An aviation obstacle light cluster synchronization control system and method with a virtual master light self-election mechanism

Abstract

The application discloses an aviation obstruction light cluster synchronization control system and method with a virtual master light self-election mechanism, and belongs to the technical field of aviation obstruction light control. A cluster construction module is used to group multiple aviation obstruction light nodes into a distributed peer-to-peer communication cluster. A state monitoring module is used to periodically broadcast heartbeat messages by each node, generate and maintain a cluster member state table. A virtual master light election module is used to generate a unique virtual master light in the cluster based on a distributed consensus mechanism when the current virtual master light working state is abnormal and there is no node in the reconnect detection state in the cluster. A synchronization control module is used to broadcast synchronization flashing instructions from the virtual master light to the remaining nodes. A node recovery module is used to make the original virtual master light rejoin as a slave node and not trigger a new election when there is an existing effective virtual master light. The technical problems of relying on a fixed master node, single point failure requiring manual intervention and cluster brain split are solved.

Description

Technical Field

[0001] This invention relates to the field of aviation obstruction light control technology, specifically to an aviation obstruction light cluster synchronization control system and method with a virtual master light self-election mechanism. Background Technology

[0002] Aviation obstruction lights are crucial navigational aids for identifying tall obstacles such as high-rise buildings, power transmission towers, chimneys, and communication towers. With the increasing number of skyscrapers in cities and the gradual opening of low-altitude airspace, multiple obstruction lights must flash synchronously to create a unified warning signal. Current mainstream technology employs a fixed master-slave architecture: a master light issues a synchronization command, and the remaining lights receive and flash synchronously. While this solution is simple in structure and widely used, the existing fixed master-slave architecture still has the following technical shortcomings in actual operation: 1. A single point of failure can cause system paralysis. The master light is fixedly bound to the physical node. Once the master light goes offline due to lightning strikes, power failures, or other reasons, all slave nodes will be unable to work. For example, if the master light is damaged by lightning, all other slave nodes will be unable to work, and the civil aviation department will issue a notice for rectification within a specified period. Second, the system lacks automatic fault recovery capabilities. After the main control light fails, manual climbing is required for repair, which has a long response time and poses a significant safety hazard. For example, after the main control light is damaged, manual climbing is required for repair, which has a long recovery time, during which there are no aviation warnings in the airway. Third, the main controller selection is static and lacks dynamic optimization. The main controller light is preset during installation and cannot be switched based on signal strength, health status, etc. There is also no reasonable processing logic after the fault node is restored. For example, the light with the weakest signal is fixed as the main controller because it has the smallest identifier, which causes many surrounding auxiliary lights to frequently lose synchronization. Summary of the Invention

[0003] The purpose of this invention is to provide a synchronous control system and method for an aviation obstruction light cluster with a virtual master light self-election mechanism. This system achieves dynamic decoupling of the virtual master light role through a distributed cluster architecture, automatic detection of master light faults through heartbeat detection, automatic election and role switching of healthy nodes through distributed consensus voting, and reasonable management after fault recovery through forced degradation of recovering nodes. This addresses the technical problems of existing fixed master-slave architectures, such as the fixed binding of physical nodes to the master light, leading to system paralysis due to single-point failures; the lack of automatic recovery capability after a master light failure, requiring manual maintenance; and the static master light election process, which cannot dynamically optimize the election based on node health.

[0004] The objective of this invention can be achieved through the following technical solution: This application provides an aviation obstruction light cluster synchronization control system with a virtual master light self-election mechanism, comprising: The cluster building module is used to build a distributed cluster of multiple aviation obstruction light nodes, with each node communicating with the others in a peer-to-peer manner. The status monitoring module is used for each node to periodically broadcast heartbeat messages, receive heartbeat messages broadcast by other nodes, and monitor the online status and health score of each node in real time, and generate and maintain a cluster member status table; the cluster member status table includes at least the node identifier, current role identifier, online status, health score and node reconnection detection status of each node; the reconnection detection status is used to indicate whether the node is in the process of link reconnection detection. The virtual main light election module is used to read the cluster member status table and filter healthy nodes in the cluster. If the current virtual main light is in an abnormal working state and there are no nodes in the reconnection detection state in the cluster, a unique virtual main light is elected from the healthy nodes in the cluster based on a distributed consensus mechanism. The virtual main light is a logical control role that issues synchronous blinking commands. The abnormal working state is defined as the current virtual main light node's cumulative heartbeat loss reaching a preset heartbeat threshold. Furthermore, the present invention also provides a synchronous control system for an aviation obstruction light cluster with a virtual master light self-election mechanism, comprising the following steps: S1. After each node powers on and starts up, it broadcasts a probe message to detect and discover online nodes in the cluster, establishes a two-way peer-to-peer communication link to complete the cluster networking, and uniformly configures the initial role of all nodes as slave nodes. S2. After the cluster network is completed, each node confirms that there is no valid virtual main light in the cluster and no node is in the reconnection detection state according to the local cluster member status table. Then, it waits for a preset random backoff time and verifies through mutual voting among nodes. After obtaining confirmation from more than half of the online healthy nodes, the first and only virtual main light node in the cluster is elected. S3. During cluster operation, each node periodically broadcasts heartbeat messages and updates the local cluster member status table in real time based on the received heartbeat messages. If the current virtual main light node is in the reconnection detection state, the missing heartbeat of that node will not be counted in the failure count. When the cumulative number of missing heartbeats reaches the preset heartbeat threshold, the current virtual main light node is determined to be in an abnormal working state. S4. When an abnormal working status of the current virtual main light node is detected and no node in the reconnection probe state is in the local cluster member status table, an election voting request is initiated after the random backoff timer expires. Each node completes the voting verification based on its online status, reconnection probe status, and health score. After obtaining confirmation from more than half of the online healthy nodes, the corresponding node is switched to the new virtual main light and the election results are broadcast to the entire cluster network. S5. The selected virtual master light node generates a synchronization timing reference signal using its local clock as the cluster synchronization timing reference, configures unified flashing mode parameters, and broadcasts the synchronization flashing command carrying the synchronization timing reference signal and flashing mode parameters to all slave nodes in the cluster. After receiving the command, each slave node calibrates its local clock according to the synchronization timing reference signal and performs synchronized flashing of the lights according to the flashing mode parameters. S6. When it is detected that the original virtual master node has resumed heartbeat broadcasting, ended the reconnection probe state and is stably online, if there is a valid virtual master node in the cluster, the restored node will be locked and configured as a slave node, and a new round of cluster election will not be triggered; at the same time, a node role change notification will be broadcast to the entire network, so that the other nodes will update the online status, role identifier and reconnection probe status of the corresponding node in the local cluster member status table.

[0005] The beneficial effects of this invention are as follows: This invention uses a distributed cluster architecture to form a peer-to-peer communication network with multiple aviation obstruction light nodes. The virtual master light is dynamically decoupled from the physical node as a logical role. With the help of heartbeat timeout perception and automatic election, it solves the problem of single-point paralysis caused by master light binding in a fixed master-slave architecture and realizes the dynamic migration of the master light role among healthy nodes. The status monitoring module enables each node to periodically broadcast heartbeat messages and monitor health scores in real time. When the cumulative heartbeat loss reaches a preset threshold, the main light is automatically identified as abnormal and an election is initiated. This solves the problem of lack of automatic recovery capability after the main light fails and the need for manual climbing for maintenance, and realizes automatic detection and automatic recovery of main light failure. By using a health threshold screening and voting decision mechanism to elect the master light only from healthy nodes, and by forcibly demoting fault recovery nodes to slave nodes, the problems of static master light election being unable to dynamically select the best and faulty nodes preempting the master light after recovery causing oscillations are solved. This achieves dynamic selection of the master light and reasonable management of recovery nodes. Attached Figure Description

[0006] To better understand and implement this application, the technical solution is described in detail below with reference to the accompanying drawings.

[0007] Figure 1 A schematic diagram of the structure of an aviation obstruction light cluster synchronization control system with a virtual master light self-election mechanism provided in this application; Figure 2 This is a flowchart illustrating a method for synchronous control of an aviation obstruction light cluster with a virtual master light self-election mechanism, as provided in this application. Detailed Implementation

[0008] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, exemplary embodiments will be described in detail below, examples of which are illustrated in the accompanying drawings. In the following description, when referring to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0009] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0010] The following detailed description of the specific implementation methods, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided in detail.

[0011] Example 1, please refer to Figure 1 This embodiment provides a synchronous control system for aviation obstruction lights cluster with a virtual master light self-election mechanism, which is deployed in aviation obstruction light intensive application scenarios; The scenario is equipped with multiple aviation obstruction light nodes (this embodiment uses five as an example; the number can be expanded according to the coverage area in actual applications and is not limited thereto); each aviation obstruction light node integrates an LED light group, an embedded control board, a GNSS timing module (supporting GPS / BeiDou dual-mode, used to provide absolute time reference), an industrial-grade communication module (supporting IEEE 802.11 series wireless protocols and Modbus TCP industrial Ethernet protocol), and environmental sensors (including but not limited to temperature and humidity sensors, vibration sensors, and power voltage monitoring modules); the embedded control board can use an industrial-grade microcontroller (such as ARM). The communication module can be configured with a Cortex-M core or other types of processors with equivalent computing power. The specific frequency band and speed of the communication module can be configured according to the on-site electromagnetic environment. Each aviation obstruction light node is independently equipped with a power management unit and a communication protocol conversion module. All nodes are networked via a hybrid industrial Ethernet and wireless Mesh network to form a distributed peer-to-peer communication cluster. Each node communicates with others in a peer-to-peer manner, and any node has the qualification to become a candidate for virtual main light. There is no preset fixed master control node within the cluster; the virtual main light is a logical control role dynamically elected based on a distributed consensus mechanism and is not fixedly bound to any physical node. The system also includes a ground monitoring terminal (running on an industrial control computer) connected to each node for real-time display of cluster member status, main light identification, and synchronous flashing status information. In this embodiment, the node identification can use a device MAC address or a custom UID / IP address. The preset coverage area is determined based on the nominal communication distance of the communication module and corrected through on-site testing. The system includes a cluster construction module, a status monitoring module, a virtual master light election module, a synchronization control module, and a node recovery module; Furthermore, the cluster building module is used to form a distributed cluster of multiple aviation obstruction light nodes, and the nodes communicate with each other in a peer-to-peer manner. This module includes a node discovery unit, a link establishment unit, and a role initialization unit, which work together in sequence to complete the cluster networking and initial identity configuration. The node discovery unit is used to send probe messages to the preset coverage area via UDP broadcast protocol at preset intervals after each node is powered on. The probe messages contain the node identifier and communication port information of the node itself. It also receives response messages from other nodes, which contain the node identifier and communication address of the corresponding node. In this way, it collects the node identifiers and communication addresses of the currently online aviation obstruction light nodes within the preset coverage area and generates a local neighbor node list. If no response is received from a discovered node after multiple consecutive broadcasts, the node is marked as a node to be confirmed and its broadcast is monitored until it resumes broadcasting. Specifically, in this embodiment, the preset period is set to 2 seconds (i.e., the probe message broadcast period), and its configuration is based on the following: Considering the balance between fault detection timeliness requirements (the detection delay of a node joining the cluster should be controlled within 30 seconds) and network bandwidth consumption, in a 5-node scale, the single-node broadcast traffic corresponding to a 2-second period is approximately 0.5kbps, and the total broadcast traffic of 5 nodes is approximately 2.5kbps, which is far lower than the available bandwidth of industrial Ethernet and wireless Mesh networks. Therefore, this value satisfies the detection timeliness without constituting a network bottleneck. In other embodiments, the preset period can be dynamically scaled proportionally according to the number of nodes to ensure network stability under large-scale clusters. The number of consecutive packet loss judgments is set to 3 times, based on the following: In an industrial Wi-Fi environment, the measured value of a single UDP broadcast packet loss rate is approximately 0.1%~1%, and the probability of 3 consecutive losses is one in a million to one in a billion, which is sufficient to eliminate misjudgments caused by accidental packet loss. The link establishment unit is used to send a link establishment request to the target node based on the node identifier and communication address; if a confirmation response is received from the peer, a bidirectional peer-to-peer communication link is established between the local node and the target node, and the cluster communication frequency, data transmission rate and data encryption rules corresponding to the bidirectional peer-to-peer communication link are determined; if the request times out without a response, the link establishment request is repeatedly initiated, and after retries exceeding a preset number of requests, the current link establishment operation is terminated and the target node is recorded as currently unreachable, and the reconnection detection status of the target node in the cluster member status table stored locally by the local node is marked as reconnection detection in progress; Specifically, in this embodiment, the preset number of requests is set to 3 (i.e., the number of link establishment retries). According to the three-retry coverage criterion, 3 retransmissions are sufficient to cover most transient network failures, and failures after 3 attempts are considered permanent link failures. The link establishment request timeout is configured to 200 milliseconds (default value), based on the fact that the average round-trip latency of industrial Ethernet is 3-10 milliseconds, and a margin of 100 times is taken to cover the worst case of protocol stack processing and wireless mesh multi-hop forwarding. If the round-trip latency of multiple consecutive link establishment requests exceeds 100 milliseconds during operation, the system automatically adjusts the timeout based on the most recent measured timeout. The dynamic value of round-trip delay statistics; after the link is established, a two-way heartbeat keep-alive mechanism is maintained between the local node and the target node. If no application layer heartbeat confirmation is received from the other party for three consecutive times, the link is determined to be interrupted and the reconnection process is automatically triggered; after the link establishment operation is terminated, the state detection subunit immediately starts listening, and the listening start time is the time when the link establishment is terminated; wherein, the unreachable flag and the reconnection detection state are two representations of the same state in the reconnection detection. The existence of the unreachable flag is equivalent to the reconnection detection state of the node being in the reconnection detection. Both are set and cleared synchronously in the cluster member status table stored locally on the local node. The role initialization unit is used to generate an initial role identifier locally on the node after the bidirectional peer-to-peer communication link is established, and configure the initial role of the node as a slave node; after the link is established, all nodes join the cluster as slave nodes and wait for the election result of the virtual master election module; in this embodiment, the initial term number of each node is uniformly set to a preset initial value (such as 0), the synchronization sequence number is uniformly set to a preset initial value, and the initial role identifier is a slave node. As an output of the cluster construction phase, through the coordinated efforts of the node discovery unit, link establishment unit, and role initialization unit, and via a three-layer mechanism of probe message broadcasting, link establishment and retry, and unified initial role configuration, all nodes in the cluster are networked as slave nodes, providing a two-way peer-to-peer communication link foundation and a unified initial state for the subsequent virtual master election. Furthermore, the link establishment unit also includes a status detection subunit, a reconnection triggering subunit, and an information update subunit, which are used to handle reconnection detection and status recovery after a node goes offline. When the link is interrupted due to a fault, these three subunits work together to complete the detection of unreachable nodes and link reconstruction. The state detection subunit is used to read the currently unreachable node information recorded locally by the node and continuously monitor whether the target node rebroadcasts the probe message within a preset time window. In this embodiment, the preset time window is set to 10 seconds (i.e., the reconnection probe time window), which is based on the five-times probe cycle coverage criterion. Based on the preset cycle of 2 seconds, five times the preset cycle is taken (covering 5 probe opportunities) to ensure that if the target node does not resume broadcasting within 5 consecutive probe cycles, it is highly likely to be in a physically offline state. The continuous monitoring is implemented in an event-driven manner. The node registers a probe message receiving callback function within the preset time window. Once the probe message of the target node is received, the callback is triggered without polling. The reconnection trigger subunit is used to resend a link establishment request to the target node when the status detection subunit receives a probe message broadcast by the target node within the preset time window, and restore the bidirectional peer-to-peer communication link with the target node after receiving the confirmation response from the peer, remove the unreachable mark of the target node recorded locally by the node, update the online status of the node in the cluster member status table stored locally by the node to online, and update the reconnection detection status of the node to not being in reconnection detection. The information update subunit is used to update the currently unreachable node information to offline node information and clear the reconnection probe status mark of the node in the cluster member status table when the status detection subunit has not received the probe message broadcast by the target node within the preset time window. Specifically, in this embodiment, the three complete migration paths of the reconnection detection state are as follows: After the link establishment unit fails to establish a link, the node's reconnection detection state switches from normal to "reconnecting in progress"; while in progress, after the status detection subunit receives a probe message within the preset time window and the reconnection triggering subunit successfully establishes a link, the state is updated to "not in progress"; when the preset time window expires and the information update subunit still has not received a probe message, the reconnection detection state flag is cleared. This is to prevent individual nodes from affecting the normal operation of the cluster due to prolonged periods in the "reconnection in progress" state. In addition, this embodiment also includes a forced clearing mechanism for reconnection status timeout: when the duration of the reconnection probe status of any node exceeds a preset absolute timeout threshold (set to 30 seconds in this embodiment, based on a six-times reconnection probe window, i.e., 6 × 10 seconds = 60 seconds halved, ensuring that reconnection attempts have been fully executed in most cases), the system automatically determines that the node is in a permanent offline state, forcibly clears its reconnection probe status and updates it to offline, so as to allow the cluster to perform fault recovery election normally when a majority of nodes are healthy; this timeout threshold can be remotely configured through the ground monitoring terminal according to the stability of the on-site network, and is not limited to a fixed value; Specifically, the link establishment unit achieves automatic reconnection detection and state self-healing of unreachable nodes through the coordinated cooperation of the state detection subunit, reconnection triggering subunit, and information update subunit, and avoids long-term blockage of the election process by the reconnection state through the timeout forced clearing mechanism. Furthermore, after the cluster construction module completes the network setup and all nodes join the cluster as slave nodes, the status monitoring module starts working. This module is used for each node to periodically broadcast heartbeat messages, receive heartbeat messages broadcast by other nodes, and monitor the online status and health score of each node in real time, generating and maintaining a cluster member status table. The cluster member status table includes at least the node identifier, current role identifier, online status, health score, and node reconnection detection status for each node. This cluster member status table is the core data basis for the subsequent election module to screen candidate nodes and make voting decisions. The status monitoring module includes a heartbeat sending unit, a heartbeat receiving unit, and a timeout determination unit. The heartbeat sending unit is used by each node to broadcast heartbeat messages to other nodes in the distributed cluster at fixed intervals; the heartbeat message includes the node identifier, current role identifier, timestamp, and node health score of the node; as a preferred extension, this embodiment additionally carries term number and synchronization sequence number for distributed time-series auxiliary synchronization, and the extended fields are not mandatory. Specifically, in this embodiment, the fixed period is set to 1 second (i.e., the heartbeat period), based on the requirement that the response time of critical equipment failures in the aviation obstruction light system should not exceed 3 seconds. Based on this, the fixed period of 1 second, combined with three consecutive loss detections, results in a fault detection time of approximately 3 seconds (3 times × 1 second), precisely meeting the above requirement. This setting ensures timely fault detection while avoiding network bandwidth consumption caused by excessively high heartbeat frequencies. Before sending a heartbeat message, each node calculates a health score based on its current operating status, with an initial value of 100 points, which is then progressively reduced using a weighted deduction system. When a node fails to send probe messages or link establishment requests to the same target node three times in a row, 5 points are deducted each time. According to the principle of minor failure penalty, a single communication failure is not considered a node failure and only a minor point is deducted. Only after multiple consecutive failures will the points be accumulated to the unhealthy state. When the node power supply voltage exceeds the rated range (±10% of the rated voltage of 220 volts, i.e., below 198 volts or above 242 volts), 10 points will be deducted. Power supply abnormalities have a significant impact on equipment operation, and the deduction value is relatively high. When the internal temperature of a node exceeds 70℃ (based on the upper limit of the operating temperature range of LED driver chips and communication modules of 85℃, 70℃ is taken as the warning threshold to leave a 20% margin), 15 points will be deducted. Exceeding the temperature limit may cause the equipment to reduce its frequency or be permanently damaged, which is a serious fault. When the bit error rate of the node communication module exceeds one in a thousand (the bit error rate is obtained by internal statistics of the communication module; based on the limit of the forward error correction capability of the wireless communication module, that is, the receiver cannot correctly recover the original data when the bit error rate is higher than one in a thousand), 8 points will be deducted; The minimum score is 0. A node is considered unhealthy if its score is below the preset health threshold of 60 and will not participate in the election or voting confirmation. The preset health threshold of 60 is set based on the principle of eliminating multiple faults: when a node simultaneously experiences power abnormality (deduct 10 points), excessive temperature (deduct 15 points), and high communication error rate (deduct 8 points), a total of 33 points are deducted, leaving 67 points, which is still higher than 60 points. Only when a node simultaneously experiences power abnormality, excessive temperature, high communication error rate, and multiple consecutive communication failures (deduct another 15 points), a total of 48 points are deducted, leaving 52 points, which is lower than 60 points, is it considered unhealthy. This avoids misjudging a healthy node due to a single momentary abnormality. Specifically, in this embodiment, the complete calculation and update rules for the health score are as follows: The counting object for three consecutive failures in the deduction trigger condition is: the communication between this node and the same target node fails three times in a row, that is, the node does not receive a response after sending a probe message or connection request, and the failure event occurs consecutively in time without any successful communication in between; if there is at least one successful communication (receiving a response) in between, the consecutive failure count is cleared and re-accumulated. When multiple faults exist simultaneously, the deductions for each fault are calculated cumulatively, with a maximum total deduction of 100 points, meaning a minimum score of 0 points. Negative values ​​are not allowed. The health score has an automatic recovery mechanism: when the fault is eliminated (such as the temperature returning to the normal range or the power supply voltage stabilizing), the deducted points are recovered according to the following rules: every 10 complete heartbeat cycles (i.e., 10 seconds) is a recovery cycle, and the recovery range in a single instance shall not exceed 5 points, until the score is restored to 100 points; recovery is only performed on the premise that the node is continuously in a healthy state (without any deduction conditions triggered) and no new deduction events occur within 3 consecutive recovery cycles; the starting time of the recovery cycle count is the first complete 10-second cycle after the node has eliminated all the previous deduction conditions and entered a no-deduction state; if any deduction condition is triggered again in any recovery cycle, the accumulated number of consecutive recovery cycles is cleared to zero, and the count restarts after all deduction conditions are eliminated; Furthermore, the heartbeat receiving unit is used to receive heartbeat messages broadcast by other nodes in the distributed cluster, and generate and update the cluster member status table locally on the local node; if the term number in the received heartbeat message is greater than the current term number stored locally on the local node, then the current term number stored locally on the local node is updated to the term number in the heartbeat message, and the current role identifier of the local node is updated to slave node; if the synchronization sequence number in the received heartbeat message is greater than the current synchronization sequence number stored locally on the local node, then the node that sent the heartbeat message is requested to synchronize the flashing timing data. Specifically, in this embodiment, the heartbeat receiving unit is also used to detect the changing trend between the received health score and the historical score stored locally by the node: if the health score of a node continues to decline within three consecutive heartbeat cycles and the decline exceeds 20 points (20% of the full score of 100), then a health deterioration warning is marked in the node status table to assist the candidate node screening unit in prioritizing the exclusion of the node during the election; after receiving the broadcast heartbeat message, the node first checks whether the node identifier in the message is its own node identifier. If so, the message is discarded directly without any status table update to prevent the node from misprocessing the heartbeat message it sent. The timeout determination unit is used to record the number of times the current node fails to receive a heartbeat message from the current virtual main light node in the distributed cluster. When the reconnection detection status of the node in the cluster member status table stored locally by the current node is "in reconnection detection", the missing heartbeat is not counted in the failure count. When the cumulative failure count reaches the preset heartbeat threshold, the current virtual main light node is determined to be in an abnormal working state. The online status of the node in the cluster member status table stored locally by the current node is updated to offline, and a virtual main light failure event is generated. The virtual main light failure event is the direct signal that triggers the virtual main light election module to start fault re-election. In this embodiment, the preset heartbeat threshold is set to 3 times, which is based on the fault detection response time of 3 seconds divided by the fixed period of 1 second, i.e., 3 times. Specifically, the status monitoring module, through the cooperation of the heartbeat sending unit, the heartbeat receiving unit, and the timeout determination unit, realizes real-time monitoring of the online status and health of each node in the cluster through a periodic heartbeat broadcast and timeout accumulation determination mechanism. When the cumulative heartbeat loss of the current virtual main light node reaches the preset heartbeat threshold, it automatically generates a virtual main light failure event, providing a trigger signal for automatic election after the main light failure. Specifically, the missing heartbeats in the reconnection detection of the timeout determination unit are not counted as failures. This embodiment demonstrates this through a specific scenario as follows: Suppose the current virtual master light node is marked as being in reconnection detection due to a network interruption. During this period, the slave nodes do not receive heartbeat messages from the master light five times consecutively. However, since the master light's reconnection detection status in the cluster member status table is "in reconnection detection," these five missing heartbeats are not counted as failures and do not trigger an election. Only when the master light's reconnection detection status is cleared (i.e., becomes not in reconnection detection) or is forcibly cleared after a timeout, will subsequent missing heartbeats begin to be counted as failures. This design is based on the fault boundary isolation principle: a node in reconnection detection is in the recovery attempt phase, which may be a temporary communication interruption rather than a permanent failure. Counting its missing heartbeats as failures at this time would trigger unnecessary elections, leading to frequent master light switching. Furthermore, based on the virtual main light failure event generated by the aforementioned status monitoring module, the virtual main light election module initiates a fault re-election process. After detecting that the current virtual main light is in an abnormal working state, the module first performs a pre-verification to confirm that the cluster has the conditions for election, and then elects a new virtual main light from healthy nodes based on the distributed consensus mechanism. The specific pre-verification process is as follows: Read the cluster member status table stored locally on this node, and traverse the reconnection probe status field corresponding to all nodes in the table; If any node is in the reconnection probe state and its duration does not exceed the preset absolute timeout threshold (30 seconds in this embodiment), then it is determined that there is a valid reconnection probe state in the cluster, and the election process is paused, waiting for the state to be cleared or forcibly cleared after timeout before being retried. If all nodes are in a state of not being in reconnection probe, or are in a state of reconnection probe but have exceeded the preset absolute timeout threshold, then the cluster is considered to have no valid reconnection probe state and is allowed to enter the election process. After the pre-verification is passed, the cluster member status table is read and healthy nodes are filtered. Based on the distributed consensus mechanism, a unique virtual main light is elected from among the healthy nodes. The virtual main light is a logical control role that issues synchronous blinking commands. This module includes a candidate node filtering unit, an election initiation unit, a voting adjudication unit, a role switching unit, and an initial election unit, which are respectively responsible for candidate set generation, election initiation, voting adjudication, role switching, and initial election upon power-on. The candidate node filtering unit is used to read the cluster member status table, filter nodes whose online status is normal and whose health score is higher than the preset health threshold (set to 60 points in this embodiment, based on the same majority fault superposition exclusion criterion as mentioned above, nodes are excluded only if the cumulative deduction exceeds 40 points, to ensure that normal nodes are not misjudged), and generate a local election candidate set; the specific meaning of the online status being normal is: the online status field value of the node in the cluster member status table is online, and the difference between the timestamp of the last heartbeat message received from the node and the current time is less than 3 times the fixed period (i.e., 3 seconds), so as to exclude the situation where the status table shows online but there has been no heartbeat for a long time in reality; The election initiation unit is used to: first complete two pre-state checks (detecting an abnormal working state of the current virtual main light and confirming that there are no nodes in the cluster with a valid reconnection probe state); after both conditions are met, start a preset random backoff timer; and after the preset random backoff timer expires, broadcast an election voting request to the nodes in the local election candidate set; the election voting request includes the node identifier, current role identifier, and health score of the node; in this embodiment, the election voting request also carries the term number and synchronization sequence number; Specifically, in this embodiment, the timeout value of the preset random backoff timer ranges from 0 to 1000 milliseconds (i.e., the preset random backoff time). Its algorithm is based on the heartbeat cycle benchmark method, using the fixed cycle of 1 second as a benchmark. The upper limit of the backoff time is set to 1 times the fixed cycle, ensuring that all nodes complete backoff and begin election within a maximum of 1 second, avoiding election delays exceeding the tolerable fault recovery time of the aviation obstruction light system due to excessively long backoff times; the lower limit is 0 milliseconds, allowing some nodes to immediately initiate elections to accelerate the recovery process; the random number generation algorithm uses the linear congruent method, based on the heartbeat cycle benchmark method. The lower 16 bits of the node's MAC address are used as a seed to ensure that the backoff time of each node is differentiated, reducing the probability of election conflicts. If multiple nodes initiate elections simultaneously, resulting in scattered voting and failure to elect a leader, each node will wait for a longer random backoff time in the next election round. The upper limit of the backoff time is automatically adjusted to 1.5 times that of the previous round, with a maximum of 3000 milliseconds. This upper limit is based on the three-backoff doubling convergence criterion. If a leader cannot be elected after a maximum of 3 rounds of backoff doubling, it is determined that the cluster has a serious network partition or a failure of half of the nodes. An election is forced after 3000 milliseconds to prevent infinite delay. It should be noted that this adaptive backoff expansion mechanism is only applicable to fault re-election after the master light fails during cluster operation; the backoff time doubling logic is not enabled during the initial power-on election of the cluster. Before initiating an election, the election initiating unit increments its current term number by 1 and temporarily switches its role identifier to candidate status. The point at which the number of currently online healthy nodes is counted is the instant the election initiating unit starts the preset random backoff timer. This value is fixed throughout the election and is used to determine the majority. Changes in node status during the voting process do not modify the vote counting criteria. Nodes initiating elections vote for themselves during the election process. Their voting criteria are: their online status is online, their reconnection detection status is not in reconnection detection, and their health score is not lower than the preset health threshold. When all three conditions are met, the node broadcasts a vote request to other nodes and also includes itself in the vote confirmation response count. The voting adjudication unit is used to receive election voting requests broadcast by other nodes in the distributed cluster, perform three-layer legal basic serial verification, and add an additional layer of preferred extended verification in this embodiment (one vote per term to prevent split-brain). The complete verification rules are as follows: Basic verification: If the initiating node is in an offline state in the status table, a rejection response will be returned directly. If the initiating node's reconnection probe status is "in progress", return a rejection response directly. If the health score of the initiating node is lower than the preset health threshold (60 points), a rejection response will be returned directly. This embodiment preferably uses extended verification: Check if the local node has already returned a vote confirmation response to other nodes within the current term number. If it has, return a rejection response directly to ensure that each node only casts one vote in each term, effectively preventing cluster split-brain. The three basic verification layers and the one preferred extended verification layer are executed sequentially. If any condition is triggered, the verification will terminate and a rejection response will be returned. Only after all verifications pass can a voting confirmation response be returned. This embodiment also adds two additional verification layers: verifying whether the term number of the initiating node is lower than the current term number stored locally on this node. If it is lower, a rejection response will be returned. Verifying whether the synchronization sequence number of the initiating node is lower than the current synchronization sequence number stored locally on this node will also be returned. The two additional verification layers do not change the above basic voting judgment logic. The role switching unit is used to update the current role identifier of the current node to the virtual master light when the cumulative number of received vote confirmation responses exceeds half of the number of currently online healthy nodes in the distributed cluster, and broadcasts the election result notification to the remaining nodes in the distributed cluster. In this embodiment, the calculation rule for the vote threshold exceeding half is as follows: if the number of currently online healthy nodes is N, the minimum number of votes required is N divided by 2 and rounded up by 1, i.e., floor(N / 2)+1. If the number of online healthy nodes in the cluster is even and a network partition occurs in a 2-to-2 manner, then either side can only obtain a maximum of 2 votes, which does not meet the requirement of exceeding half, thereby preventing split-brain. If the election deadlock is caused by network partition, each node re-initiates the election after the backoff timer expires. If the election fails for 3 consecutive rounds, the cluster split detection is triggered, and each node actively doubles the upper limit of the backoff time until the network recovers and converges again. The election result notification includes the node identifier of the new master light, the current term number, and the synchronization sequence number. After receiving the notification, the remaining nodes update the role identifier of the corresponding node in their local cluster member status table to the master light. The initialization election unit is used to read the local cluster member status table after the distributed cluster is powered on for the first time and the node network is completed. When it is detected that there is no valid virtual master light role configured in the cluster, no running virtual master light node, and no node in the reconnection detection state, each node waits for the preset random backoff time (i.e., the timeout value of the preset random backoff timer ranges from 0 to 1000 milliseconds), filters nodes with local health that meet the standard (not lower than the preset health threshold of 60 points) and are online normally to generate a candidate set, broadcasts an election voting request carrying its own node identifier, role identifier, and health score to the other nodes in the cluster, and elects the first unique virtual master light node in the cluster by mutual voting verification between nodes and adjudication based on more than half of the votes, thus completing the construction of the initial synchronization control architecture of the cluster. The backoff time doubling logic is not enabled during the initial election when the cluster is powered on for the first time. Specifically, the virtual main light election module, through the coordinated cooperation of the candidate node screening unit, election initiation unit, voting adjudication unit, role switching unit, and initialization election unit, and through the preset health threshold screening, random backoff to prevent conflict, three-layer basic verification and one-layer preferred extension verification voting adjudication and more than half confirmation mechanism, realizes the dynamic decoupling of the virtual main light role and physical nodes. When the main light fails, the election and switching of the new main light can be completed automatically without manual intervention. Among them, the preferred extension verification ensures that each node casts only one vote in each term, effectively preventing cluster split brain. The triggering conditions for initialization election and fault re-election are distinguished as follows: Initialization election is only executed once when the cluster is powered on for the first time, the network is completed, there is no main light, and there is no reconnection probe; Fault re-election is executed when the cluster is running, the main light is abnormal, and there is no effective reconnection probe; the two scenarios do not overlap, initialization election is only executed once when there has never been a main light, and subsequent main light failures are all classified as fault re-election. Furthermore, the synchronization control module is activated after the virtual master light election is completed and the valid virtual master light node is determined. It is used to broadcast a synchronization blinking instruction from the elected virtual master light node to the other nodes in the cluster and control all nodes to perform light blinking according to the same timing sequence. The synchronization control module includes a timing reference generation unit, a blinking parameter configuration unit, an instruction broadcasting unit, and a slave node synchronization execution unit. The timing reference generation unit is used to generate a unified synchronization timing reference signal for the cluster after the virtual master light election is completed and the valid virtual master light node is determined, using the local clock of the current virtual master light node as the cluster synchronization timing reference. The synchronization timing reference signal includes at least a time reference and a pulse counting reference. As a preferred synchronization method in this embodiment, since each node has integrated a GNSS timing module, the GNSS second pulse is used to calibrate the local clock of the virtual master light node. The selected virtual master light node and all slave nodes can directly use the second pulse output by GNSS as the absolute time reference, so that each node in the cluster can achieve synchronization. The node naturally possesses a globally unified time source, with synchronization accuracy reaching the microsecond level. As another implementation method, the system can automatically downgrade to using the local crystal oscillator clock of the virtual master light node as a reference, generating a synchronization reference pulse with a period of 1 second through a frequency divider. Each rising edge of the pulse corresponds to a timing beat. The timing beat counter is initially set to 0, increments by 1 after each pulse cycle, and wraps back to 0 to start counting again when the counter value reaches 86400. All slave nodes calibrate their own local pulse counters based on the received timing beat counter value. The above two methods can be dynamically switched according to the GNSS signal quality, and this embodiment does not limit this to a single method. Specifically, in this embodiment, the GNSS signal quality determination and the switching mechanism between the two synchronization modes are as follows: GNSS signal loss judgment criteria: GNSS signal loss is judged when any of the following conditions are met: no valid GNSS second pulse signal is received for 3 consecutive seconds; the strength of the received GNSS signal is continuously lower than -130dBm for more than 5 seconds; the above thresholds are set according to the nominal sensitivity parameters of the GNSS receiving module, and -130dBm is the lower limit of the tracking sensitivity of most industrial-grade GNSS modules. GNSS local clock degradation switching: After the lock is lost, the system automatically switches to the local crystal oscillator clock as the timing reference; a phase continuous transition mechanism is adopted during the switching, and the current difference between the local crystal oscillator counter and the GNSS second pulse counter is recorded at the moment of switching. This difference is used as the initial phase offset of the local crystal oscillator mode to ensure that the timing tick counter does not jump before and after the switching. Local clock GNSS recovery switching: When the GNSS signal is restored (a valid second pulse is received for 10 consecutive seconds and the signal strength is not lower than -130dBm), the system automatically switches back from local clock mode to GNSS mode; the recovery switching also adopts a phase continuous transition mechanism, using the counter difference at the moment of switching as the offset to avoid sudden timing changes; the 10-second recovery confirmation time is based on the majority filtering criteria - 10 consecutive seconds of valid signal can eliminate false recovery caused by brief interference; Furthermore, the flashing parameter configuration unit is used to uniformly broadcast preset flashing mode parameters to all slave nodes in the distributed cluster through the current virtual master light node. The preset flashing mode parameters include flashing period, duty cycle, and flash count. These parameters are configurable and can be flexibly set according to aviation obstruction light standards and on-site control requirements, and are not limited to specific values. In this embodiment, the flashing period is set to 1 second, which is based on the industry standard requirements for aviation obstruction light flashing frequency. A 1-second period corresponds to 60 flashes per minute, covering the commonly used flashing frequency range of 20 to 60 flashes per minute. The duty cycle is set to 50%, which is based on the effective light emission duration requirements of aviation obstruction light intensity. The flash count is set to continuous flashing with no fixed number of flashes, and is terminated when a stop command is issued by the ground monitoring terminal. The preset flashing mode parameters can be dynamically adjusted by the ground monitoring terminal during operation, and after adjustment, the virtual master light node rebroadcasts them across the entire network. The instruction broadcasting unit is used to broadcast a synchronization blinking instruction carrying the synchronization timing reference signal and the preset blinking mode parameters to all slave nodes in the distributed cluster after the cluster synchronization timing reference signal is generated and the preset blinking mode parameters are configured. To ensure that the instruction arrives reliably, this embodiment can adopt an acknowledgment and retransmission mechanism. After the master light broadcasts, it waits for the acknowledgment response from each slave node. Slave nodes that do not receive an acknowledgment response will be recorded and retransmission will be triggered. The maximum number of retransmissions is set to 3, for example, and the retransmission interval is set to 500 milliseconds. The slave node synchronization execution unit is deployed on each slave node within the cluster. It receives the synchronization flashing command issued by the virtual master light node, calibrates the local clock of the node according to the synchronization timing reference signal, and drives the local aviation obstruction light to complete the synchronization flashing action strictly according to the preset flashing mode parameters. In the preferred method of using GNSS second pulse as reference, the slave node directly locks to the GNSS second pulse to achieve high-precision synchronization. When downgrading to software synchronization mode, the slave node calculates the local clock deviation by comparing the deviation between the local timing beat and the reference issued by the master light, and gradually adjusts the frequency division coefficient of the local crystal oscillator using a proportional-integral adjustment algorithm, so that the local timing beat counter gradually converges to the master light reference in multiple cycles. When the deviation is less than 5 milliseconds, synchronization is considered to have been achieved, and the current frequency division coefficient is locked until the next synchronization calibration. The 5-millisecond setting is based on the visual criterion for synchronous flashing of aviation obstruction lights. The human eye cannot distinguish a flashing time difference of less than 50 milliseconds, and 5 milliseconds is one-tenth of 50 milliseconds, leaving sufficient margin. Specifically, in this embodiment, the specific parameters and convergence criteria of the proportional-integral adjustment algorithm are as follows: The proportional gain is 0.5, determined based on the optimal proportional gain corresponding to a typical second-order system damping ratio of 0.707; the integral gain is 0.1, taken as 5 times the proportional gain based on the integral time constant, to ensure the convergence of steady-state error. The adjustment cycle is one adjustment operation triggered every time a synchronization reference pulse is received, that is, once every 1 second, to keep in sync with the heartbeat cycle; The convergence determination condition is as follows: after three consecutive adjustment operations, the absolute value of the deviation between the local timing clock counter and the main lamp timing clock counter is less than 5 milliseconds. At this time, it is determined that convergence has been achieved and the current division coefficient is locked. If the deviation exceeds 10 milliseconds again after locking (i.e., loss of synchronization), the lock is automatically released and PI adjustment is restarted. The above-mentioned 10-millisecond loss of synchronization determination threshold is set to twice the synchronization accuracy threshold of 5 milliseconds, leaving sufficient hysteresis margin to avoid frequent locking and unlocking due to instantaneous fluctuations. The slave node synchronous execution unit does not perform light flashing during calibration, that is, the LED light group remains in the off state; it only starts flashing according to the main light sequence after calibration is completed; the flashing interruption time during calibration shall not exceed 3 seconds, in accordance with the operational requirement that aviation obstruction lights shall not be continuously extinguished for more than 3 seconds under low visibility conditions; Specifically, the synchronization control module, through the coordinated operation of the timing reference generation unit, the blinking parameter configuration unit, the instruction broadcasting unit, and the slave node synchronization execution unit, generates timing signals based on the local clock of the virtual main light, uniformly configures the preset blinking mode parameters, and issues synchronization instructions through a reliable broadcasting mechanism. Each slave node calibrates its local clock according to the timing reference signal and then executes light blinking, thus realizing that all nodes in the cluster execute light blinking according to the same timing. Furthermore, during the normal operation of the aforementioned synchronization control module, if the original virtual main light goes offline due to a fault and then comes back online, the node recovery module is used to enable the original virtual main light node to rejoin the cluster as a slave node when there is already a valid virtual main light in the distributed cluster, without triggering a new election; the node recovery module includes a recovery detection unit, an identity confirmation unit, and a role synchronization unit. The recovery detection unit continuously monitors the message broadcast status of the offline original virtual master node. When it detects that the node has resumed periodic broadcasting heartbeat messages and ended the reconnection probe state, the determination method is that the reconnection probe status field of the node in the cluster member status table is updated to not be in reconnection probe, the node is determined to be online again, and a node recovery event is generated to notify other nodes in the distributed cluster. The recovery detection unit also verifies whether the node identifier of the recovered node is consistent with the node identifier of the original virtual master node, and whether the current term number of the node is lower than or equal to the term number of the current master. If the term number of the recovered node is higher than that of the current master, according to the principle of monotonically increasing term number, it means that the node may have incremented its term number and attempted to initiate an election during the partitioning period. At this time, the current master should actively synchronize the higher term number and perform demotion, and then invite the node to join as a slave node. The identity verification unit is used to read the cluster member status table stored locally on the node and verify whether there is a valid virtual master light node in the distributed cluster. If there is already a valid virtual master light node in the cluster, the recovery node is forcibly locked and configured as a slave node, without triggering or starting a new round of virtual master light election. The identity verification unit verifies whether there is a valid master light in the cluster by sending a confirmation request to the current virtual master light node. If the current master light replies with confirmation within a preset response time (in this embodiment, the preset response time is 1 second), it is determined that there is a valid master light in the cluster, and the recovery node performs a downgrade operation. If there is no reply after the confirmation request is sent to the current virtual master light, it is determined that there is no valid virtual master light in the cluster. The cluster-wide no-master-light fault scenario is not handled by the identity verification unit, but is uniformly handled by the initialization election unit or fault re-election process of the virtual master light election module. The role synchronization unit is used to broadcast a node role change notification to the remaining nodes in the distributed cluster after the role identifier of the recovery node is updated to that of a slave node. This allows the remaining nodes in the cluster to update the online status, role identifier, and reconnection detection status of the node in the cluster member status table stored locally. The role change notification includes the latest role identifier, latest term number, latest synchronization sequence number, and current online status of the recovery node. After receiving the notification, the remaining nodes update the corresponding fields in their local cluster member status table. Specifically, the node recovery module, through the coordinated operation of the recovery detection unit, the identity verification unit, and the role synchronization unit, enables the fault-recovered node to rejoin the cluster as a slave node without triggering a new round of election, thus ensuring the stable operation of the cluster after the node fault is recovered.

[0012] Example 2, please refer to Figure 2 This embodiment provides a method for synchronous control of an aviation obstruction light cluster with a virtual master light self-election mechanism. Compared with Embodiment 1, the core difference of this embodiment is that it focuses on describing the complete method flow for realizing virtual master light election and cluster synchronization control using this system. The method includes the following steps: S1. After each node powers on and starts up, it sends probe messages to the preset coverage area via UDP broadcast protocol at the preset period to detect and discover other online nodes in the cluster; based on the node identifier and communication address obtained from the probe response, it sends a link establishment request to the target node, and establishes a bidirectional peer-to-peer communication link after receiving the confirmation response from the other end; after all nodes complete the link establishment, the initial role of all nodes is uniformly configured as slave nodes, and the initial term number and synchronization sequence number are uniformly set to preset initial values ​​(such as 0), thus completing the cluster networking. Specifically, the preset period and the number of consecutive loss judgments can be flexibly configured according to the number of nodes and the network environment; after the link establishment timeout, the preset number of requests will be retried. After the preset number of requests is exceeded, the link establishment operation will be terminated and the target node will be recorded as an unreachable node, and its reconnection detection status will be marked as reconnection detection in progress; after the link is established, a two-way heartbeat keep-alive mechanism will be maintained between the local node and the target node. If no heartbeat confirmation is received for several consecutive times, the link will be determined to be interrupted and the reconnection process will be automatically triggered; the reconnection detection status has an absolute timeout forced clearing mechanism. When the reconnection detection status of any node lasts for more than the preset absolute timeout threshold (30 seconds in this embodiment), it will be forcibly cleared and updated to offline. S2. After the cluster network is completed, each node confirms that there is no valid virtual main light role in the cluster and no node is in the reconnection detection state according to the local cluster member status table. Then, it waits for a preset random backoff time (i.e., the timeout value of the preset random backoff timer ranges from 0 to 1000 milliseconds). Through mutual voting verification between nodes, after obtaining confirmation from more than half of the online healthy nodes, the first unique virtual main light node in the cluster is elected. Specifically, candidate nodes are selected based on their online status being normal and their health score being no lower than the preset health threshold of 60 points. Each node's voting decision unit performs three-layer basic verification based on its online status, reconnection detection status, and health score. A voting confirmation response is returned only when the initiating node is online, not in reconnection detection status, and has a health score no lower than the preset health threshold. Each node casts only one vote per term. The vote threshold is more than half of the current number of online healthy nodes (floor(N / 2)+1) to ensure that at most one node meets the election criteria at any given time. S3. During cluster operation, each node periodically broadcasts heartbeat messages at the fixed period (1 second in this embodiment) and updates the local cluster member status table in real time based on the received heartbeat messages; if the current virtual main light node is in the reconnection detection state, the missing heartbeat of the node is not counted in the failure count; when the cumulative number of missing heartbeats reaches the preset heartbeat threshold (3 times in this embodiment), it is determined that the current virtual main light node is in an abnormal working state and a virtual main light failure event is generated; Specifically, the health score adopts a weighted deduction system. Nodes with a score below the preset health threshold of 60 are considered unhealthy nodes and will not participate in the election of candidates and voting confirmation. S4. When an abnormal working status of the current virtual main light node is detected, a pre-verification is performed, which involves reading the cluster member status table stored locally on this node, traversing the reconnection probe status field corresponding to all nodes in the table, and confirming that there are no nodes with a valid reconnection probe status (i.e., all nodes are either not in reconnection probe or are in reconnection probe but have exceeded the preset absolute timeout threshold and have been forcibly cleared). After the pre-verification passes, wait for the preset random backoff timer to expire, and then initiate an election voting request. The voting adjudication unit of each node performs three layers of basic verification and one layer of preferred extended anti-duplicate voting verification based on the online status, reconnection probe status, and health score. After obtaining confirmation from more than half of the online healthy nodes, the corresponding node is switched to the new virtual main light, and the election results are broadcast to the entire cluster network. Specifically, if multiple nodes initiate elections simultaneously, resulting in scattered voting and failure to elect a master light, then each node will automatically adjust its backoff time limit to 1.5 times that of the previous round in the next election, with a maximum of 3000 milliseconds; the election result notification includes the node identifier of the new master light, the current term number, and the synchronization sequence number, and the other nodes will update their local cluster member status table after receiving the notification. S5. The selected virtual master light node generates a synchronization timing reference signal using its local clock as the cluster synchronization timing reference (preferably using the GNSS second pulse to calibrate the local clock as the reference; if the GNSS is lost, it is downgraded to the local crystal oscillator clock reference), configures unified preset flashing mode parameters, and broadcasts the synchronization flashing command carrying the synchronization timing reference signal and the preset flashing mode parameters to all slave nodes in the cluster; after receiving the command, each slave node calibrates its local clock according to the synchronization timing reference signal and performs the light synchronization flashing action according to the preset flashing mode parameters; Specifically, the synchronous flashing command adopts a confirmation and retransmission mechanism (the retransmission limit is set to 3 times, and the interval is set to 500 milliseconds) to ensure reliable arrival; in GNSS second pulse synchronization mode, the slave node directly locks to GNSS to achieve microsecond-level synchronization; in software synchronization mode, a proportional-integral adjustment algorithm is used to bring the deviation to a tolerance range of 5 milliseconds; the flashing action can be temporarily suspended during calibration, and normal flashing will begin according to the main light sequence after calibration is completed. S6. Fault Node Recovery: When it is detected that the original virtual master node has resumed heartbeat broadcasting, ended the reconnection probe state, and is stably online, if there is a valid virtual master node in the cluster, the recovered node will be locked and configured as a slave node, and a new round of cluster election will not be triggered; at the same time, a node role change notification will be broadcast to the entire network, so that the other nodes will update the online status, role identifier, and reconnection probe status of the corresponding node in the local cluster member status table. Specifically, if the term number of the recovery node is higher than that of the current master node, the current master node will actively synchronize with the higher term number and be downgraded to a slave node, and then invite the node to join as a slave node to ensure that the term number remains monotonically increasing throughout the cluster.

[0013] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A synchronous control system for aviation obstruction lights clusters with a virtual master light self-election mechanism, characterized in that, include: The cluster building module is used to build a distributed cluster of multiple aviation obstruction light nodes, with each node communicating with the others in a peer-to-peer manner. The status monitoring module is used for each node to periodically broadcast heartbeat messages, receive heartbeat messages broadcast by other nodes, and monitor the online status and health score of each node in real time, and generate and maintain a cluster member status table; the cluster member status table includes at least the node identifier, current role identifier, online status, health score and node reconnection detection status of each node; the reconnection detection status is used to indicate whether the node is in the process of link reconnection detection. The virtual main light election module is used to read the cluster member status table and filter healthy nodes in the cluster. If the current virtual main light is in an abnormal working state and there are no nodes in the reconnection detection state in the cluster, a unique virtual main light is elected from the healthy nodes in the cluster based on a distributed consensus mechanism. The virtual main light is a logical control role that issues synchronous blinking commands. The abnormal working state is defined as the current virtual main light node's cumulative heartbeat loss reaching a preset heartbeat threshold. The synchronization control module is used to broadcast a synchronization blinking command from the selected virtual master light node to the other nodes in the cluster, and control all nodes to perform light blinking in the same timing sequence. The node recovery module is used to enable the original virtual master node to rejoin the cluster as a slave node when a valid virtual master already exists in the distributed cluster, without triggering a new election.

2. The aviation obstruction light cluster synchronization control system with a virtual master light self-election mechanism according to claim 1, characterized in that, The cluster construction module includes: The node discovery unit is used to periodically broadcast detection messages after each node is powered on and receive response messages from other nodes, thereby collecting the node identifiers and communication addresses of the currently online aviation obstruction light nodes within the preset coverage area. The link establishment unit is used to send a link establishment request to the target node based on the node identifier and communication address; if a confirmation response is received from the peer, a bidirectional peer-to-peer communication link is established between the local node and the target node, and the cluster communication frequency, data transmission rate and data encryption rules corresponding to the communication link are determined; if the request times out without a response, the link establishment request is repeatedly initiated, and after retries exceeding a preset number of requests, the current link establishment operation is terminated and the target node is recorded as currently unreachable, and the reconnection detection status of the target node in the cluster member status table stored locally by the local node is marked as reconnection detection in progress; The role initialization unit is used to generate an initial role identifier locally on the node after the communication link is established, and to configure the initial role of the node as a slave node.

3. A synchronous control system for an aviation obstruction light cluster with a virtual master light self-election mechanism according to claim 2, characterized in that, The link establishment unit also includes: The status detection subunit is used to read the currently unreachable node information recorded locally by this node and continuously listen to whether the target node rebroadcasts the probe message within a preset time window; The reconnection trigger subunit is used to resend a link establishment request to the target node when the status detection subunit receives a probe message broadcast by the target node within a preset time window, and restore the bidirectional peer-to-peer communication link with the target node after receiving the confirmation response from the peer, remove the unreachable mark of the target node recorded locally by this node, update the online status of the node in the cluster member status table stored locally by this node to online, and update the reconnection detection status of the node to not being in reconnection detection. The information update subunit is used to update the currently unreachable node information to offline node information and clear the reconnection probe status mark of the node in the cluster member status table when the status detection subunit has not received the probe message broadcast by the target node within a preset time window.

4. A synchronous control system for an aviation obstruction light cluster with a virtual master light self-election mechanism according to claim 1, characterized in that, The status monitoring module includes: The heartbeat sending unit is used by each node to broadcast heartbeat messages to other nodes in the distributed cluster at fixed intervals. The heartbeat message includes the node identifier, current role identifier, timestamp, and node health score of the node. The heartbeat receiving unit is used to receive heartbeat messages broadcast by other nodes in the distributed cluster, and generate and update the cluster member status table locally on the local node. The timeout determination unit is used to record the number of times the current virtual main light node in the distributed cluster fails to receive a heartbeat message consecutively. When the reconnection detection status of the node in the cluster member status table stored locally by the current node is "in reconnection detection", the missing heartbeat is not counted in the failure count. When the cumulative failure count reaches a preset heartbeat threshold, the current virtual main light node is determined to be in an abnormal working state. The online status of the node in the cluster member status table stored locally by the current node is updated to offline, and a virtual main light failure event is generated.

5. A synchronous control system for an aviation obstruction light cluster with a virtual master light self-election mechanism according to claim 1, characterized in that, The virtual main light election module includes: The candidate node filtering unit is used to read the cluster member status table, filter the nodes whose online status is normal and whose health score is higher than the preset health threshold, and generate a local election candidate set. An election initiation unit is used to broadcast an election voting request to the nodes in the local election candidate set after detecting that the current virtual main light is in an abnormal working state, that there are no nodes in the reconnection probe state in the cluster member status table stored locally by this node, and that the preset random backoff timer has expired. The election voting request includes the node identifier, current role identifier, and health score of this node. The voting adjudication unit is used to receive the broadcast election voting request, verify the health and online status of the initiating node according to the cluster member status table stored locally on the node, and return a voting confirmation response if the verification passes, otherwise return a rejection response. The role switching unit is used to update the current role identifier of the node to virtual master lamp and broadcast the election result notification to the other nodes in the distributed cluster when the cumulative number of received vote confirmation responses exceeds half of the number of currently online healthy nodes in the distributed cluster.

6. A synchronous control system for an aviation obstruction light cluster with a virtual master light self-election mechanism according to claim 5, characterized in that, The step of verifying the health and online status of the initiating node based on the cluster member status table stored locally on this node includes: If the online status of the initiating node in the cluster member status table stored locally on this node is offline, then a rejection response is returned directly. If the reconnection probe status in the cluster member status table stored locally by the initiating node is "in reconnection probe", then a rejection response is returned directly. If the health score of the initiating node in the cluster member status table stored locally on this node is lower than the preset health threshold, a rejection response is returned. If the initiating node has an online status in the cluster member status table stored locally on this node, a reconnection probe status of not being in reconnection probe, and a health score not lower than a preset health threshold, then a voting confirmation response is returned.

7. A synchronous control system for an aviation obstruction light cluster with a virtual master light self-election mechanism according to claim 1, characterized in that, The synchronization control module includes: The timing reference generation unit is used to generate a unified synchronization timing reference signal for the cluster after the virtual main light election is completed and the valid virtual main light node is determined, using the local clock of the current virtual main light node as the cluster synchronization timing reference. The synchronization timing reference signal includes at least a time reference and a pulse counting reference. The blinking parameter configuration unit is used to broadcast preset blinking mode parameters to all slave nodes in the distributed cluster through the current virtual master light node. The blinking mode parameters include blinking period, duty cycle and blink count. The instruction broadcasting unit is used to broadcast the synchronization blinking instruction, which carries the synchronization timing reference signal and blinking mode parameters, to all slave nodes in the distributed cluster after the cluster synchronization timing reference signal is generated and the blinking mode parameters are configured. The slave node synchronization execution unit is deployed on each slave node in the cluster. It is used to receive the synchronization flashing command issued by the virtual master light node, calibrate the local clock of the node according to the synchronization timing reference signal, and drive the local aviation obstruction light to complete the synchronization flashing action strictly according to the flashing mode parameters.

8. A synchronous control system for an aviation obstruction light cluster with a virtual master light self-election mechanism according to claim 1, characterized in that, The node recovery module includes: The recovery detection unit is used to continuously monitor the message broadcast status of the offline original virtual master light node. When it is detected that the node resumes periodic broadcast heartbeat messages and ends the reconnection probe status, it is determined that the node has recovered online and a node recovery event is generated to notify other nodes in the distributed cluster. The identity verification unit is used to read the cluster member status table stored locally on this node and verify whether there is a valid virtual master node in the distributed cluster. If there is already a valid virtual master node in the cluster, the recovery node is forcibly locked and configured as a slave node, without triggering or starting a new round of virtual master election in the cluster. The role synchronization unit is used to broadcast a node role change notification to the remaining nodes in the distributed cluster after the role identifier of the recovered node is updated to that of a slave node, so that the remaining nodes in the cluster update the online status, role identifier, and reconnection detection status of the node in the cluster member status table stored locally.

9. A synchronous control system for an aviation obstruction light cluster with a virtual master light self-election mechanism according to claim 1, characterized in that, The virtual main light election module also includes: An initialization election unit is used to allow each node to read its local cluster member status table after the distributed cluster is powered on for the first time and the node network is completed. When it is detected that there is no valid virtual master light role configured in the cluster, no running virtual master light node, and no node in the reconnection detection state, each node waits for a preset random backoff time, filters nodes with qualified local health and normal online status to generate a candidate set, and broadcasts an election voting request carrying its own node identifier, role identifier, and health score to the other nodes in the cluster. The first and only virtual master light node in the cluster is generated by mutual voting verification between nodes and the election is based on a majority vote, thus completing the construction of the initial synchronization control architecture of the cluster.

10. A method for synchronous control of an aviation obstruction light cluster with a virtual master light self-election mechanism, applied to the system described in any one of claims 1-9, characterized in that, Includes the following steps: S1. After each node powers on and starts up, it broadcasts a probe message to detect and discover online nodes in the cluster, establishes a two-way peer-to-peer communication link to complete the cluster networking, and uniformly configures the initial role of all nodes as slave nodes. S2. After the cluster network is completed, each node confirms that there is no valid virtual main light in the cluster and no node is in the reconnection detection state according to the local cluster member status table. Then, it waits for a preset random backoff time and verifies through mutual voting among nodes. After obtaining confirmation from more than half of the online healthy nodes, the first and only virtual main light node in the cluster is elected. S3. During cluster operation, each node periodically broadcasts heartbeat messages and updates its local cluster member status table in real time based on the received heartbeat messages. If the current virtual main light node is in the reconnection detection state, the missing heartbeat of the node will not be counted in the failure count; When the cumulative number of missing heartbeats reaches the preset heartbeat threshold, the current virtual main light node is determined to be in an abnormal working state. S4. When an abnormal working status of the current virtual main light node is detected and no node in the reconnection probe state is in the local cluster member status table, an election voting request is initiated after the random backoff timer expires. Each node completes the voting verification based on its online status, reconnection probe status, and health score. After obtaining confirmation from more than half of the online healthy nodes, the corresponding node is switched to the new virtual main light and the election results are broadcast to the entire cluster network. S5. The selected virtual master light node generates a synchronization timing reference signal using its local clock as the cluster synchronization timing reference, configures unified flashing mode parameters, and broadcasts the synchronization flashing command carrying the synchronization timing reference signal and flashing mode parameters to all slave nodes in the cluster. After receiving the instruction, each slave node calibrates its local clock according to the synchronization timing reference signal and performs synchronized flashing of the lights according to the flashing mode parameters; S6. When it is detected that the original virtual master node has resumed heartbeat broadcasting, ended the reconnection probe state and is stably online, if there is a valid virtual master node in the cluster, the restored node will be locked and configured as a slave node, and a new round of cluster election will not be triggered; at the same time, a node role change notification will be broadcast to the entire network, so that the other nodes will update the online status, role identifier and reconnection probe status of the corresponding node in the local cluster member status table.

Images