorithm module, a failure detection unit, a pixel drive control unit, a status feedback verification unit, and a packaging substrate environment monitoring module.

[0022] The cellular redundant interconnect network, serving as the physical layer communication and power transmission base for the entire system, is embedded within the packaging substrate of the MiniLED display module. This network is configured to provide at least one main driving path and multiple redundant backup interconnect paths for each MiniLED chip. In the specific spatial topology logic, each path is interconnected in a highly symmetrical hexagonal cellular topology. This structure not only possesses extremely high tiling efficiency geometrically but also forms an interconnect architecture with strong local connectivity and global reachability in circuit logic.

[0023] The honeycomb redundant interconnect network further includes a main metal trace layer, redundant metal trace layers, and interlayer insulating vias. The main metal trace layer uses a highly conductive nano-copper or gold alloy material, and its linewidth and spacing are precisely matched to the pixel pitch of the MiniLED chips to ensure extremely low parasitic resistance under full-load current conditions. The redundant metal trace layer is located vertically below or above the main metal trace layer and is electrically interconnected with the main trace layer through the interlayer insulating vias. The anode or cathode pin of each MiniLED chip is not directly connected to a single power bus, but is mounted through multiple cross nodes in the honeycomb redundant interconnect network. In this architecture, even if the main metal trace layer experiences a physical break at a specific location or fails due to electromigration, current can still be transmitted through other hexagonal edge paths in the redundant metal trace layers.

[0024] The programmable switch matrix is ​​integrated at all intersection nodes and critical path inflection points of the cellular redundant interconnection network. The programmable switch matrix consists of a massive number of miniature electronic switches configured to dynamically switch the on or off states of each interconnection path segment according to logic instructions from the control terminal, achieving instantaneous reconfiguration of current paths at the logic level.

[0025] The microelectronic switches in the programmable switch matrix are physically implemented using microelectromechanical systems (MEMS) switches or high-performance thin-film transistor (TFT) switches. For the MEMS switch implementation, a miniature mechanical cantilever beam structure is internally incorporated. The cantilever beam is driven to physically contact or separate from the contact electrodes via electrostatic attraction or electromagnetic force, achieving near-zero resistance conduction and extremely high turn-off impedance, which is significant for reducing the overall power consumption of the display module. For the TFT switch implementation, it is integrated into the low-temperature polysilicon layer of the packaging substrate, possessing a fast response speed capable of completing switching actions within nanoseconds. Each microelectronic switch is equipped with an independent latch unit to store the current switch state bit, ensuring that the interconnect topology remains stable even if the control signal momentarily disappears.

[0026] The failure detection unit is configured to monitor the electrical health status of each interconnect line, solder joint, and critical node within the MiniLED display module in real time and at high frequency. Through an integrated sensor array, the failure detection unit can accurately capture precursory signals of failure such as abnormal voltage, sudden current drops, and impedance drift. When an open-circuit or short-circuit fault is detected in a certain path, the failure detection unit immediately generates a failure event signal containing the fault type, fault severity, and precise coordinate information, and quickly locates the physical position of the failed node within the cellular network.

[0027] The failure detection unit integrates a differential voltage sampling circuit. This circuit acquires the potential difference of each interconnection path segment in real time by setting sampling probes at both ends of each interconnection path. The failure detection unit has a built-in reference voltage source and a precision comparator. When the absolute value of the difference between the measured potential difference and the preset theoretical voltage drop based on normal ohmic losses exceeds a preset tolerance threshold, it is determined that the interconnection path segment has aging or impedance abnormalities. The failure detection unit also includes a current loop integrity verification module, which uses periodic test pulses to inspect redundant paths in an idle state, ensuring that the backup path is in good condition when needed. This proactive inspection mechanism improves the system's instantaneous response success rate in the face of sudden failures.

[0028] The distributed game theory algorithm module, serving as the system's decision-making core, is configured to initiate a computational process based on complex network reconstruction upon receiving the failure event signal. This module treats the failed node and its neighboring MiniLED beads, associated redundant interconnect nodes, and available switching path segments as agents with independent decision-making capabilities. Based on a non-cooperative game model, the module initiates negotiation among these agents regarding resource allocation and path occupancy. Following the Nash equilibrium principle, it iteratively calculates an interconnect topology reconstruction scheme that achieves the most uniform global current distribution, minimizes total energy loss, and optimizes the brightness consistency of surrounding pixels.

[0029] In the computational logic of the distributed game theory algorithm module, each agent possesses a predefined utility function. The input parameters of this utility function include at least: the current driving current required by the pixel corresponding to the neighboring node, the real-time thermal load state of the branch containing the neighboring node, the electromagnetic interference increment to adjacent nodes after the interconnection path reconstruction, and the total physical length of the reconstructed path. The game process involves finding a combination of strategies that achieves the overall optimal equilibrium point without any agent being able to unilaterally change the reconstruction strategy to obtain higher efficiency. Compared to traditional heuristic search or greedy algorithms, this game theory-based computational approach can better handle complex, high-density failure scenarios and effectively avoids new overload points caused by local repairs.

[0030] The distributed game algorithm module also embeds a local routing protocol based on the ad-hoc self-organizing network concept. This means that when making path reconstruction decisions, not all agents need to participate in global communication, but only exchange information within a specific number of hops around the fault point. For example, each agent only exchanges its current load weight and available path bandwidth with its neighboring agents within one or two hops. This localized communication strategy compresses the algorithm's convergence time and keeps the communication overhead of the system bus at a low level, enabling real-time applications in large-scale MiniLED arrays.

[0031] The pixel drive control unit is configured to receive the final reconstruction scheme output by the distributed game algorithm module and convert it into a low-level switch control sequence. The pixel drive control unit sends corresponding control commands to the programmable switch matrix via a high-bandwidth dedicated instruction bus, precisely driving each microelectronic switch to perform physical-level on / off operations, guiding current to avoid failure areas, and completing dynamic path rerouting.

[0032] To ensure a smooth and seamless display, the pixel drive control unit is configured to maintain a high degree of synchronization with the timing controller of the display drive system. The pixel drive control unit issues a reconstruction command during the vertical blanking period of each frame refresh cycle. Because the action time of the microelectronic switch is much shorter than the duration of the blanking period, the switching process of the current path is visually completely transparent and does not cause screen flicker or ghosting. Simultaneously, the pixel drive control unit also has a current pre-compensation function. At the instant the reconstruction path is turned on, it fine-tunes the output voltage of the drive circuit to compensate for the increased impedance voltage drop of the new path, ensuring that the brightness of the pixels remains strictly consistent before and after reconstruction.

[0033] The state feedback verification unit is configured to perform closed-loop verification after the interconnect topology reconstruction is completed. The state feedback verification unit collects the actual current data of each branch after reconstruction using current shunt sampling resistors distributed throughout the substrate, and compares this data in real time with the target current distribution map expected by the distributed game algorithm module. If the deviation between the actual current distribution and the expected solution exceeds a preset tolerance range of 5%, the state feedback verification unit sends a correction request to the distributed game algorithm module, triggering a secondary game optimization process. This closed-loop mechanism effectively addresses reconstruction deviations caused by substrate ambient temperature fluctuations or component parameter drift, further improving the system's robustness.

[0034] The packaging substrate environmental monitoring module is configured to monitor the temperature distribution, humidity, and mechanical stress level inside the MiniLED module in real time. These environmental parameters are synchronized to the distributed game algorithm module in real time. In areas with excessively high ambient temperatures, the algorithm automatically lowers the utility weight of the corresponding agent, causing the reconstruction path to shift towards a cooler area. This achieves system-level thermal management optimization through logical means, preventing secondary electrical failures caused by localized overheating.

[0035] The construction of the honeycomb redundant interconnect network is further refined. At the physical wiring level, to achieve the hexagonal topology, the packaging substrate employs multilayer printed circuit technology or glass-based packaging technology. Each metal interconnect layer is divided into multiple repeating lattice units, with a MiniLED chip bonding platform at the center of each lattice unit. The main path extends along the horizontal axis of the lattice unit, while at least three spare interconnect paths are distributed along the hypotenuse of the hexagon, connecting to redundant nodes of adjacent lattice units. Through this dense three-dimensional interconnect, even if a micro-crack occurs in the packaging substrate causing partial metal layer breakage, the system can still find alternative current guiding paths from the remaining effective layers.

[0036] In the logic of handling multi-point concurrent failures in the distributed game theory algorithm module, the system establishes a parallel game model based on a conflict resolution mechanism. When the failure detection unit simultaneously reports multiple non-adjacent failure points, the system determines whether their influence ranges overlap based on the physical distance between the failure points. For non-overlapping failure regions, the distributed game theory algorithm module is configured to start multiple parallel computing threads to perform local game decisions simultaneously. For complex failure groups with overlapping influence ranges, the system automatically escalates the game level, introducing a temporary "coordinating agent" role responsible for allocating common redundant path resources among multiple local game processes to prevent new failures or current backflow caused by path overlap. This hierarchical and categorized processing strategy ensures that the system still has a very high self-repair success rate when facing large-scale process defects or extreme service environments.

[0037] The driving logic of the programmable switch matrix is ​​further detailed. To reduce transient inrush current during switching operations, the pixel drive control unit employs a stepped drive voltage technique. When controlling the microelectronic switch to turn on, the drive voltage does not jump to its maximum value instantaneously, but rather rises gradually according to a preset slope, causing the switch contact resistance to decrease smoothly. This effectively suppresses voltage spikes caused by inductive effects, protecting the fragile epitaxial layer of the MiniLED chip from damage. The system has built-in anti-deadlock logic. Before generating a reconfiguration command, it automatically checks whether each switch combination will cause a short circuit between the power supply and ground. If the check fails, the current command will be forcibly rejected and the algorithm will be required to recalculate.

[0038] In the differential sampling logic of the failure detection unit, an adaptive filtering technique is introduced to eliminate interference from environmental electromagnetic noise on the sampling results. The sampled data first passes through a low-pass filter composed of high-speed operational amplifiers to filter out high-frequency switching noise, and then undergoes multiple sampling and averaging processes in the digital domain. The failure detection unit is configured to dynamically adjust the sampling threshold based on the average brightness of the displayed image: when displaying a bright image (i.e., a large total current), the absolute differential voltage threshold for determining failure is increased; while when displaying a dark image, a high-sensitivity sampling mode is switched. This dynamic threshold strategy reduces the false alarm rate and ensures the detection accuracy of the system under various display conditions.

[0039] When calculating the optimal solution for brightness consistency, the distributed game theory algorithm module also calls a pre-set lookup table database of MiniLED photoelectric characteristics. This database stores the brightness response curves of various batches of MiniLED chips under different current drives and operating temperatures. During the game decision-making process, the algorithm not only considers path resistance but also simulates the predicted luminous efficacy output after the reconstructed drive current acts on a specific chip. By adjusting the equivalent impedance of the reconstructed path or the drive pulse width, it achieves active brightness balance at the physical level. This deep integration, extending from circuit reconstruction to optical quality control, reflects the system's ultimate pursuit in ensuring display performance.

[0040] For applications requiring high refresh rates, such as 240Hz or higher, the distributed game theory algorithm module employs a hardware-accelerated architecture. The Nash equilibrium iterative calculation in game theory is mapped to dedicated hardware operators, leveraging the speed advantage of parallel processing units in large-scale matrix operations to compress the software computation process, which originally required milliseconds, to the microsecond level. This allows the system to complete the entire process from detection to repair within a single frame, even when playing fast-moving scenes, ensuring that the user is unaware of any fault occurrence.

[0041] Example 2: Based on the architecture of Example 1 above, this example provides a dynamic self-healing interconnection system for MiniLED display modules based on a hierarchical clustering game strategy to cope with the massive node computing pressure brought by ultra-high resolution (such as 8K and above).

[0042] In Embodiment 2, the distributed game algorithm module is further optimized into a hierarchical processing architecture, including regional game units and a global coordination unit. The dynamic self-healing interconnect system for the MiniLED display module divides the entire display array into multiple independent display clusters, each containing a predetermined number of MiniLED beads and corresponding redundant interconnect networks.

[0043] When the failure detection unit detects a fault within a display cluster, the local game within that cluster is initiated by the regional game unit. The regional game unit is configured to utilize only the programmable switch resources within that display cluster for path reconstruction. Because the computational scale is limited to a smaller cluster, the convergence speed is improved. However, when the number of failure points within a display cluster becomes too large, causing its own redundant paths to be insufficient to meet the reconstruction requirements (i.e., cluster resources are exhausted), the regional game unit sends a resource request signal to the global coordination unit.

[0044] Upon receiving a signal, the global coordination unit incorporates several adjacent display clusters into the game space. Through a cross-cluster programmable switching matrix, it utilizes idle redundant paths in neighboring, normally functioning display clusters. In this mode, the game agents include not only pixel nodes but also "cluster head agents" representing each display cluster. These cluster head agents engage in a higher-dimensional, non-cooperative game to determine how to allocate cross-regional current transfer quotas. This hierarchical architecture effectively reduces computational complexity from exponential to quasi-linear, enabling stable deployment in large-scale, ultra-high-density MiniLED backlighting or direct-view products.

[0045] In terms of hardware physical layer implementation, the packaging substrate in Example 2 uses a glass substrate process. The glass substrate has better flatness and a lower coefficient of thermal expansion, which is beneficial for maintaining precise impedance consistency in high-density wiring. The honeycomb redundant interconnect network is formed on the glass substrate using photolithography, and the roughness of its conductor edges is strictly controlled at the nanometer level to reduce skin effect losses of high-frequency drive signals. The programmable switch matrix uses an LTPS (low-temperature polycrystalline silicon) transistor array integrated on the glass.

[0046] In Example 2, the state feedback verification unit adds an optical sensing branch. The system deploys a miniature photodiode array at the edge of the MiniLED module or below the light-shielding layer. When path reconstruction occurs, the photodiode array captures the intensity of scattered light near the reconstruction area. If the optical feedback indicates a visible brightness difference in that area, the global coordination unit will intervene, fine-tuning the brightness weight parameters in the game utility function to drive the regional game unit to perform a second iteration until optical consistency reaches a preset standard. This verification mechanism, based on both electrical and optical feedback, provides dual assurance for display quality.

[0047] In Example 2, the distributed game theory algorithm module also incorporates predictive repair logic. By analyzing historical data recorded by the failure detection unit, the system uses a machine learning model to assess the aging trend of each path. For nodes that, while not yet completely failed, exhibit abnormal impedance growth rates, the game theory algorithm initiates "preventive reconfiguration" in advance, smoothly migrating the current to a healthier path before a true failure occurs. This shift from "post-failure repair" to "pre-failure prevention" improves the module's mean time between failures (MTBF).

[0048] Example 3: This example provides a dynamic self-healing interconnect system for MiniLED display modules with energy self-balancing characteristics. Example 3 focuses on the stability of the power distribution network in high-current, large-size modules.

[0049] In Embodiment 3, the cellular redundant interconnect network is designed as a multi-layer impedance-balanced architecture. In addition to conventional signal transmission paths, a dedicated power compensation layer is embedded in the network. The programmable switch matrix is ​​not only used for signal path reconstruction but also responsible for dynamically adjusting the equivalent width of each branch. By activating multiple redundant path segments in parallel, the effective impedance of local circuits can be effectively reduced, suppressing voltage drop losses in high-brightness display mode.

[0050] In Example 3, the distributed game algorithm module employs a Nash equilibrium optimization model with a penalty term. A "hot aggregation penalty term" is introduced into the utility function. When the algorithm detects that a backup path's current density is approaching its physical limit due to frequent borrowing, the penalty term rapidly increases, forcing the game to shift to other backup paths, even if these paths may be longer or have slightly higher impedance. This strategy prevents cascading overheating failures caused by the self-healing process.

[0051] For industrial or automotive-grade applications, the failure detection unit in Example 3 enhances its vibration and electromagnetic interference resistance. It employs differential shielded cables to transmit sampling signals and introduces a hardware-level redundancy voting mechanism. Specifically, each failure point is monitored by three independent voltage detection circuits. The system only triggers a reconfiguration process when two or more detection circuits provide a consistent failure determination, thus preventing malfunctions in harsh environments.

[0052] In Embodiment 3, the pixel drive control unit integrates a load prediction engine. This engine can parse the image data stream to be displayed and predict the high-brightness area in the next frame. If the high-brightness area happens to cover the path being repaired or reconstructed, the pixel drive control unit will communicate with the distributed game algorithm module in advance to complete a stronger path reinforcement before the image is actually displayed, ensuring the stability of the reconstructed topology under the instantaneous impact of high current.

[0053] Example 3 also provides an emergency manual intervention mode. Through a reserved external debugging interface, maintenance personnel can bypass the distributed game theory algorithm module and directly manually program the locking state of the programmable switch matrix. This provides a final technical means for system final inspection on the production line or disaster recovery in extreme situations.

[0054] In the distributed game algorithm module, the game process is specifically detailed into the following steps: First, initialize the policy space. When a failure signal is triggered, the system automatically defines all effective switching nodes within the physical radius centered on the failure point, and uses the currently switchable on / off combinations of these nodes as the initial game policy set. Second, calculate utility evaluation. For each combination in the policy set, calculate the resulting change in current distribution. Calculate the sum of squares of the difference between the measured current and the standard current of each affected pixel node, using this as a negative feedback index to measure brightness consistency. Simultaneously, calculate the total resistance value of the reconstructed path, using it as an index to measure power consumption. Third, iterate the game. The agent updates its own policy based on the policy selection of neighboring nodes in the previous round to maximize its own utility. Utility maximization here is defined as: restoring the normal driving current of the current node as much as possible without increasing the current load of neighboring nodes. Fourth, convergence determination and scheme generation. When the policy changes of all agents are less than a preset small constant, or the number of iterations reaches the upper limit, the game is determined to have reached Nash equilibrium, and the current optimal switching combination is output.

[0055] The programmable switch matrix in Embodiment 3 also employs self-holding technology. Each microelectronic switch has a built-in tiny ferroelectric storage unit. This means that even when the entire display module is completely powered off, the repaired topology information is still physically stored inside the switch. When the module is powered on again, the system does not need to re-perform failure detection and game theory calculations, and can immediately restore to the previous repaired state, achieving the characteristic of "power-on repair" and shortening the power-on self-test time.

[0056] For large-size MiniLED video wall applications, this system also supports cross-module interconnection repair. At the edges of each module, a honeycomb redundant interconnect network provides physical interconnection interfaces. When a large-area failure occurs at the edge of a module, rendering its own redundancy insufficient, the pixel drive control unit can request assistance from the redundant networks of adjacent modules to provide a current loop through the cross-module communication link. This cross-boundary collaboration mode further enhances the overall survivability of the giant display system.

[0057] The state feedback verification unit also integrates long-term aging monitoring logic. It records the changes in electrical parameters before and after each refactoring in non-volatile memory, forming a "module health profile". The distributed game algorithm module refers to this profile to avoid "sub-healthy" paths that show early signs of aging in future repair decisions, achieving a higher level of reliability management.

[0058] Furthermore, the distributed game theory algorithm module supports multi-priority task scheduling. In some critical task displays (such as medical images or aviation instruments), the system can set specific pixel areas as "highest priority." When these areas fail, the game theory algorithm will grant these areas absolute resource control, even at the cost of sacrificing the display quality of non-critical areas (such as reducing brightness or temporarily borrowing their drive paths) to ensure the electrical continuity of the core areas. This business logic-based resource scheduling makes the system more industry-specific.

[0059] In summary, this invention successfully transforms the originally static and rigid circuit interconnection into a dynamic and adaptive intelligent network by constructing a honeycomb-shaped flexible interconnection base at the physical layer and introducing a highly intelligent game theory algorithm at the logic layer. It not only fundamentally solves the persistent problems of islanding effects and current crosstalk in MiniLED display modules, but also improves product yield, reliability, and visual consistency through hardware and software collaboration, paving the way for the development of ultra-large size and ultra-high resolution display technologies.

[0060] Those skilled in the art will understand that although the present invention has been described in detail through the above embodiments, various modifications and optimizations can be made to the system configuration, module functions, hardware selection, and algorithm details of the present invention without departing from the spirit and scope of the invention. For example, the game algorithm can be replaced with other types of distributed optimization algorithms, or the microelectronic switch can be integrated into a packaging substrate of different material systems.

[0061] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A dynamic self-healing interconnection system for MiniLED display modules, characterized in that, include: Cellular redundant interconnection network, programmable switch matrix, distributed game algorithm module, failure detection unit and pixel drive control unit; The honeycomb redundant interconnection network is embedded inside the packaging substrate of the MiniLED display module. It is configured to provide a main path and multiple backup interconnection paths for each MiniLED chip. Each path is interconnected in a hexagonal topology to form a physical interconnection architecture with local connectivity and global reachability. The programmable switch matrix, integrated at the cross nodes of the cellular redundant interconnection network, is composed of microelectronic switches and is configured to dynamically switch the on or off state of each interconnection path according to control commands, so as to realize the reconstruction of the current path at the physical level. The failure detection unit is configured to monitor the electrical status of each interconnection line and solder joint in real time, and generate a failure event signal containing the location information of the failure node when an open circuit or short circuit fault is detected. The distributed game algorithm module is configured to, upon receiving the failure event signal, treat the failure node and its adjacent MiniLED beads and their associated redundant nodes as independent intelligent agents, initiate path reconstruction negotiation based on a non-cooperative game model, and calculate an interconnection topology reconstruction scheme that balances the uniformity of global current distribution and the consistency of brightness based on the Nash equilibrium principle. The pixel drive control unit is configured to receive the reconstruction scheme and send corresponding switch control commands to the programmable switch matrix to drive each micro electronic switch to perform on / off operations.

2. The dynamic self-healing interconnection system for MiniLED display modules according to claim 1, characterized in that, The cellular redundant interconnect network includes a main metal trace layer, a redundant metal trace layer, and insulating vias disposed between the layers. The main metal trace layer uses highly conductive nano-copper or gold alloy material, and its line width and line spacing are matched according to the pixel spacing of the MiniLED lamp beads. The redundant metal trace layer is located in the vertical projection direction of the main metal trace layer and is electrically connected to the main metal trace layer through the interlayer insulating vias. In the honeycomb redundant interconnection network, the anode pin or cathode pin of each MiniLED lamp bead is mounted on different interconnection path segments through at least three independent cross nodes to ensure that in the event of physical breakage or electromigration failure of any single path segment, there are still at least two feasible current transmission paths inside the system.

3. The dynamic self-healing interconnection system for MiniLED display modules according to claim 1, characterized in that, The microelectronic switches in the programmable switch matrix are microelectromechanical system switches or high-performance thin-film transistor switches. When a microelectromechanical system switch is used, its internal structure includes a micromechanical cantilever beam and contact electrodes, and is configured to drive the micromechanical cantilever beam to physically contact or separate from the contact electrodes through electrostatic attraction or electromagnetic force. When a high-performance thin-film transistor switch is used, it is integrated within the low-temperature polysilicon layer of the packaging substrate, and each high-performance thin-film transistor switch is electrically connected to an independent latch unit. The latch unit is used to store the current switch state bit information to maintain the static stability of the interconnect topology. When the pixel drive control unit controls the micro electronic switch to be turned on, it adopts a stepped drive voltage technology to make the drive voltage rise slowly at a preset slope, thereby reducing the rate of change of contact resistance at the moment of switch action.

4. The dynamic self-healing interconnection system for MiniLED display modules according to claim 1, characterized in that, The failure detection unit integrates a differential voltage sampling circuit, a current loop integrity verification module, and an adaptive filtering module. The differential voltage sampling circuit obtains the real-time potential difference of each interconnection path segment by setting sampling probes at both ends of each interconnection path segment, and compares the real-time potential difference with a preset reference voltage. When the absolute value of the difference between the real-time potential difference and the preset voltage drop exceeds the preset fault threshold, it is determined that there is an impedance abnormality in the interconnection path segment; the current loop integrity verification module is configured to use periodic test pulses to inspect the redundant interconnection paths in the idle state; the adaptive filtering module is configured to dynamically adjust the sampling sensitivity according to the current average brightness of the MiniLED display module, and filter out high current noise interference by increasing the absolute voltage difference threshold for determining failure when displaying a bright image.

5. The dynamic self-healing interconnection system for MiniLED display modules according to claim 1, characterized in that, The distributed game algorithm module has embedded local routing logic based on the self-organizing network communication protocol, which enables each agent to exchange state information only with its neighboring agents within one or two hops. In the computation process of the distributed game algorithm module, each agent corresponds to a utility function. The input parameters of the utility function include: the real-time driving current intensity value required for the corresponding pixel, the thermal load value of the branch, the predicted increment of electromagnetic interference generated by the reconstruction path to adjacent nodes, and the total physical length of the reconstruction path. The distributed game algorithm module searches for a strategy combination that maximizes the sum of the utility functions of each agent through multiple iterations. During the iteration process, each agent updates its own on / off state of its switch node based on the path selection strategy of its neighboring nodes in the previous iteration, until the change in the strategies of all agents is less than the preset convergence constant, at which point the Nash equilibrium state is determined to be reached.

6. The dynamic self-healing interconnection system for MiniLED display modules according to claim 1, characterized in that, It also includes a state feedback verification unit, which collects the actual current data of each interconnected branch after reconstruction by means of current shunt sampling resistors distributed at various locations on the packaging substrate, and compares them in real time with the expected target current distribution output by the distributed game algorithm module. If the percentage deviation between the actual current data and the expected target current distribution exceeds the preset tolerance range of 5%, the state feedback verification unit sends a correction request to the distributed game algorithm module to trigger the secondary game optimization process. The state feedback verification unit is also connected to a non-volatile memory, which is used to record the electrical parameter changes before and after each reconstruction to build a module health profile. The distributed game algorithm module retrieves the module health profile in subsequent reconstruction decisions and avoids sub-healthy paths marked as abnormal impedance growth rates in the profile.

7. The dynamic self-healing interconnection system for MiniLED display modules according to claim 1, characterized in that, It also includes a packaging substrate environment monitoring module, which is configured to monitor the temperature field distribution, humidity level and mechanical stress deformation data inside the MiniLED display module in real time, and synchronize the environmental monitoring data to the distributed game algorithm module; In areas where the ambient temperature exceeds a preset safe temperature threshold, the distributed game algorithm module automatically lowers the utility weight of the agents in the corresponding area, causing the reconstruction path to shift towards a cooling area with a lower ambient temperature. The pixel drive control unit is hardware synchronized with the timing controller of the display drive system and is configured to issue on / off control commands to the programmable switch matrix only during the vertical blanking period of each frame image refresh cycle. The pixel drive control unit has a current pre-compensation function, which can offset the additional impedance voltage drop introduced by the reconstruction path by fine-tuning the output voltage of the drive circuit.

8. The dynamic self-healing interconnection system for MiniLED display modules according to claim 1, characterized in that, The distributed game algorithm module adopts a hierarchical processing architecture, including regional game units and a global coordination unit; The dynamic self-healing interconnection system for MiniLED display modules divides the entire display array into multiple independent display clusters, each display cluster containing a predetermined number of MiniLED beads and corresponding redundant interconnection network segments. When the failure detection unit locates a failure node within a display cluster, the corresponding regional game unit first performs local game calculations using the cluster's path resources. If the available redundant path resources within the display cluster are exhausted and the current loop cannot be closed, the regional game unit sends a resource request signal to the global coordination unit. The global coordination unit then borrows idle redundant paths from adjacent normally displaying display clusters through a cross-cluster programmable switch matrix and performs high-dimensional game calculations between cluster head agents representing different display clusters to allocate cross-regional current transmission quotas.

9. The dynamic self-healing interconnection system for MiniLED display modules according to claim 1, characterized in that, The distributed game algorithm module supports a parallel processing mode for multiple concurrent failures. When the failure detection unit detects multiple failure nodes that are not physically adjacent, the distributed game algorithm module determines whether there is an overlapping area based on the radius of influence of each failure point. For failure regions with non-overlapping impact areas, the distributed game algorithm module starts multiple parallel computing threads to simultaneously perform local game decisions; for failure groups with overlapping impact areas, the distributed game algorithm module introduces a coordinating agent role and allocates common redundant path resources among multiple local game processes through a conflict resolution mechanism. The state feedback verification unit is also connected to a miniature photodiode array, which is used to capture the intensity of scattered light near the reconstruction region. If the deviation value of the scattered light intensity reflects a brightness difference that is visible to the naked eye, the global coordination unit will forcibly intervene and fine-tune the brightness weight parameter in the game utility function.

10. The dynamic self-healing interconnection system for MiniLED display modules according to claim 1, characterized in that, The packaging substrate adopts a glass-based packaging process, and the honeycomb redundant interconnect network is formed by photolithography on the glass substrate, with the roughness of the wire edges controlled within the nanometer range. Each microelectronic switch in the programmable switch matrix has a built-in ferroelectric storage unit, which is used to physically save the repaired topology state information when the system is powered off. The pixel drive control unit also integrates a load prediction engine, which is configured to parse the image data stream to be displayed and predict the distribution of high brightness areas in the next frame. If the high brightness area overlaps with the path being reconstructed, the pixel drive control unit performs path reinforcement operation before the image is displayed. A cross-module physical interconnection interface is provided at the edge of the display module, which is configured to request the redundant network of the adjacent module to provide a current compensation loop through the cross-module communication link when the resources of a single module are insufficient.

Images