LED street lamp alarm information generation method and system

Abstract

The invention discloses an LED street lamp alarm information generation method and system, and the method comprises the steps: carrying out the cluster division of all LED street lamps in a set region based on a preset clustering rule, and forming at least one street lamp cluster; running state data and environment sensing data collected by the multi-mode sensor of each LED street lamp in real time are obtained, and warning information corresponding to the current LED street lamp is generated based on the running state information and the environment sensing data; broadcasting the alarm information to other LED street lamps which do not generate the alarm information in the cluster through a preset cluster multicast address, so that the other LED street lamps which do not generate the alarm information receive the corresponding alarm information; and after the LED street lamps in the cluster receive the alarm information, analyzing the alarm information so as to match a preset power supply switching strategy according to the alarm information, and correspondingly supplying power to the mains supply or the storage battery according to the power supply switching strategy. The problems that in the prior art, an LED street lamp is poor in reliability and safety are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of LED street lamps, in particular to an LED street lamp alarm information generation method and system. BACKGROUND

[0002] With the continuous acceleration of urbanization, as a core component of urban public infrastructure, the lighting continuity, operation reliability and emergency disposal capability of street lamps are directly related to the safety of pedestrians and vehicles, and have an important influence on urban public safety and the quality of life of residents. However, the traditional LED street lamp and the existing part of the emergency street lamp still face many technical bottlenecks in actual application, and it is difficult to adapt to the needs of complex and variable urban environment and sudden conditions.

[0003] For example, in the aspect of sudden condition response, the traditional street lamp generally relies on single power supply, and when power failure, regional power failure and other situations occur, there is a lack of effective backup power supply mechanism, which is easy to cause lighting interruption, leading to darkness on the road, and bringing great safety hazards to night travel. At the same time, in the extreme weather such as rainstorm and flood, the street lamp poles in low-lying areas of the city are easy to be soaked by water, and the traditional street lamp lacks effective water level monitoring and protection means, and often causes short circuit fault due to immersion, which not only causes single lamp damage, but also may affect the normal operation of the surrounding street lamp system, further expand the fault range, and may cause electric shock, involving the safety of pedestrians. In addition, in the long-term operation process of street lamp, the lamp and control components are easy to cause temperature abnormity due to overload, poor heat dissipation and other reasons, and the traditional street lamp lacks real-time temperature monitoring and alarm mechanism, which may cause fire or permanent damage of components due to overheating, seriously threatening public safety.

[0004] Therefore, the traditional LED street lamp has the problems of poor reliability and safety in the face of many sudden situations. SUMMARY

[0005] Therefore, the present application aims to provide an LED street lamp alarm information generation method and system, which aims to solve the problems of poor reliability and safety of LED street lamp in the prior art.

[0006] The present application provides an LED street lamp alarm information generation method, which is applied to the scene of LED street lamp, each LED street lamp can communicate with each other, and each LED street lamp includes an LED light source arranged at the top of the lamp pole and a multi-modal sensor arranged on the lamp pole, the LED light source can be powered by connecting the power supply or connecting the battery arranged in the lamp pole, and the method comprises the following steps: Based on the preset clustering rule, all LED street lamps in the set area are divided into clusters to form at least one street lamp cluster; The system acquires real-time operating status data and environmental perception data from the multimodal sensors of each LED street light, and generates alarm information corresponding to the current LED street light based on the operating status information and environmental perception data. The alarm information is broadcast to other LED streetlights in the same cluster that have not generated alarm information through a preset cluster multicast address, so that the other LED streetlights that have not generated alarm information can receive the corresponding alarm information; When the LED streetlights in the cluster receive an alarm message, they analyze the alarm message to match the preset power supply switching strategy, and then switch to mains power or battery power according to the corresponding power supply switching strategy.

[0007] Furthermore, in the aforementioned method for generating LED street light alarm information, the step of dividing all LED street lights within a designated area into clusters based on preset clustering rules to form at least one street light cluster includes: Collect static basic data and dynamic operation data of each LED street light, and construct a three-dimensional feature set including geographical topology, equipment association, and operating environment characteristics; The feature set is Z-score standardized to eliminate dimensional differences. Core features with a correlation greater than the alarm threshold are selected by mutual information entropy. Redundant data is removed to form a feature matrix. Different weights are assigned to the geographical topology, device association, and operating environment features in the feature matrix, and the comprehensive feature vector of each street light is obtained by weighted summation. Based on the difference of the comprehensive feature vectors of each pair of streetlights, the comprehensive distance representing the correlation between each pair of streetlights is obtained, and the neighborhood radius of each streetlight is determined. If the number of streetlights in the neighborhood of a certain streetlight is greater than the threshold and the average device matching degree of these streetlights is greater than the preset value, then the streetlight is determined as the cluster center. Using the cluster center as the center, all streetlights that meet the condition of having a comprehensive distance less than the threshold are grouped into a cluster. For streetlights without qualified neighbors, the K-nearest neighbor algorithm is used to calculate the average distance between them and the surrounding clusters, and they are assigned to the cluster with the smallest distance. In the end, the clustering of all LED streetlights in the set area is realized.

[0008] Furthermore, in the above-mentioned method for generating LED street light alarm information, the step of generating the alarm information corresponding to the current LED street light based on operating status information and environmental perception data includes: Modal classification is performed on the acquired operational status data and environmental perception data. Differentiated denoising strategies are adopted for the noise characteristics of different modal data. The synchronization and alignment of the two types of data are completed based on timestamps, and a binary data matrix containing device status dimension and environmental dimension is constructed. The core feature vectors of the binary data matrix are extracted. After eliminating the difference in dimensions by Z-score standardization, the fusion weights of the two types of features are dynamically allocated by the environment adaptive weight formula. A feature interaction model is constructed through a cross-modal attention mechanism to focus on feature combinations that are strongly associated with equipment faults. The cosine similarity between the feature combination and the preset fault mode library is calculated, and candidate abnormal feature sets with similarity greater than the preset threshold are selected. A Bayesian network model is used to perform probabilistic reasoning on the candidate anomaly feature set to obtain the corresponding fault type and probability in order to generate alarm information. The prior probability of the network is trained based on historical fault data of the same type of LED street light.

[0009] Furthermore, in the aforementioned LED street light alarm information generation method, the step of broadcasting the alarm information to other LED street lights within the same cluster that have not generated alarm information via a preset cluster multicast address, so that these other LED street lights receive the corresponding alarm information, includes: According to preset rules, a leader street light is elected in the street light cluster, and alarm information is broadcast to the leader node in the cluster through a preset cluster multicast address. The leader node broadcasts alarm information to other LED streetlights in its cluster that have not generated alarm information via a preset cluster multicast address, so that the other LED streetlights that have not generated alarm information can receive the corresponding alarm information.

[0010] Furthermore, in the aforementioned method for generating LED street light alarm information, the step of electing a leader street light within the street light cluster according to preset rules includes: Obtain the coordinate information of each street light in the cluster, and map each street light to a preset coordinate system based on the coordinate information; A longitudinal baseline is formed by connecting the northernmost and southernmost streetlights within the cluster, and a transverse baseline is formed by connecting the easternmost and westernmost streetlights. The intersection of the two baselines is recorded as the geometric center of the cluster. Based on the geometric center of the cluster, the corresponding leader node determination area is determined, and one of the sub-regions of the leader node determination area divided by the baseline is selected as the target leader node determination area according to the preset rules. Draw a circle with the coordinates of each street light as the center and the standard communication radius of the street light model as the radius to obtain the communication coverage circle of each street light. Draw an auxiliary line through the geometric center in the area determined by the target leader node at a preset angle to the longitudinal baseline. Select the preset number of street lights with the most intersections between the auxiliary line and each communication coverage circle as candidate leader nodes. Taking the geometric center of the cluster as the origin, calculate the extreme radius of each candidate leader node, select the candidate node with the smallest extreme radius as the primary leader node, and then calculate the overlap area of ​​the communication coverage circles of the primary leader node and other candidate nodes. Select the node with the smallest overlap area and a distance from the primary leader node greater than a preset multiple of the communication radius as the backup leader node.

[0011] Furthermore, in the aforementioned LED street light alarm information generation method, the step of broadcasting the alarm information to the leader node within the cluster via a preset cluster multicast address includes: The LED street light that generates alarm information locates the main leader node through coordinate matching. If its own communication coverage circle overlaps with the coverage circle of the main leader node, it directly transmits the alarm information to the main leader node. If there is no overlap, other streetlights within its own coverage circle are identified as relay nodes, and the relay nodes relay the data to the main leader node. After receiving the alarm, the master leader node draws a circle with itself as the center, dividing the cluster into an inner circle and an outer circle. The streetlights in the inner circle directly receive the broadcast, while the streetlights in the outer circle are forwarded by the streetlights designated by the master leader node at the edge of the coverage circle, ensuring that alarm information is covered by the entire cluster.

[0012] Furthermore, in the aforementioned LED street light alarm information generation method, the step of selecting one sub-region from the sub-regions of the leader node determination area divided by the baseline as the target leader node determination area according to preset rules includes: The sub-region attribute dataset is obtained by collecting the number of valid streetlights and the coordinates of each streetlight within each sub-region and calculating the distribution concentration.

[0013] Based on the standard communication coverage radius of LED streetlights of the same model, the maximum number of nodes that a single light can theoretically directly cover is determined. Based on this, the optimal range of the number of effective streetlights in a sub-region is set to filter sub-regions where the number of effective streetlights is within the optimal range. At the same time, sub-regions with a distribution concentration greater than the threshold are retained as candidate sets. For the generated candidate set, the sub-region with the largest number of valid streetlights is selected first. If the number is the same, the distribution concentration is compared and the sub-region with higher distribution concentration is selected to finally determine the target sub-region.

[0014] Another objective of this invention is to provide an LED street light alarm information generation system, applicable to LED street light scenarios. Each LED street light can communicate with each other, and each LED street light includes an LED light source mounted on the top of the light pole and a multimodal sensor deployed on the light pole. The LED light source can be powered by connecting to mains power or by connecting to a battery installed inside the light pole. The system includes: The partitioning module is used to divide all LED streetlights in a set area into clusters based on preset clustering rules, forming at least one streetlight cluster. The acquisition module is used to acquire real-time operating status data and environmental perception data collected by the multimodal sensors of each LED street light, and generate alarm information corresponding to the current LED street light based on the operating status information and environmental perception data. The broadcast module is used to broadcast alarm information to other LED streetlights in the same cluster that have not generated alarm information through a preset cluster multicast address, so that other LED streetlights that have not generated alarm information can receive the corresponding alarm information; The switching module is used to analyze the alarm information when the LED streetlights in the cluster receive alarm information, and match the preset power supply switching strategy according to the alarm information, and switch to mains power or battery power according to the corresponding power supply switching strategy.

[0015] Another object of the present invention is to provide a readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0016] Another object of the present invention is to provide an electronic device including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the program to implement the steps of the method described above.

[0017] This invention constructs a regionalized street light collaborative network by clustering LED streetlights within a designated area based on preset clustering rules, breaking through the limitations of the traditional "distributed and isolated operation" architecture of streetlights. Furthermore, multimodal sensors deployed on the light poles simultaneously collect streetlight operating status data (such as lamp temperature and circuit load) and environmental perception data (such as water depth around the light poles). Based on the fusion of these two types of data, accurate alarm information is generated. The multimodal sensors cover the dual-dimensional monitoring needs of "equipment operation + external environment," enabling real-time detection of sudden risks such as excessive water levels, abnormal temperatures, and power outages, avoiding the risk of monitoring gaps. This system addresses the issue of missed fault detection due to power outages. It then broadcasts the alarm information via a preset multicast address to other streetlights within the cluster that have not generated alarms, achieving synchronous sharing of alarm information within the area. This breaks the limitation of traditional streetlights where "a single light alarms, but the surrounding area does not respond," allowing surrounding streetlights to detect risks in advance (such as water accumulation or extended line faults), reserving time for subsequent protective actions and preventing the fault from escalating. Finally, after receiving the alarm information, the streetlights within the cluster analyze the alarm type and match it with a preset power supply switching strategy (such as automatically switching to battery power when the mains power is interrupted, and cutting off non-critical circuits while maintaining battery lighting when the water level exceeds the standard). This solves the problems of poor reliability and safety in existing LED streetlight technologies. Attached Figure Description

[0018] Figure 1This is a flowchart of the LED street light alarm information generation method in the first embodiment of the present invention; Figure 2 This is a structural block diagram of the LED street light alarm information generation system in the third embodiment of the present invention.

[0019] The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0020] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Several embodiments of the invention are illustrated in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0021] It should be noted that when a component is said to be "fixed to" another component, it can be directly on the other component or there may be an intervening component. When a component is said to be "connected to" another component, it can be directly connected to the other component or there may be an intervening component. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the specification of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0023] Example 1 Please see Figure 1 The image shows an LED street light alarm information generation method according to the first embodiment of the present invention, which is applied to LED street light scenarios. Each LED street light can communicate with each other, and each LED street light includes an LED light source set on the top of the light pole and a multi-modal sensor deployed on the light pole. The LED light source can be powered by connecting to the mains power or by connecting to a battery set inside the light pole. The method includes steps S10 to S13.

[0024] Step S10: Based on preset clustering rules, all LED streetlights within the designated area are divided into clusters to form at least one streetlight cluster.

[0025] The core purpose of clustering is to group streetlights with similar geographical distribution, equipment characteristics, and operating environments into one group, which facilitates the accurate transmission and coordinated response of subsequent alarm information and avoids the inefficiency caused by decentralized management. For example, streetlights in low-lying areas of the same city that use the same model of controller can be grouped into a cluster to facilitate centralized alarm processing during rainstorms and floods, thereby improving the targeting of urban lighting management.

[0026] For example, the step of dividing all LED streetlights within a designated area into clusters based on preset clustering rules to form at least one streetlight cluster includes: Collect static basic data and dynamic operation data of each LED street light, and construct a three-dimensional feature set including geographical topology, equipment association, and operating environment characteristics; The feature set is Z-score standardized to eliminate dimensional differences. Core features with a correlation greater than the alarm threshold are selected by mutual information entropy. Redundant data is removed to form a feature matrix. Different weights are assigned to the geographical topology, device association, and operating environment features in the feature matrix, and the comprehensive feature vector of each street light is obtained by weighted summation. Based on the difference of the comprehensive feature vectors of each pair of streetlights, the comprehensive distance representing the correlation between each pair of streetlights is obtained, and the neighborhood radius of each streetlight is determined. If the number of streetlights in the neighborhood of a certain streetlight is greater than the threshold and the average device matching degree of these streetlights is greater than the preset value, then the streetlight is determined as the cluster center. Using the cluster center as the center, all streetlights that meet the condition of having a comprehensive distance less than the threshold are grouped into a cluster. For streetlights without qualified neighbors, the K-nearest neighbor algorithm is used to calculate the average distance between them and the surrounding clusters, and they are assigned to the cluster with the smallest distance. In the end, the clustering of all LED streetlights in the set area is realized.

[0027] Specifically, static basic data can include fixed attribute information such as the installation coordinates, equipment model, controller version, and rated power of the streetlights, while dynamic operating data can cover dynamic change data such as real-time power supply status, operating temperature, and historical fault records. Geographic topology features reflect the spatial relationship between streetlights (such as their distance from each other and whether they are located on the same road segment), equipment association features reflect the consistency of the hardware configuration of the streetlights (such as the same type of light source), and operating environment features include environmental attributes such as the terrain (whether it is low-lying) and climate conditions (average annual rainfall) of the area where the streetlights are located. The construction of the three-dimensional feature set can comprehensively characterize the characteristics of streetlights from three dimensions: space, equipment, and environment, providing data support for subsequent accurate clustering and avoiding insufficient cluster association caused by single-dimensional division. Next, Z-score standardization is performed on the feature set to eliminate dimensional differences. Core features with a correlation greater than a threshold to the alarm are selected using mutual information entropy. Redundant data is then removed to form a feature matrix. Z-score standardization converts feature data of different dimensions (e.g., distance in meters, temperature in degrees Celsius) into standardized data with a mean of 0 and a standard deviation of 1, eliminating the interference of data scale differences on the clustering results and ensuring balanced weighting of each feature in the calculation. Mutual information entropy measures the correlation strength between each feature and the alarm event. For example, water level sensing data has a very high correlation with flooding alarms, while street light installation time has a very low correlation with alarms. By selecting and retaining core features (e.g., geographical distance, equipment model matching degree, water level sensitivity coefficient) and removing redundant data (e.g., installation time, manufacturer), the feature matrix can be simplified, computational complexity reduced, and clustering accuracy improved. Then, different weights are assigned to the geographical topology, equipment association, and operating environment features in the feature matrix, and a comprehensive feature vector for each street light is obtained by weighted summation. The weight allocation is determined based on the importance of each feature to cluster collaborative emergency response. For example, the geographical topology feature has the highest weight (e.g., 0.4) because street lights that are spatially close are most likely to be affected by the same sudden event (e.g., localized rainstorms, power outages in road sections); the equipment association feature has the next highest weight (e.g., 0.3) because equipment of the same model has consistent failure modes and emergency response mechanisms, facilitating collaborative handling; the operating environment feature has a weight of 0.3 because street lights in the same environment face more similar types of risks (e.g., all low-lying areas face the risk of flooding). By weighted summation, the three-dimensional features are integrated into a single comprehensive feature vector, giving each street light a unique quantitative representation, which facilitates subsequent calculation of association.

[0028] Then, based on the difference in the comprehensive feature vectors of every two streetlights, a comprehensive distance representing the correlation between them is obtained. This determines the neighborhood radius of each streetlight. If the number of streetlights in a streetlight's neighborhood with a comprehensive distance less than a threshold is greater than a preset number, and the average device matching degree of these streetlights is greater than a preset value, then that streetlight is determined as a cluster center. The smaller the comprehensive distance, the more similar the characteristics of the two streetlights, and the stronger the correlation. The neighborhood radius can be set according to urban road planning (e.g., 500 meters, ensuring coverage of streetlights on the same or adjacent road segments). The preset number (e.g., 5) and device matching degree threshold (e.g., 0.8) are used to select sufficiently representative cluster centers, avoiding the selection of isolated streetlights or streetlights with large characteristic differences as centers, ensuring that the cluster centers can truly reflect the core characteristics of a cluster. For example, if a streetlight has 6 streetlights in its neighborhood with a comprehensive distance less than the threshold, and the average device matching degree of these streetlights is 0.85, then it becomes a cluster center. For example, a comprehensive distance can be obtained based on the Euclidean distance, Manhattan distance, or cosine distance of the features, while the equipment matching degree can be based on core components that are strongly related to emergency functions; then, the characteristics of these components (high scores are given for consistent qualitative characteristics such as model number, and high scores are given for similar quantitative characteristics such as capacity / power) are quantitatively evaluated; finally, the components are weighted and summed according to their importance to obtain the equipment matching degree in the range of 0-1, with the closer the value is to 1, the higher the matching degree. All of this is understandable to those skilled in the art and will not be elaborated here.

[0029] Finally, using the cluster center as the center, all streetlights that meet the condition of a comprehensive distance less than a threshold are grouped into one cluster. For streetlights without qualifying neighborhoods, the K-nearest neighbor algorithm is used to calculate the average distance between them and their surrounding clusters, assigning them to the cluster with the smallest distance. This ultimately achieves the clustering of streetlights across the entire area. Streetlights without qualifying neighborhoods are usually isolated (such as scattered streetlights in suburban areas). The K-nearest neighbor algorithm (K value can be set to 3) calculates the average distance between them and multiple surrounding clusters to ensure that they are assigned to the cluster with the most similar characteristics, avoiding omissions. The beneficial effects of this clustering scheme are high accuracy in the partitioning results, strong spatial correlation, equipment consistency, and environmental similarity of streetlights within the cluster, providing an efficient grouping basis for subsequent alarm broadcasting and collaborative emergency response, reducing interference from irrelevant information, and lowering the complexity of urban lighting management.

[0030] Step S11: Obtain the real-time operating status data and environmental perception data collected by the multi-modal sensors of each LED street light, and generate the alarm information corresponding to the current LED street light based on the operating status information and environmental perception data.

[0031] The operational status data can include equipment-related data such as the working voltage, current, and brightness parameters of the LED light source, while the environmental perception data covers external environmental data such as water level, ambient temperature, and humidity around the light pole. By comprehensively analyzing the two types of data, abnormal potential hazards of the streetlights can be fully captured, such as excessive water level, overheating of the light source, and abnormal power supply. Generating alarm information is a key step in achieving early warning of risks and buying time for subsequent emergency response.

[0032] Specifically, the steps for generating alarm information for the current LED streetlight based on operational status information and environmental perception data include: First, the acquired operational status data and environmental perception data are classified into different modalities. Differentiated denoising strategies are adopted for the noise characteristics of different modalities. Then, the synchronization and alignment of the two types of data are completed based on timestamps, and a binary data matrix containing device status dimension and environmental dimension is constructed. Modal classification divides data into equipment operation modes (such as light source current, battery voltage, and heat dissipation system status) and environmental perception modes (such as water level, ambient temperature, and humidity). Different modes have different noise characteristics. For example, water level data in environmental perception data is susceptible to impulse noise from raindrops, which can be denoised using median filtering. Current data in operation status data is susceptible to Gaussian noise from circuit interference, which can be denoised using mean filtering or wavelet filtering. Differential denoising can preserve effective data to the greatest extent. Timestamp synchronization and alignment matches equipment data and environmental data collected at the same time. For example, light source current data collected when lights need to be turned on is matched with ambient temperature data collected at the same time to avoid analysis errors caused by data time misalignment. Specifically, different data collection times are determined due to different sunset times in different seasons. In practice, the accurate sunset time is also calculated based on latitude and longitude to determine the data collection time. The construction of a binary data matrix realizes the fusion and correlation of equipment status and environmental factors, providing a complete data foundation for subsequent fault analysis.

[0033] Next, the core feature vectors of the binary data matrix are extracted. After eliminating the difference in dimensions by Z-score standardization, the environmental adaptive weight formula is introduced to dynamically allocate the fusion weights of the two types of features. A feature interaction model is constructed through a cross-modal attention mechanism to focus on feature combinations that are strongly associated with equipment faults. The cosine similarity between the feature combination and the preset fault mode library is calculated, and candidate abnormal feature sets with similarity greater than the preset threshold are selected. Core feature vector extraction extracts key indicators (such as battery voltage fluctuation amplitude, water level rise rate, and temperature exceedance duration) from binary matrices; Z-score standardization ensures that core features of different dimensions can directly participate in the calculation; the environment-adaptive weight formula dynamically adjusts the weights of equipment features and environmental features according to real-time environmental conditions. For example, the weight of environmental features (water level) is increased during heavy rain, and the weight of equipment features (supply voltage) is increased during peak power periods, making feature fusion more consistent with the actual scenario; the cross-modal attention mechanism can automatically focus on strongly correlated feature combinations. For example, the combination "water level rise rate 0.5m / h + circuit current fluctuation 10%" is highly correlated with short-circuit faults, and this mechanism can strengthen the weight of such combinations; the preset fault mode library is built based on a large amount of historical fault data of the same type of streetlights, containing feature combinations corresponding to various faults (such as "temperature exceeds 60℃ + brightness decay 30% = light source overheating fault"). Cosine similarity is used to measure the degree of matching between real-time feature combinations and fault modes in the library, and a preset threshold (such as 0.7) is used to filter out candidate feature sets of suspected faults to reduce misjudgments.

[0034] Finally, a Bayesian network model is used to perform probabilistic reasoning on the candidate anomaly feature set to obtain the corresponding fault type and probability, thereby generating alarm information. The prior probability of the network is trained based on historical fault data of the same model of LED street light. Bayesian network is a graphical model based on probabilistic reasoning, which can handle uncertainty problems. The prior probability is obtained by training with historical fault data (e.g., "the probability of short circuit fault when water level exceeds the standard is 85%)). After inputting the candidate anomaly feature set, the model can calculate the posterior probability of various faults. For example, the probability of short circuit fault is 92% and the probability of overheating fault is 15% for "water level rise rate of 0.5m / h + circuit current fluctuation of 10%". Finally, the fault type is determined to be short circuit fault and alarm information is generated (e.g., "Alarm type: risk of short circuit due to water immersion in light pole; fault probability: 92%; recommendation: immediately switch to battery power and disconnect non-critical circuits").

[0035] Step S12: Broadcast the alarm information to other LED streetlights in the same cluster that have not generated alarm information through a preset cluster multicast address, so that the other LED streetlights that have not generated alarm information can receive the corresponding alarm information.

[0036] Because streetlights within a cluster share similar operating environments and equipment characteristics, an anomaly in one streetlight (such as water immersion or partial power outage) may spread to other streetlights in the same cluster. By broadcasting within the cluster, surrounding streetlights can be informed of the risks in advance, thus preventing cascading failures. For example, if a streetlight detects that the water level exceeds the standard, other streetlights in the cluster can receive the alarm and prepare for power switching in advance. Compared to individual light responses, clustered broadcasting can significantly improve the synchronization of emergency responses.

[0037] Step S13: When the LED streetlights in the cluster receive an alarm message, the alarm message is analyzed to match the preset power supply switching strategy according to the alarm message, and the power supply is switched to mains power or battery power according to the power supply switching strategy.

[0038] When an LED street light in the cluster receives an alarm message, it analyzes the message to match a preset power supply switching strategy, and then switches between mains power and battery power accordingly. For example, if the alarm message indicates a mains power outage or a risk of short circuit due to water immersion, the street light can automatically switch to battery power to ensure continuous LED illumination, prevent the road from falling into darkness, ensure the safety of pedestrians and vehicles, and reduce traffic accidents and economic losses caused by lighting interruptions, significantly improving the reliability and safety of the street light system.

[0039] In summary, the LED street light alarm information generation method in the above embodiments of the present invention, by dividing LED street lights within a set area into clusters based on preset clustering rules, constructs a regionalized street light collaborative network, breaking through the limitations of the traditional street light "distributed isolated operation" architecture; then, through multimodal sensors deployed on the light poles, it simultaneously collects street light operating status data (such as lamp temperature and circuit load) and environmental perception data (such as water level depth around the light pole), and generates accurate alarm information based on the fusion of the two types of data—the multimodal sensors cover the dual-dimensional monitoring needs of "equipment operation + external environment", and can capture in real time water level exceeding the standard, abnormal temperature, and mains power interruption. To prevent missed fault detection due to unforeseen risks, alarm information is broadcast via a pre-defined cluster multicast address to other streetlights within the cluster that have not yet generated alarms. This enables synchronous sharing of alarm information within the area, breaking the limitation of traditional streetlights where "a single light alarms, but the surrounding area does not respond." This allows surrounding streetlights to detect risks in advance (such as water accumulation or extended line faults), providing time for subsequent protective actions and preventing the fault from escalating. Finally, after receiving the alarm information, the streetlights within the cluster analyze the alarm type and match it with a pre-defined power supply switching strategy (such as automatically switching to battery power when the mains power is interrupted, and cutting off non-critical circuits while maintaining battery lighting when the water level exceeds the standard). This solves the problems of poor reliability and safety in existing LED streetlight technologies.

[0040] Example 2 This embodiment also proposes a method for generating LED street light alarm information. The difference between the LED street light alarm information generation method in this embodiment and the LED street light alarm information generation method in Embodiment 1 is as follows: The step of broadcasting the alarm information to other LED streetlights in the same cluster that have not generated alarm information via a preset cluster multicast address, so that the other LED streetlights that have not generated alarm information receive the corresponding alarm information, includes: According to preset rules, a leader street light is elected in the street light cluster, and alarm information is broadcast to the leader node in the cluster through a preset cluster multicast address. The leader node broadcasts alarm information to other LED streetlights in its cluster that have not generated alarm information via a preset cluster multicast address, so that the other LED streetlights that have not generated alarm information can receive the corresponding alarm information.

[0041] The cluster multicast address is a dedicated communication address pre-assigned to each cluster, used to limit the broadcast range of alarm information, avoid cross-cluster interference, and improve communication efficiency. The purpose of electing a leader streetlight is to establish a communication scheduling core within the cluster, avoiding communication conflicts and resource waste caused by multiple streetlights broadcasting simultaneously. The leader streetlight must have strong communication capabilities, stable operation, and a central geographical location, enabling it to efficiently receive and forward alarm information. For example, the streetlight with the widest communication coverage and the fewest historical failures within a cluster can be selected as the leader node. Streetlights generating alarm information first send the information to the leader node, rather than broadcasting it directly to all streetlights, reducing the communication load on individual streetlights and ensuring the stability of alarm information transmission. As the communication hub within the cluster, the leader node possesses unified scheduling capabilities, enabling it to synchronously and efficiently forward alarm information to all streetlights within the cluster that have not triggered alarms, avoiding omissions or delays in information transmission. For example, if a streetlight in a cluster detects a power outage and generates an alarm, after the streetlight sends the alarm to the leader node, the leader node immediately broadcasts it to the remaining 19 streetlights in the cluster that have not generated alarms, ensuring that all streetlights are simultaneously aware of the power supply anomaly and prepare to switch to battery power in a timely manner.

[0042] In addition, in some optional embodiments of the present invention, the step of electing a leader streetlight within the streetlight cluster according to preset rules includes: Obtain the coordinate information of each street light in the cluster, and map each street light to a preset coordinate system based on the coordinate information; A longitudinal baseline is formed by connecting the northernmost and southernmost streetlights within the cluster, and a transverse baseline is formed by connecting the easternmost and westernmost streetlights. The intersection of the two baselines is recorded as the geometric center of the cluster. Based on the geometric center of the cluster, the corresponding leader node determination area is determined, and one of the sub-regions of the leader node determination area divided by the baseline is selected as the target leader node determination area according to the preset rules. Draw a circle with the coordinates of each street light as the center and the standard communication radius of the street light model as the radius to obtain the communication coverage circle of each street light. Draw an auxiliary line through the geometric center in the area determined by the target leader node at a preset angle to the longitudinal baseline. Select the preset number of street lights with the most intersections between the auxiliary line and each communication coverage circle as candidate leader nodes. Taking the geometric center of the cluster as the origin, calculate the extreme radius of each candidate leader node, select the candidate node with the smallest extreme radius as the primary leader node, and then calculate the overlap area of ​​the communication coverage circles of the primary leader node and other candidate nodes. Select the node with the smallest overlap area and a distance from the primary leader node greater than a preset multiple of the communication radius as the backup leader node.

[0043] First, the coordinate information of each street light in the cluster is obtained, and then each street light is mapped to a preset coordinate system based on the coordinate information. The coordinate information can be obtained through the GPS module built into the street light (such as latitude and longitude data). The preset coordinate system adopts a Cartesian coordinate system, converting latitude and longitude into planar coordinate values ​​(such as the X-axis representing the east-west direction and the Y-axis representing the north-south direction), which facilitates the calculation of the positional relationship and geometric center between street lights, and provides a quantitative basis for the subsequent selection of the location of the leader node.

[0044] Next, a vertical baseline is formed by connecting the northernmost and southernmost streetlights within the cluster, and a horizontal baseline is formed by connecting the easternmost and westernmost streetlights. The intersection of the two baselines is recorded as the geometric center of the cluster. The vertical and horizontal baselines constitute the spatial boundary framework of the cluster, while the geometric center is the core spatial location of the cluster. The leader node is selected in the vicinity of the geometric center to maximize coverage of all streetlights in the cluster and reduce signal attenuation or delay caused by excessive communication distance.

[0045] Then, based on the geometric center of the cluster, the corresponding leader node determination area is determined. Following preset rules, one sub-region from the leader node determination area divided by the baseline is selected as the target leader node determination area. The baseline divides the cluster into four sub-regions (e.g., northeast, southeast, northwest, and southwest sub-regions). The leader node determination area revolves around the geometric center. The selection of the target sub-region needs to comprehensively consider factors such as the distribution density of streetlights and operational stability within the sub-region. The core objective is to select an area with reasonable streetlight distribution and excellent communication conditions as the location of the leader node, ensuring effective communication coverage for the leader node.

[0046] Next, circles are drawn with each streetlight's coordinates as the center and the standard communication radius of that streetlight model as the radius, resulting in the communication coverage circle for each streetlight. An auxiliary line is then drawn through the geometric center in the area defined by the target leader node, forming a preset angle with the longitudinal baseline. A preset number of streetlights whose communication coverage circles intersect the auxiliary line the most are selected as candidate leader nodes. The standard communication radius is determined by the performance of the streetlight's wireless communication module (e.g., 300 meters). The communication coverage circle visually reflects the communication coverage range of a single streetlight. The preset angle can be set to 30 degrees or 45 degrees. The auxiliary line is used to filter streetlights whose communication coverage range overlaps significantly with the core area of ​​the cluster. The preset number (e.g., 3) limits the number of candidate nodes, avoiding excessive selection complexity. For example, if the auxiliary line intersects the communication coverage circles of 3 streetlights the most times (8, 7, and 6 times respectively), then these 3 streetlights become candidate leader nodes, ensuring that the candidate nodes have strong communication coverage capabilities.

[0047] Finally, taking the cluster's geometric center as the origin, the extreme radius of each candidate leader node is calculated. The candidate node with the smallest extreme radius is selected as the primary leader node. Then, the overlap area of ​​the communication coverage circles between the primary leader node and the other candidate nodes is calculated. The node with the smallest overlap area and a distance greater than a preset multiple of the communication radius from the primary leader node is selected as the backup leader node. The extreme radius is the straight-line distance from the origin (geometric center) to the candidate node. The candidate node with the smallest extreme radius is closest to the cluster core and its communication coverage is most likely to cover the entire cluster, thus it is selected as the primary leader node. The requirements of a preset multiple of the communication radius (e.g., 1.5 times) and the smallest overlap area are to ensure that the communication coverage areas of the primary and backup leader nodes do not overlap too much, avoiding the backup node being unable to effectively take over when the primary node fails. For example, if the extreme radius of the primary leader node is 100 meters, and a backup node is 500 meters away from the primary node (greater than 1.5 × 300 = 450 meters), and has the smallest overlap area of ​​the coverage circles, then it is selected as the backup.

[0048] Furthermore, the step of broadcasting the alarm information to the leader node within the cluster via a preset cluster multicast address includes: The LED street light that generates alarm information locates the main leader node through coordinate matching. If its own communication coverage circle overlaps with the coverage circle of the main leader node, it directly transmits the alarm information to the main leader node. If there is no overlap, other streetlights within its own coverage circle are identified as relay nodes, and the relay nodes relay the data to the main leader node. After receiving the alarm, the master leader node draws a circle with itself as the center, dividing the cluster into an inner circle and an outer circle. The streetlights in the inner circle directly receive the broadcast, while the streetlights in the outer circle are forwarded by the streetlights designated by the master leader node at the edge of the coverage circle, ensuring that alarm information is covered by the entire cluster.

[0049] First, the LED streetlight that generates the alarm information locates the main leader node through coordinate matching. If its own communication coverage circle overlaps with the main leader node's coverage circle, the alarm information is directly transmitted to the main leader node. Coordinate matching determines the positional relationship between the two by comparing their own coordinates with those of the main leader node. Overlapping communication coverage circles mean that the distance between them is less than the standard communication radius (e.g., 300 meters). In this case, the alarm information can be directly transmitted through the wireless communication module. This method features a short transmission path, high speed, and low loss, ensuring that the alarm information quickly reaches the main leader node.

[0050] If a streetlight's own communication coverage circle does not overlap with the coverage circle of the main leader node, then another streetlight within its own coverage circle is selected as a relay node, which relays the signal to the main leader node. When the distance between the two is greater than the standard communication radius (e.g., 400 meters), direct transmission signal attenuation is severe, requiring relay transmission through a relay node. The relay node must be a normally functioning streetlight within its own communication coverage circle, and its communication coverage circle must overlap with the coverage circle of the main leader node or the next relay node to form a transmission link. For example, if streetlight A, which generates an alarm, is 400 meters away from the main leader node C, and streetlight B is within A's coverage circle, and B is 280 meters away from C (overlapping coverage circles), then A first transmits the alarm to B, and then B forwards it to C, ensuring that the alarm information can overcome distance barriers to reach the main leader node.

[0051] After receiving alarm information, the master leader node draws a circle with itself as the center, dividing the cluster into an inner circle and an outer circle. Streetlights in the inner circle directly receive the broadcast, while those in the outer circle are forwarded by streetlights designated by the master leader node at the edge of its coverage circle, ensuring full cluster coverage of alarm information. Inner circle streetlights are within the direct communication range of the master leader node and can directly receive the broadcast, resulting in high transmission efficiency. Outer circle streetlights are farther from the master leader node and require forwarding nodes to extend their coverage. The forwarding nodes designated by the master leader node are streetlights at the edge of its coverage circle, extending their coverage to the outer circle area, forming a secondary broadcast. For example, if the master leader node's coverage circle radius is 300 meters, the inner circle consists of streetlights ≤300 meters away, which directly receive the broadcast; the outer circle consists of streetlights >300 meters away. The master leader node designates three streetlights at the edge of its coverage circle (300 meters away) as forwarding nodes, which then broadcast the alarm information to all streetlights in the outer circle, ensuring no omissions.

[0052] In addition, in some optional embodiments of the present invention, the step of selecting one sub-region from the sub-regions of the leader node determination region divided by the baseline as the target leader node determination region according to a preset rule includes: The sub-region attribute dataset is obtained by collecting the number of valid streetlights and the coordinates of each streetlight within each sub-region and calculating the distribution concentration.

[0053] Based on the standard communication coverage radius of LED streetlights of the same model, the maximum number of nodes that a single light can theoretically directly cover is determined. Based on this, the optimal range of the number of effective streetlights in a sub-region is set to filter sub-regions where the number of effective streetlights is within the optimal range. At the same time, sub-regions with a distribution concentration greater than the threshold are retained as candidate sets. For the generated candidate set, the sub-region with the largest number of valid streetlights is selected first. If the number is the same, the distribution concentration is compared and the sub-region with higher distribution concentration is selected to finally determine the target sub-region.

[0054] First, taking each sub-region as a unit, the number of valid streetlights and the coordinates of each streetlight within that region were collected to calculate the distribution concentration, thus obtaining the sub-region attribute dataset. The number of valid streetlights refers to the total number of streetlights within the sub-region that are operating normally (no faults, normal communication). A higher number of valid streetlights indicates a more stable streetlight operation in that region, and thus higher reliability for the leading node region. Distribution concentration is obtained by calculating the variance of the coordinates of all streetlights within the sub-region. A smaller variance indicates a more concentrated streetlight distribution, facilitating communication coverage for the leading node. For example, a sub-region with a streetlight coordinate X-axis variance of 50 and a Y-axis variance of 40 has a high distribution concentration, while another sub-region with an X-axis variance of 200 and a Y-axis variance of 180 has a more dispersed distribution. The collection of the sub-region attribute dataset provides quantitative indicators for subsequent selection, avoiding selection bias caused by subjective judgment.

[0055] Next, based on the standard communication coverage radius of LED streetlights of the same model, the maximum number of nodes that a single streetlight can theoretically directly cover is determined. Using this as a benchmark, an optimal range for the number of effective streetlights in a sub-region is set. Sub-regions with the number of effective streetlights within this optimal range are selected, while sub-regions with a distribution concentration greater than a threshold are retained as a candidate set. The standard communication coverage radius (e.g., 300 meters) determines the maximum communication coverage range of a single streetlight, from which the maximum number of nodes that a single streetlight can theoretically directly cover can be calculated (e.g., a maximum of 10 streetlights can be directly communicated with). The optimal range is set to 60%-100% of the theoretical maximum number of nodes (e.g., 6-10). Sub-regions with the number of effective streetlights within this range will not suffer from insufficient communication coverage due to too few streetlights, nor from communication congestion due to too many streetlights. The distribution concentration threshold (e.g., variance ≤ 100) is used to filter sub-regions with concentrated distribution, ensuring that the leading node can efficiently cover the streetlights within that area. For example, if a sub-region has 8 valid streetlights (in the range of 6-10) and a distribution concentration variance of 80 (≤100), it will be included in the candidate set; if a sub-region has 5 valid streetlights (less than 6) or a distribution concentration variance of 120 (>100), it will be removed.

[0056] Finally, for the generated candidate set, the sub-region with the largest number of valid streetlights is selected first. If the number is the same, the distribution concentration is compared, and the sub-region with higher distribution concentration is selected to determine the target sub-region. Prioritizing the sub-region with the largest number of valid streetlights is because more valid streetlights mean a more stable operating environment and more abundant communication resources in that region, providing better support for the leader node. When multiple sub-regions have the same number of valid streetlights, the sub-region with higher distribution concentration (smaller variance) is more advantageous because concentrated streetlight distribution results in higher communication coverage efficiency for the leader node. For example, if the candidate set has two sub-regions, region A with 8 valid streetlights and a variance of 80, and region B with 8 valid streetlights and a variance of 70, then region B is selected as the target sub-region; if region A has 9 valid streetlights and region B has 8, then region A is selected. The selection of the target sub-region is based on quantitative indicators, is scientific and reasonable, and ensures that the selected region has sufficient valid streetlights and superior distribution conditions, providing a good deployment environment for the leader node and further improving the efficiency of alarm broadcasting and collaborative emergency response within the cluster.

[0057] In summary, the LED street light alarm information generation method in the above embodiments of the present invention, by dividing LED street lights within a set area into clusters based on preset clustering rules, constructs a regionalized street light collaborative network, breaking through the limitations of the traditional street light "distributed isolated operation" architecture; then, through multimodal sensors deployed on the light poles, it simultaneously collects street light operating status data (such as lamp temperature and circuit load) and environmental perception data (such as water level depth around the light pole), and generates accurate alarm information based on the fusion of the two types of data—the multimodal sensors cover the dual-dimensional monitoring needs of "equipment operation + external environment", and can capture in real time water level exceeding the standard, abnormal temperature, and mains power interruption. To prevent missed fault detection due to unforeseen risks, alarm information is broadcast via a pre-defined cluster multicast address to other streetlights within the cluster that have not yet generated alarms. This enables synchronous sharing of alarm information within the area, breaking the limitation of traditional streetlights where "a single light alarms, but the surrounding area does not respond." This allows surrounding streetlights to detect risks in advance (such as water accumulation or extended line faults), providing time for subsequent protective actions and preventing the fault from escalating. Finally, after receiving the alarm information, the streetlights within the cluster analyze the alarm type and match it with a pre-defined power supply switching strategy (such as automatically switching to battery power when the mains power is interrupted, and cutting off non-critical circuits while maintaining battery lighting when the water level exceeds the standard). This solves the problems of poor reliability and safety in existing LED streetlight technologies.

[0058] Example 3 Please see Figure 2The image shows an LED street light alarm information generation system proposed in the third embodiment of the present invention, applied in an LED street light scenario. Each LED street light can communicate with each other, and each LED street light includes an LED light source mounted on the top of the light pole and a multimodal sensor deployed on the light pole. The LED light source can be powered by connecting to mains power or by connecting to a battery installed inside the light pole. The system includes: The partitioning module 100 is used to partition all LED streetlights in a set area into clusters based on preset clustering rules, forming at least one streetlight cluster. The acquisition module 200 is used to acquire the real-time operating status data and environmental perception data collected by the multimodal sensors of each LED street light, and generate alarm information corresponding to the current LED street light based on the operating status information and environmental perception data. The broadcast module 300 is used to broadcast alarm information to other LED streetlights in the same cluster that have not generated alarm information through a preset cluster multicast address, so that other LED streetlights that have not generated alarm information can receive the corresponding alarm information; The switching module 400 is used to analyze the alarm information when the LED streetlights in the cluster receive alarm information, and match the preset power supply switching strategy according to the alarm information, and switch to mains power or battery power according to the power supply switching strategy.

[0059] The functions or operation steps implemented by the above modules are largely the same as those in the above method embodiments, and will not be repeated here.

[0060] Example 4 In another aspect, the present invention provides a readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the steps of the method described in any one of Embodiments 1 to 2 above.

[0061] Example 5 In another aspect, the present invention provides an electronic device, the electronic device including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the program to implement the steps of any one of the methods described in Embodiments 1 to 2 above.

[0062] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0063] Those skilled in the art will understand that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequential list of executable instructions for implementing logical functions, and can be embodied in any computer-readable storage medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable storage medium" can mean any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0064] More specific examples (a non-exhaustive list) of computer-readable storage media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable storage media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0065] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0066] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0067] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for generating alarm information for LED streetlights, characterized in that, In scenarios where LED streetlights are applied, each LED streetlight can communicate with each other. Each LED streetlight includes an LED light source mounted on the top of the light pole and a multi-modal sensor deployed on the light pole. The LED light source can be powered by connecting to mains power or by connecting to a battery installed inside the light pole. The method includes: Based on preset clustering rules, all LED streetlights within a designated area are divided into clusters to form at least one streetlight cluster. The system acquires real-time operating status data and environmental perception data from the multimodal sensors of each LED street light, and generates alarm information corresponding to the current LED street light based on the operating status information and environmental perception data. The alarm information is broadcast to other LED streetlights in the same cluster that have not generated alarm information through a preset cluster multicast address, so that the other LED streetlights that have not generated alarm information can receive the corresponding alarm information; When the LED streetlights in the cluster receive an alarm message, they analyze the alarm message to match the preset power supply switching strategy, and then switch to mains power or battery power according to the corresponding power supply switching strategy.

2. The LED street light alarm information generation method according to claim 1, characterized in that, The step of dividing all LED streetlights within a designated area into clusters based on preset clustering rules to form at least one streetlight cluster includes: Collect static basic data and dynamic operation data of each LED street light, and construct a three-dimensional feature set including geographical topology, equipment association, and operating environment characteristics; The feature set is Z-score standardized to eliminate dimensional differences. Core features with a correlation greater than the alarm threshold are selected by mutual information entropy. Redundant data is removed to form a feature matrix. Different weights are assigned to the geographical topology, device association, and operating environment features in the feature matrix, and the comprehensive feature vector of each street light is obtained by weighted summation. Based on the difference of the comprehensive feature vectors of each pair of streetlights, the comprehensive distance representing the correlation between each pair of streetlights is obtained, and the neighborhood radius of each streetlight is determined. If the number of streetlights in the neighborhood of a certain streetlight is greater than the threshold and the average device matching degree of these streetlights is greater than the preset value, then the streetlight is determined as the cluster center. Using the cluster center as the center, all streetlights that meet the condition of having a comprehensive distance less than the threshold are grouped into a cluster. For streetlights without qualified neighbors, the K-nearest neighbor algorithm is used to calculate the average distance between them and the surrounding clusters, and they are assigned to the cluster with the smallest distance. In the end, the clustering of all LED streetlights in the set area is realized.

3. The LED street light alarm information generation method according to claim 1, characterized in that, The step of generating the alarm information corresponding to the current LED street light based on the operating status information and environmental perception data includes: Modal classification is performed on the acquired operational status data and environmental perception data. Differentiated denoising strategies are adopted for the noise characteristics of different modal data. The synchronization and alignment of the two types of data are completed based on timestamps, and a binary data matrix containing device status dimension and environmental dimension is constructed. The core feature vectors of the binary data matrix are extracted. After eliminating the difference in dimensions by Z-score standardization, the fusion weights of the two types of features are dynamically allocated by the environment adaptive weight formula. A feature interaction model is constructed through a cross-modal attention mechanism to focus on feature combinations that are strongly associated with equipment faults. The cosine similarity between the feature combination and the preset fault mode library is calculated, and candidate abnormal feature sets with similarity greater than the preset threshold are selected. A Bayesian network model is used to perform probabilistic reasoning on the candidate anomaly feature set to obtain the corresponding fault type and probability in order to generate alarm information. The prior probability of the network is trained based on historical fault data of the same type of LED street light.

4. The LED street light alarm information generation method according to claim 1, characterized in that, The step of broadcasting the alarm information to other LED streetlights in the same cluster that have not generated alarm information via a preset cluster multicast address, so that the other LED streetlights that have not generated alarm information receive the corresponding alarm information, includes: According to preset rules, a leader street light is elected in the street light cluster, and alarm information is broadcast to the leader node in the cluster through a preset cluster multicast address. The leader node broadcasts alarm information to other LED streetlights in its cluster that have not generated alarm information via a preset cluster multicast address, so that the other LED streetlights that have not generated alarm information can receive the corresponding alarm information.

5. The LED street light alarm information generation method according to claim 4, characterized in that, The step of electing a leader streetlight within the streetlight cluster according to preset rules includes: Obtain the coordinate information of each street light in the cluster, and map each street light to a preset coordinate system based on the coordinate information; A longitudinal baseline is formed by connecting the northernmost and southernmost streetlights within the cluster, and a transverse baseline is formed by connecting the easternmost and westernmost streetlights. The intersection of the two baselines is recorded as the geometric center of the cluster. Based on the geometric center of the cluster, the corresponding leader node determination area is determined, and one of the sub-regions of the leader node determination area divided by the baseline is selected as the target leader node determination area according to the preset rules. Draw a circle with the coordinates of each street light as the center and the standard communication radius of the street light model as the radius to obtain the communication coverage circle of each street light. Draw an auxiliary line through the geometric center in the area determined by the target leader node at a preset angle to the longitudinal baseline. Select the preset number of street lights with the most intersections between the auxiliary line and each communication coverage circle as candidate leader nodes. Taking the geometric center of the cluster as the origin, calculate the extreme radius of each candidate leader node, select the candidate node with the smallest extreme radius as the primary leader node, and then calculate the overlap area of ​​the communication coverage circles of the primary leader node and other candidate nodes. Select the node with the smallest overlap area and a distance from the primary leader node greater than a preset multiple of the communication radius as the backup leader node.

6. The LED street light alarm information generation method according to claim 5, characterized in that, The step of broadcasting alarm information to the leader node within the cluster via a preset cluster multicast address includes: The LED street light that generates alarm information locates the main leader node through coordinate matching. If its own communication coverage circle overlaps with the coverage circle of the main leader node, it directly transmits the alarm information to the main leader node. If there is no overlap, other streetlights within its own coverage circle are identified as relay nodes, and the relay nodes relay the data to the main leader node. After receiving the alarm, the master leader node draws a circle with itself as the center, dividing the cluster into an inner circle and an outer circle. The streetlights in the inner circle directly receive the broadcast, while the streetlights in the outer circle are forwarded by the streetlights designated by the master leader node at the edge of the coverage circle, ensuring that alarm information is covered by the entire cluster.

7. The LED street light alarm information generation method according to claim 6, characterized in that, The step of selecting one sub-region from the sub-regions of the leader node determination area divided by the baseline as the target leader node determination area according to preset rules includes: The sub-region attribute dataset is obtained by collecting the number of valid streetlights and the coordinates of each streetlight within each sub-region and calculating the distribution concentration. Based on the standard communication coverage radius of LED streetlights of the same model, the maximum number of nodes that a single light can theoretically directly cover is determined. Based on this, the optimal range of the number of effective streetlights in a sub-region is set to filter sub-regions where the number of effective streetlights is within the optimal range. At the same time, sub-regions with a distribution concentration greater than the threshold are retained as candidate sets. For the generated candidate set, the sub-region with the largest number of valid streetlights is selected first. If the number is the same, the distribution concentration is compared and the sub-region with higher distribution concentration is selected to finally determine the target sub-region.

8. An LED street light alarm information generation system, characterized in that, In LED street lighting applications, each LED street light can communicate with each other. Each LED street light includes an LED light source mounted on the top of the light pole and multimodal sensors deployed on the light pole. The LED light source can be powered by connecting to mains power or by connecting to a battery installed inside the light pole. The system includes: The partitioning module is used to divide all LED streetlights in a set area into clusters based on preset clustering rules, forming at least one streetlight cluster. The acquisition module is used to acquire real-time operating status data and environmental perception data collected by the multimodal sensors of each LED street light, and generate alarm information corresponding to the current LED street light based on the operating status information and environmental perception data. The broadcast module is used to broadcast alarm information to other LED streetlights in the same cluster that have not generated alarm information through a preset cluster multicast address, so that other LED streetlights that have not generated alarm information can receive the corresponding alarm information; The switching module is used to analyze the alarm information when the LED streetlights in the cluster receive alarm information, and match the preset power supply switching strategy according to the alarm information, and switch to mains power or battery power according to the corresponding power supply switching strategy.

9. A readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor, when executing the program, implements the steps of the method as described in any one of claims 1 to 7.

Images