AIoT Device Control Method and System under Cloud-Edge Collaboration Architecture

Abstract

The present application provides an AIoT device control method and system under a cloud-edge collaboration architecture, relating to the field of intelligent control technologies. The method includes: clustering and dividing a target road according to night brightness monitoring records, and arranging a brightness perception layer for each independent road section; monitoring multiple real-time ambient brightnesses; monitoring and obtaining road traffic flow and vehicle images through a data perception layer; predicting the brightness of vehicle headlights at an edge computing layer; optimizing the standard night driving brightness according to the multiple real-time ambient brightnesses and the predicted brightness of vehicle headlights to obtain multiple adapted driving brightnesses, and dynamically regulating the street lamp system of the target road. By means of the present application, the technical problem in the prior art that due to the inability of AIoT street lamp control to dynamically adjust the brightness according to real-time traffic and environmental conditions, over-illumination or insufficient illumination of street lamps occurs, affecting driving safety, can be solved. By dynamically regulating the street lamps according to the actual requirements of road sections, the energy utilization efficiency and driving safety are improved.

Description

Technical Field

[0001] This application relates to the field of intelligent control technology, and particularly to an AIoT device control method and system under a cloud-edge collaboration architecture. Background Art

[0002] In urban construction, AIoT devices are widely used in fields such as traffic management, environmental monitoring, and intelligent street lamp control. As a typical application of AIoT, intelligent street lamps are equipped with devices such as sensors, cameras, and communication modules, which can sense environmental information, traffic flow, etc. in real time, and dynamically adjust according to this data to achieve energy conservation, improve road safety, and optimize management efficiency. Traditional intelligent street lamp control usually adjusts the brightness of street lamps based on fixed time periods or ambient light intensity sensors, but cannot consider dynamic factors such as traffic flow, weather changes, and traffic patterns on the road section. As a result, during peak traffic hours or in bad weather, the brightness of street lamps is insufficient, affecting driving safety; while during low traffic flow or high ambient brightness periods, there may be over-illumination, wasting energy.

[0003] In summary, there is a technical problem in the prior art that due to the inability of AIoT street lamp control to dynamically adjust the brightness according to real-time traffic and environmental conditions, it leads to over-illumination or under-illumination of street lamps, further affecting energy utilization efficiency and driving safety. Summary of the Invention

[0004] The purpose of this application is to provide an AIoT device control method and system under a cloud-edge collaboration architecture to solve the technical problem in the prior art that due to the inability of AIoT street lamp control to dynamically adjust the brightness according to real-time traffic and environmental conditions, it leads to over-illumination or under-illumination of street lamps, further affecting energy utilization efficiency and driving safety.

[0005] To achieve the above purpose, this application provides an AIoT device control method and system under a cloud-edge collaboration architecture.

[0006] In a first aspect, the present application provides an AIoT device control method under a cloud-edge collaboration architecture. The AIoT device control method under the cloud-edge collaboration architecture is implemented through an AIoT device control system under the cloud-edge collaboration architecture. Among them, the AIoT device control method under the cloud-edge collaboration architecture includes: clustering and dividing the target road according to the night brightness monitoring records of the target road, determining multiple independent road segments, and arranging a brightness perception layer for each independent road segment; monitoring and obtaining the real-time ambient brightness of the multiple independent road segments through the brightness perception layer; monitoring and obtaining the road traffic flow and vehicle images through the data perception layer at a predetermined position, where an edge computing layer is also embedded at the predetermined position; in the edge computing layer, analyzing and obtaining the predicted headlight brightness according to the road traffic flow and vehicle images; optimizing the standard night driving brightness according to the multiple real-time ambient brightness and the predicted headlight brightness to obtain multiple adapted driving brightnesses, and dynamically regulating the street lamp system of the target road.

[0007] Optionally, if no passing vehicles are detected at the predetermined position, the street lamp system of the target road is dynamically regulated according to a predetermined minimum brightness, where the predetermined position is in front of the target road and is separated from the target road by a predetermined distance.

[0008] Optionally, the brightness perception layer, the data perception layer and the edge computing layer are communicatively connected, where a brightness sensor is embedded in the brightness perception layer and a monitoring camera is embedded in the data perception layer.

[0009] Optionally, obtain the night brightness monitoring records of the target road, where the night brightness monitoring records include multiple night brightness monitoring data, and each night brightness monitoring data includes ambient brightness data at multiple predetermined monitoring coordinates, and the adjacent predetermined monitoring coordinates are separated by a predetermined length; divide the multiple night brightness monitoring data according to multiple predetermined moonlight intensity intervals to determine multiple sample brightness monitoring data sets; cluster and divide the target road according to the multiple sample brightness monitoring data sets respectively to determine multiple road segment division results; fuse and cluster the multiple road segment division results in sequence to obtain multiple independent road segments.

[0010] Optionally, divide the target road according to the predetermined length to determine multiple sub-sections, and sequentially select a first sub-section and a second sub-section; randomly select a first sample brightness monitoring data set, and respectively extract the first brightness data set and the second brightness data set of the first sub-section and the second sub-section, and perform deviation calculation to obtain a first brightness deviation set; if all the first brightness deviation sets are less than a predetermined deviation scalar, then merge the first sub-section and the second sub-section and set it as a first aggregated section, where the brightness data of the aggregated section is the mean value of the brightness data of several sub-sections; if there is data greater than or equal to the predetermined deviation scalar in the first brightness deviation set, then set the first sub-section as the first aggregated section, and start iterative clustering with the second sub-section as the starting point until all sub-sections are traversed to obtain a first road section division result, and add it to the multiple road section division results.

[0011] Optionally, sequentially obtain multiple first aggregated sections in the multiple road section division results, where each first aggregated section includes several first sub-sections; respectively count the frequencies of the first sub-sections appearing in the multiple first aggregated sections, and add those with frequencies greater than a predetermined threshold to the first independent section, and add those with frequencies less than or equal to the predetermined threshold to the second independent section; start fusion clustering with the second independent section as the starting point to obtain multiple independent sections.

[0012] Optionally, at the edge computing layer, use a vehicle recognition plugin to perform model recognition based on the vehicle image to obtain the vehicle model, where the vehicle recognition plugin is constructed based on a convolutional neural network, trained at the cloud computing layer, and deployed to the edge computing layer after convergence; match the headlight information in the database based on the vehicle model to obtain the vehicle high beam brightness and the vehicle low beam brightness; at the edge computing layer, use a headlight state prediction plugin to respectively predict and output the headlight state probabilities of the multiple independent sections according to the current time node and the road traffic flow, where the headlight state probabilities include the low beam probability and the high beam probability; calculate and obtain multiple predicted headlight brightnesses according to the vehicle high beam brightness, the vehicle low beam brightness, the multiple low beam probabilities and the multiple high beam probabilities.

[0013] Optionally, at the cloud computing layer, according to the historical monitoring data of the first independent section, collect a sample time node set and a sample traffic flow set, and mark the proportions of the vehicle using high and low beams under different sample time nodes and sample traffic flows to obtain a sample low beam probability set and a sample high beam probability set; use the sample time node set, the sample traffic flow set, the sample low beam probability set and the sample high beam probability set to train a feedforward neural network until convergence to obtain a first headlight state prediction branch; sequentially analyze and obtain the headlight state prediction branches of the multiple independent sections, map and construct the headlight state prediction plugin, and deploy it to the edge computing layer.

[0014] Optionally, if the low beam probability or the high beam probability is greater than a predetermined probability threshold, mark the low beam probability or the high beam probability as 1, where the predetermined probability threshold is 75%; perform weighted calculation based on the vehicle high beam brightness, vehicle low beam brightness, multiple low beam probabilities, and multiple high beam probabilities, and output multiple predicted headlight brightnesses.

[0015] In a second aspect, the present application further provides an AIoT device control system under a cloud-edge collaboration architecture for executing the AIoT device control method under the cloud-edge collaboration architecture as described in the first aspect. The AIoT device control system under the cloud-edge collaboration architecture includes: a clustering and partitioning module for clustering and partitioning the target road according to the night brightness monitoring records of the target road, determining multiple independent road segments, and deploying a brightness perception layer for each independent road segment; a brightness monitoring module for monitoring and obtaining the multiple real-time ambient brightnesses of the multiple independent road segments through the brightness perception layer; a road condition monitoring module for monitoring and obtaining the road traffic flow and vehicle images through the data perception layer at a predetermined position, where an edge computing layer is also embedded in the predetermined position; a headlight brightness prediction module for analyzing and obtaining multiple predicted headlight brightnesses according to the road traffic flow and vehicle images in the edge computing layer; and a brightness regulation module for optimizing the standard night driving brightness according to the multiple real-time ambient brightnesses and multiple predicted headlight brightnesses to obtain multiple adapted driving brightnesses, and dynamically regulating the street lamp system of the target road.

[0016] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0017] By clustering and partitioning the target road according to the night brightness monitoring records of the target road, determining multiple independent road segments, and deploying a brightness perception layer for each independent road segment; monitoring and obtaining the multiple real-time ambient brightnesses of the multiple independent road segments through the brightness perception layer; monitoring and obtaining the road traffic flow and vehicle images through the data perception layer at a predetermined position, where an edge computing layer is also embedded in the predetermined position; analyzing and obtaining the predicted headlight brightness according to the road traffic flow and vehicle images in the edge computing layer; optimizing the standard night driving brightness according to the multiple real-time ambient brightnesses and predicted headlight brightnesses to obtain multiple adapted driving brightnesses, and dynamically regulating the street lamp system of the target road. That is to say, the target road is divided into multiple independent road segments according to the night brightness monitoring records, a brightness perception layer is deployed respectively to obtain the real-time ambient brightness of each road segment, the headlight brightness is predicted according to the road traffic flow and vehicle images, and the street lamp system is dynamically regulated through the real-time ambient brightness and predicted headlight brightness, which not only avoids over-illumination but also ensures sufficient illumination, enhancing driving safety while improving energy efficiency.

[0018] The above description is only an overview of the technical solution of this application. In order to understand the technical means of this application more clearly, it can be implemented according to the content of the specification. In order to make the above and other purposes, features, and advantages of this application more obvious and understandable, the following specific embodiments of this application are specifically given. It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of this application, nor is it used to limit the scope of this application. Other features of this application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in this application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only exemplary, and for those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0020] Figure 1 It is a schematic flowchart of the AIoT device control method under the cloud-edge collaboration architecture of this application;

[0021] Figure 2 It is a schematic structural diagram of the AIoT device control system under the cloud-edge collaboration architecture of this application.

[0022] Description of the reference numerals: clustering and partitioning module 11, brightness monitoring module 12, road condition monitoring module 13, vehicle headlight brightness prediction module 14, brightness regulation module 15. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0023] By providing an AIoT device control method and system under the cloud-edge collaboration architecture, this application solves the technical problem in the prior art that due to the inability of AIoT street lamp control to dynamically adjust the brightness according to real-time traffic and environmental conditions, it leads to over-illumination or under-illumination of street lamps, further affecting the energy utilization efficiency and driving safety. The target road is divided into multiple independent sections according to the night brightness monitoring records, and a brightness perception layer is respectively arranged to obtain the real-time environmental brightness of each section. The vehicle headlight brightness is predicted based on the road traffic flow and vehicle images, and the street lamp system is dynamically regulated through the real-time environmental brightness and the predicted vehicle headlight brightness, which not only avoids over-illumination but also ensures sufficient illumination, enhancing driving safety while improving energy efficiency.

[0024] Next, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the example embodiments described herein. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts belong to the scope of protection of the present application. Additionally, it should be noted that for the sake of description, only the parts related to the present application are shown in the accompanying drawings rather than all of them.

[0025] Example 1. Please refer to the attached Figure 1 , the present application provides an AIoT device control method under a cloud-edge collaboration architecture. Among them, the AIoT device control method under the cloud-edge collaboration architecture is applied to an AIoT device control system under the cloud-edge collaboration architecture. The AIoT device control method under the cloud-edge collaboration architecture specifically includes the following steps:

[0026] S100: According to the nighttime brightness monitoring records of the target road, cluster and divide the target road to determine multiple independent road segments, and deploy a brightness perception layer for each independent road segment.

[0027] Furthermore, S100 of the present application includes:

[0028] Obtain the nighttime brightness monitoring records of the target road, where the nighttime brightness monitoring records include multiple nighttime brightness monitoring data. Each nighttime brightness monitoring data includes environmental brightness data at multiple predetermined monitoring coordinates, and the adjacent predetermined monitoring coordinates are spaced at a predetermined length; divide the multiple nighttime brightness monitoring data according to multiple predetermined moonlight intensity intervals to determine multiple sample brightness monitoring data sets; according to the multiple sample brightness monitoring data sets, cluster and divide the target road respectively to determine multiple road segment division results; perform fusion clustering on the multiple road segment division results in sequence to obtain multiple independent road segments.

[0029] Furthermore, the present application also includes the following steps:

[0030] Divide the target road according to the predetermined length, determine multiple sub-sections, and sequentially select the first sub-section and the second sub-section; randomly select the first sample brightness monitoring data set, and respectively extract the first brightness data set and the second brightness data set of the first sub-section and the second sub-section, and calculate the deviation to obtain the first brightness deviation set; if all the data in the first brightness deviation set are less than the predetermined deviation scalar, then merge the first sub-section and the second sub-section, and set it as the first aggregated section, where the brightness data of the aggregated section is the mean value of the brightness data of several sub-sections; if there is data greater than or equal to the predetermined deviation scalar in the first brightness deviation set, then set the first sub-section as the first aggregated section, and start iterative clustering with the second sub-section as the starting point until all sub-sections are traversed, obtain the first road section division result, and add it to the multiple road section division results.

[0031] Further, the present application further includes the following steps:

[0032] Sequentially obtain multiple first aggregated sections in the multiple road section division results, where each first aggregated section includes several first sub-sections; respectively count the frequencies of the first sub-sections appearing in the multiple first aggregated sections, and add those with frequencies greater than the predetermined threshold to the first independent section, and add those with frequencies less than or equal to the predetermined threshold to the second independent section; start fusing and clustering with the second independent section as the starting point to obtain multiple independent sections.

[0033] Specifically, monitor the target road at night through a photosensitive sensor, a camera or other light sensing devices to obtain a night brightness monitoring record. The night brightness monitoring record includes multiple night brightness monitoring data, and each data includes the ambient brightness values at multiple predetermined monitoring coordinates. The ambient brightness data is the brightness value measured by the sensor at the predetermined coordinate position, usually in lux (Lux), indicating the luminous flux received per unit area. The predetermined monitoring coordinate is the measurement point predetermined when the sensor or device is installed, and is the position for collecting ambient data. The adjacent predetermined monitoring coordinates are spaced by a predetermined length, which means that there is a predetermined length between two adjacent predetermined monitoring coordinates, and this predetermined length is the distance set between two adjacent monitoring points.

[0034] The predetermined moonlight intensity interval refers to multiple brightness intervals divided according to different moonlight brightness conditions (usually affected by factors such as weather, lunar phase, season, etc.). Different moonlight intensities will affect the brightness on the ground, so the influence of moonlight must be considered when conducting night brightness monitoring. The moonlight intensity interval divides different moonlight intensities into several intervals, and common intervals may include low intensity, medium intensity, and high intensity. For example, low intensity may be from 0 to 50 lux, medium intensity is from 51 to 100 lux, and high intensity is from 101 to 150 lux.

[0035] Divide multiple pieces of nighttime brightness monitoring data according to multiple predetermined moonlight intensity intervals to obtain multiple sample brightness monitoring data sets. Each sample brightness monitoring data set represents the brightness data within a specific moonlight intensity interval. For example, divide the moonlight intensity from 0 to 100 lux into three intervals: low, medium, and high. If the moonlight intensity is 50 lux during certain periods, these data can be classified into the medium-intensity moonlight interval; for periods with weaker moonlight intensity, they can be classified into the low-intensity moonlight interval.

[0036] Divide the target road into multiple sub-sections according to a predetermined length. The predetermined length is a length preset according to factors such as the overall length of the road, traffic demand, lighting requirements, etc., and may be 100 meters, 200 meters, or 500 meters, etc. The setting of the predetermined length needs to balance the division accuracy of the section and the complexity of control. Divide the target road into multiple sub-sections according to the total length of the target road and the predetermined length. Among the multiple divided sub-sections, select the first sub-section and the second sub-section in sequence. That is to say, starting from the starting point, select section by section according to the actual order of the road. For example, when the preset length is 200 meters, the first sub-section is from the starting point of the target road to the position of the first 200 meters, and the second sub-section is from the 200th meter to the 400th meter.

[0037] Randomly select the first sample brightness monitoring data set from multiple sample brightness monitoring data sets, which contains the brightness monitoring data under a certain moonlight intensity. Match the first brightness data set corresponding to the first sub-section and the second brightness data set corresponding to the second sub-section from the first sample brightness monitoring data set, and calculate the deviation between the two data sets to obtain the first brightness deviation set. Deviation calculation is used to measure the difference in brightness data between two sub-sections, and it is carried out by comparing the brightness values of the two data sets.

[0038] The first brightness deviation set indicates whether the brightness data of the first sub-section and the second sub-section are consistent or close. If the deviation between the brightness values of the two sub-sections is small, it means that the brightness conditions of the two sub-sections are similar and can be combined into one section for unified regulation. If the deviation is large, it means that the brightness changes of the two sub-sections are large and may need to be managed separately. For example, the brightness data of the first sub-section is: [20, 25, 22, 18, 23] (unit: lux), and the brightness data of the second sub-section is: [35, 38, 32, 36, 37]. Then the difference in brightness data between the first sub-section and the second sub-section is the deviation, which is [15, 13, 10, 18, 14].

[0039] The predetermined deviation scalar is a preset threshold used to determine whether the brightness of two sub - road segments is similar enough. If the deviation is less than this scalar, it is considered that the brightness data of these two sub - road segments are close and can be merged into one road segment. If all values in the deviation set are less than the predetermined deviation scalar, it is considered that the brightness of these two sub - road segments is very similar and can be merged into one aggregated road segment. The brightness data of this aggregated road segment is obtained by averaging the brightness data of several aggregated sub - road segments.

[0040] If there is data in the first brightness deviation set that is greater than or equal to the predetermined deviation scalar, that is, at least one deviation value is greater than or equal to the predetermined deviation scalar, they will not be merged. Instead, the current first sub - road segment will be kept as the aggregated road segment, and the clustering judgment will continue for the next sub - road segment. The clustering operation will continue for the remaining sub - road segments. First, keep the current sub - road segment as the basis of the aggregated road segment, then calculate the deviation with the next sub - road segment and determine whether to merge. This process will continue until all sub - road segments have been traversed and clustered.

[0041] Gradually compare the brightness differences between every two sub - road segments in order to determine whether clustering can be performed. If the brightness difference between two sub - road segments meets the conditions, they will be merged. If the deviation is greater than the predetermined scalar, set the previous road segment as the aggregated road segment and continue to use the subsequent sub - road segments as new starting points, repeating this process until all sub - road segments have been traversed. Through iterative clustering, the clustering division result of the first sample brightness monitoring data set, that is, the first road segment division result, is finally obtained, which shows the aggregated grouping of multiple sub - road segments in the target road. Each aggregated road segment contains multiple sub - road segments, and the brightness differences between these sub - road segments are small, making them suitable for adjustment under unified control.

[0042] Perform the above steps for all other sample brightness monitoring data sets in the multiple sample brightness monitoring data sets to obtain multiple road segment division results after clustering the multiple sample brightness monitoring data sets.

[0043] Obtain multiple first aggregated road segments in order from the multiple road segment division results. These are composed of multiple sub - road segments. The sub - road segments within each aggregated road segment have small brightness differences, and the complexity of street - lamp adjustment can be reduced by merging sub - road segments. Each first aggregated road segment includes one or more first sub - road segments. Perform frequency statistics on each sub - road segment. The frequency refers to the number of times a certain sub - road segment appears in different aggregated road segments, which helps to understand which sub - road segments appear frequently in multiple aggregated road segments, meaning that these road segments may be more representative and need to be given more attention. For example, if sub - road segment A appears multiple times in multiple aggregated road segments, then it may be a traffic - intensive area or a main road segment, so it should be controlled preferentially.

[0044] Count the frequency of occurrence of the first sub - section in each first aggregated section. If the frequency of occurrence in the first aggregated section is greater than a predetermined threshold, cluster the corresponding road section division result into a first independent section. Otherwise, cluster the corresponding road section division results with a lower frequency of occurrence of the first sub - section in each first aggregated section into a second independent section. The predetermined threshold is a criterion for determining which sub - sections are more important. Use the second independent section as the starting point and repeat the above steps for clustering until multiple independent sections are obtained.

[0045] Cluster - divide the target road to obtain multiple independent sections, and deploy a brightness perception layer for each independent section. Select appropriate brightness sensors according to the characteristics of the road sections, and install brightness sensors at key positions in each independent section, such as the starting point, ending point, intersection, etc. The brightness data of each independent section has high consistency, enabling these road sections to use similar control strategies.

[0046] Identify important road sections through the methods of step - by - step clustering, frequency statistics, and fusion clustering. Achieve precise road section control through clustering division and the deployment of the brightness perception layer. Each independent section will be adjusted according to specific brightness data, thus ensuring that the street - lamp system can adapt to different traffic flows and ambient light conditions.

[0047] S200: Monitor and obtain the multiple real - time ambient brightnesses of the multiple independent sections through the brightness perception layer.

[0048] Specifically, through the brightness perception layer, perform real - time monitoring on multiple independent sections to obtain the multiple real - time ambient brightnesses corresponding to the multiple independent sections, which reflect the light intensities in different areas of the road. For example, areas with heavy traffic may have higher brightness due to the headlights of vehicles and street lamps, while in sections with less or no traffic, the brightness may be lower. Through the monitoring of the brightness perception layer, obtain the ambient brightness data of each independent section in real - time, thereby achieving more precise street - lamp control, enabling the street - lamp system to make dynamic adjustments according to the actual brightness requirements of each road section, and avoiding over - illumination or under - illumination phenomena in traditional street - lamp systems. Especially in areas with dense traffic or special weather conditions, timely increase the illumination brightness to enhance road safety; while in areas with relatively smooth traffic, the brightness can be reduced to save energy.

[0049] S300: Monitor and obtain the road traffic flow and vehicle images through the data perception layer at a predetermined position, where the predetermined position also embeds an edge computing layer.

[0050] The brightness perception layer, the data perception layer are communicatively connected to the edge computing layer, where the brightness perception layer embeds brightness sensors, and the data perception layer embeds monitoring cameras.

[0051] Specifically, monitoring is carried out through the data perception layer at a predetermined location to obtain the road traffic flow and vehicle images at the predetermined location. The predetermined location is a selected location for monitoring the passing vehicles, such as urban intersections, busy sections, etc., where a data perception layer, an edge computing layer, a brightness perception layer, etc. are deployed. The traffic flow refers to the number of vehicles passing through a certain location per unit time. The vehicle image is the vehicle image data obtained through a monitoring camera, which is usually used for license plate recognition, traffic condition analysis, etc. Through the traffic flow and vehicle images, the traffic condition can be understood in real time, and then the brightness adjustment of the street lights can be optimized.

[0052] The brightness perception layer is a layer specifically used for monitoring the road environment illumination. It is embedded with brightness sensors that can measure the ambient brightness (in lux) of each area on the road in real time. The brightness data collected by these sensors will help determine whether the street light brightness needs to be adjusted to ensure that the road lighting matches the actual environmental requirements.

[0053] The data perception layer is a layer specifically used for monitoring traffic flow and vehicle behavior. It is embedded with monitoring cameras that provide necessary traffic flow data by capturing the traffic flow conditions, vehicle numbers, and vehicle speeds on the road surface in real time. Through image analysis technology, these cameras can not only identify vehicles but also analyze the traffic conditions of the section, such as congestion, vehicle speed, etc., further providing a decision-making basis for AIoT devices.

[0054] The edge computing layer refers to the technology of data processing and calculation near the data acquisition source (such as the perception layer on the road). It quickly processes the data from the brightness perception layer and the data perception layer without transmitting the data to a remote cloud server for processing. Edge computing can greatly reduce the response time and bandwidth requirements. The data processed includes traffic flow, vehicle images, ambient brightness, etc., and the street light brightness is dynamically adjusted according to these data to provide real-time traffic response.

[0055] The brightness perception layer and the data perception layer respectively monitor the target road through the embedded brightness sensors and monitoring cameras. The monitoring data is transmitted to the edge computing layer, and the edge computing layer determines whether the street lights need to be adjusted by real-time processing and analysis of these data. For example, if it is found that the traffic flow is high or the brightness of a certain area is insufficient, the edge computing layer will immediately give an instruction to adjust the street light brightness according to this information. Moreover, since the data processing occurs at the edge computing layer, it can quickly respond to changes and avoid problems caused by transmission delays.

[0056] S400: At the edge computing layer, multiple predicted headlight brightnesses are analyzed and obtained based on the road traffic flow and vehicle images.

[0057] Furthermore, S400 of this application includes:

[0058] At the edge computing layer, a vehicle recognition plugin is used to identify the vehicle model based on the vehicle image and obtain the vehicle model. Among them, the vehicle recognition plugin is constructed based on a convolutional neural network, trained at the cloud computing layer, and deployed to the edge computing layer after convergence; the headlight information is matched in the database based on the vehicle model to obtain the high beam brightness and low beam brightness of the vehicle; at the edge computing layer, a headlight state prediction plugin is used to predict and output the headlight state probabilities of multiple independent road segments according to the current time node and the road traffic flow respectively, where the headlight state probabilities include low beam probabilities and high beam probabilities; according to the high beam brightness, low beam brightness, multiple low beam probabilities and multiple high beam probabilities of the vehicle, multiple predicted headlight brightnesses are calculated.

[0059] Specifically, the road traffic flow and vehicle images monitored by the data perception layer are transmitted to the edge computing layer. At the cloud computing layer, the vehicle recognition plugin is trained and deployed to the edge computing layer to determine the vehicle model according to the vehicle image. The vehicle recognition plugin is a computer vision-based tool that can identify the vehicle model by processing the vehicle image.

[0060] The vehicle recognition plugin is a model constructed based on a convolutional neural network and can identify the vehicle model by analyzing the vehicle image. Key information is extracted from the vehicle images obtained from cameras or sensors to identify the type or brand of the vehicle, such as identifying Audi A6 or Toyota Corolla, etc. The convolutional neural network automatically extracts features (such as edges, shapes, colors, etc.) in the image through convolutional operations and can identify the objects in the image.

[0061] Vehicle models of various types and their corresponding vehicle images from various angles are obtained from public transportation databases, images collected by surveillance cameras, automobile manufacturers, etc., and the specific vehicle models are marked on each vehicle image. The images in the dataset should include vehicle photos at different time periods (especially at night), ensuring sufficient light changes in the images (such as the influence of high beam lights, low beam lights, street lighting, etc.). The images should include diverse weather conditions (such as rainy days, foggy days, etc.), as well as different traffic environments (such as dense traffic flow, different vehicle speeds, etc.). For example, obtain an existing dataset of bus images and perform data expansion in combination with actual surveillance videos. Mark the vehicle models for each image to ensure the accuracy of the marking. In night scenes, the brightness, contrast, and color saturation of the images can be adjusted to simulate different night conditions and enhance the model's ability to identify vehicles under different lighting conditions.

[0062] Design a convolutional neural network architecture suitable for vehicle recognition tasks, including multiple convolutional layers and pooling layers. The input layer is responsible for receiving vehicle images, usually RGB three-channel images, and the image size can be set to a fixed size, such as 224x224 pixels; the convolutional layer is responsible for extracting local features in the image, such as appearance, head and tail design, body contour, wheel design, texture, color, etc. The activation function is used to increase the non-linear ability of the model, and ReLU is often used as the activation function in convolutional neural networks. The pooling layer (such as the max pooling layer) is used to reduce the size of the feature map, reduce the computational amount, and enhance the robustness of the model. After the convolutional layer extracts features, the fully connected layer maps the features to the target categories (such as specific vehicle models). Finally, the vehicle model category is output.

[0063] Divide the entire dataset into a training set and a validation set. Usually, the training set accounts for 80% of the dataset, and the validation set accounts for 20%. The validation set is used to monitor the performance of the model during training to prevent overfitting. Use the training set to train the constructed convolutional neural network architecture, calculate the output value through forward propagation, and compare it with the actual label to calculate the loss. Use the cross-entropy loss function to measure the gap between the model prediction and the actual label. The training objective is to minimize the loss function and improve the accuracy of the model. During the training process, after each round (epoch) of training, the model is evaluated on the validation set once, and its accuracy or other performance metrics are calculated. Calculate the gradients and update the weights of the model through the backpropagation algorithm. This process is repeated for multiple rounds (epochs) until the loss function converges and the model can show good accuracy on both the training set and the validation set. During the training process, according to the loss calculated by the loss function, continuously adjust the parameters of the model, such as increasing the number of network layers, using a more complex convolutional neural network structure, and performing more data augmentation. After training, use the test set to perform the final evaluation of the model. Evaluate the classification effect of the model on different vehicle models by calculating metrics such as accuracy, precision, recall, and F1 score. When the classification accuracy of the model reaches 95%, stop training, and use the model obtained at this time as the vehicle recognition plugin.

[0064] This process is usually trained in the cloud computing layer using a large amount of labeled vehicle image data. Through these images, the convolutional neural network can gradually learn how to recognize various features of vehicles. The goal of training is to optimize the weights and parameters in the network until the accuracy of the model reaches a predetermined standard. When the convolutional neural network model converges, that is, when the training results reach satisfactory accuracy, the trained vehicle recognition plugin is deployed from the cloud computing layer to the edge computing layer. The plugin deployed to the edge computing layer will no longer rely on the remote server but directly process the real-time vehicle image data collected from the camera locally. For example, assume that a surveillance camera on a section of the road captures an image of a car. After being processed by the vehicle recognition plugin in the edge computing layer, the model quickly identifies the vehicle model (such as Audi A6). The execution of the model in the edge computing layer avoids the latency of transmitting image data to the remote server and improves the response speed.

[0065] According to the recognized vehicle model, a match is made in the database to obtain the corresponding headlight information, including the high beam brightness and low beam brightness of the vehicle. The database is a structured data storage system that contains relevant information about vehicles and stores detailed information corresponding to each vehicle model, including data such as headlight type, high beam brightness, and low beam brightness. These information are usually defined in the vehicle specifications provided by the vehicle manufacturer, and the brightness values may vary for different vehicle models and lighting systems. For example, the vehicle lighting information is shown in Table 1:

[0066] Table 1 Vehicle Lighting Information Table

[0067]

[0068] Surveillance cameras usually capture high-quality videos to provide real-time traffic flow data. The video data captured from the surveillance camera needs to undergo video analysis to extract useful information, especially the number of vehicles on the road. The video stream is composed of continuous frames, and each frame is a static image that captures the road conditions at a certain moment. In video analysis, the video stream is decomposed into individual image frames, and then each frame is processed. In each frame image, vehicles need to be recognized and marked through object detection algorithms (such as the YOLO algorithm). By learning vehicles of different shapes, colors, and features, the position of the vehicle is located in each frame image. By detecting the vehicles in each frame, the traffic flow is calculated. The traffic flow refers to the number of vehicles passing through a specific section per unit time. The number of detected vehicles is the number of vehicles in that frame. Through the number of vehicles in several consecutive frames, the vehicle flow within a certain period of time can be calculated. Based on the vehicle detection results in each frame, the traffic flow per unit time (such as per minute or per hour) is estimated. For example, if 200 vehicles pass through this section per minute, the output traffic flow is 200 vehicles / minute.

[0069] At the edge computing layer, the constructed headlight status prediction plugin is used to predict the headlight status probability of each independent road section according to the current time node and the traffic flow of the road, that is, the probabilities of vehicles using low beams and high beams at the current time node. Since the traffic flow and time node change in real time, the prediction plugin in the edge computing layer can continuously update the prediction results according to new data.

[0070] Based on the high beam brightness and low beam brightness obtained from the database, as well as the low beam probability and high beam probability obtained from the headlight status prediction plugin, the headlight brightness of multiple road sections is calculated. According to the brightness of the vehicle's high beam and low beam and the predicted usage probability, if the predicted usage probability of the low beam or high beam is greater than 75%, the predicted headlight brightness is determined according to the low beam or high beam. Otherwise, it is obtained by weighting according to the usage probabilities of the low beam and high beam. By analyzing the road traffic flow and vehicle images, the headlight brightness is predicted. Based on the predicted headlight usage pattern, the street light brightness is reduced during periods with low traffic flow or a high proportion of high beam usage, thereby saving energy.

[0071] Furthermore, the present application further includes the following steps:

[0072] At the cloud computing layer, according to the historical monitoring data of the first independent road section, a sample time node set and a sample traffic flow set are collected, and the proportions of vehicles using high beams and low beams at different sample time nodes and sample traffic flows are marked to obtain a sample low beam probability set and a sample high beam probability set; the sample time node set, the sample traffic flow set, the sample low beam probability set, and the sample high beam probability set are used to train a feedforward neural network until convergence to obtain a first headlight status prediction branch; the multiple headlight status prediction branches of multiple independent road sections are analyzed and obtained in sequence, and the headlight status prediction plugin is mapped and constructed and then deployed to the edge computing layer.

[0073] Specifically, at the cloud computing layer, according to the historical monitoring data of the first independent road section, including the traffic flow data, time node data, and the status of vehicles using high beams and low beams on this road section. The sample time node set is the time node data extracted from the historical monitoring data, and each time node represents a specific moment or time period. The sample traffic flow set includes the traffic flow data corresponding to multiple sample time nodes, indicating the number of vehicles passing through the road section during this time period. The traffic flow is related to the headlight usage pattern, and when the traffic flow is different, the proportions of vehicles using high beams and low beams may be different.

[0074] Use a sample time node set, a sample traffic flow set, a sample low beam probability set, and a sample high beam probability set to train a feedforward neural network. Preprocess the data set to remove or correct errors and outliers in the data set to ensure data quality. Standardize the data set to avoid the influence of different dimensions on the model training effect. Split the data set into a training set, a validation set, and a test set. The training set is used to train the model, the validation set is used to adjust the model parameters, and the test set is used to evaluate the final performance of the model.

[0075] A feedforward neural network is a classic type of artificial neural network where the information flow is unidirectional, from the input layer to the output layer, without feedback. The number of nodes in the input layer is equal to the number of features in the sample time node set and the sample traffic flow set, and the number of nodes in the output layer is equal to the dimension of the headlight state probability. Before starting training, it is necessary to initialize the weights and bias terms of the network. Usually, small random numbers are used for initialization. During training, the input data will propagate forward through the neural network. The output of each layer will be used as the input of the next layer. Select a loss function to quantify the difference between the predicted value and the actual value, such as cross-entropy loss. According to the error calculated by the loss function, backpropagate through the network to update the weights and bias of the network. Use an optimization algorithm (such as gradient descent, etc.) to adjust the weights of the network to minimize the loss function. Repeat the process of forward propagation and backpropagation, and the weights of the network will be updated in each iteration until the model performance reaches a certain threshold or the training reaches the preset number of iterations. Use the validation set to evaluate the performance of the model and check for overfitting or underfitting. According to the performance of the validation set, adjust the hyperparameters of the network, such as the learning rate, the size of the hidden layer, the number of layers, etc. When the performance of the model on the validation set is satisfactory, stop training to obtain the first headlight state prediction branch.

[0076] Repeat the above steps for other independent road segments to obtain multiple headlight state prediction branches. Integrate the headlight state prediction branches of all independent road segments into a headlight state prediction plug-in, integrate the headlight prediction models of all road segments, and intelligently predict the headlight usage status of each road segment according to real-time traffic flow, time node, and other information. The headlight state prediction plug-in is deployed to the edge computing layer. The edge computing layer is located close to the data sources (such as surveillance cameras and sensors), with low latency and fast response capabilities. At the edge computing layer, the traffic flow and time node data of each road segment can be obtained in real time, and decisions can be quickly made through the prediction plug-in to automatically adjust the brightness of the streetlights.

[0077] Furthermore, the present application further includes the following steps:

[0078] If the low beam probability or the high beam probability is greater than a predetermined probability threshold, mark the low beam probability or the high beam probability as 1, where the predetermined probability threshold is 75%; perform weighted calculation based on the vehicle high beam brightness, vehicle low beam brightness, multiple low beam probabilities, and multiple high beam probabilities, and output multiple predicted headlight brightnesses.

[0079] Specifically, the predetermined probability threshold is a set probability threshold, which is 75%, and is used to determine whether the vehicle mainly uses low beam headlights or high beam headlights. That is to say, when the low beam probability or the high beam probability is greater than or equal to 75%, it is considered that vehicles on this section of the road almost entirely use low beam headlights or high beam headlights, thus simplifying to choose to use one type of light. At this time, select the headlight type marked as 1 and directly use the corresponding headlight brightness.

[0080] If the probability of the low beam headlight or the high beam headlight is greater than the predetermined threshold, the headlight type is marked as 1, indicating that almost all vehicles use this headlight during the current period. Conversely, when it is less than the predetermined probability threshold, weight according to the probability, such as 70% * high beam brightness + 30% * low beam brightness, to obtain multiple predicted headlight brightnesses.

[0081] By dynamically adjusting the street lamp brightness according to the real-time traffic flow and the headlight usage probability, while ensuring safety, unnecessary lighting waste is avoided.

[0082] S500: Optimize the standard night driving brightness according to the multiple real-time ambient brightnesses and multiple predicted headlight brightnesses to obtain multiple adapted driving brightnesses, and perform dynamic regulation on the street lamp system of the target road.

[0083] Specifically, optimize the standard night driving brightness according to the multiple real-time ambient brightnesses and multiple predicted headlight brightnesses, so as to achieve dynamic regulation of the street lamp system of the target road. The real-time ambient brightness refers to the actual light intensity on the road, which is usually monitored by installed brightness sensors or surveillance cameras. The ambient brightness can vary due to various factors, such as whether there are street lamps on the road, whether there is natural light (such as moonlight) on the street, weather conditions, etc.

[0084] The predicted headlight brightness calculates the possible impact of the vehicle's headlight brightness (low beam and high beam) on road lighting at a specific moment and section through factors such as traffic flow, time node, vehicle model, and headlight usage probability.

[0085] The standard night driving brightness refers to the brightness level that ensures sufficient visibility for the driver during night driving, and is usually set according to factors such as road type, traffic density, and ambient light. The goal of the standard driving brightness is to ensure that the driver can clearly see the road ahead and obstacles, while avoiding energy waste and the impact on the surrounding environment caused by over-illumination.

[0086] Determine the required minimum safe brightness according to the standard driving brightness. Optimize and calculate the standard night driving brightness based on multiple real-time ambient brightnesses and multiple predicted headlight brightnesses to obtain multiple adapted driving brightness values. The actual brightness of each road section must meet the requirements of the minimum standard driving brightness. The street lamp brightness cannot exceed a preset maximum value to prevent energy waste. During periods with low traffic volume, the headlight brightness and street lamp brightness can be appropriately reduced. Taking these as constraints, continuously adjust the standard night driving brightness and calculate the fitness value corresponding to each brightness, and select the one with the highest fitness value as the adapted driving brightness, so as to obtain multiple adapted driving brightnesses for multiple independent road sections, and dynamically regulate the street lamp system of the target road accordingly.

[0087] Automatically adjust the brightness of the street lamps according to the calculated adapted driving brightness to ensure that the lighting on the road not only meets the standard driving brightness but also saves energy. For example, if the real-time ambient brightness and headlight brightness are already close to the standard driving brightness, automatically reduce the street lamp brightness; while in the case of high traffic volume or low headlight brightness, the street lamp brightness will be increased.

[0088] Furthermore, this application also includes the following steps:

[0089] If no passing vehicle is detected at the predetermined position, dynamically regulate the street lamp system of the target road according to the predetermined minimum brightness, where the predetermined position is in front of the target road and is separated from the target road by a predetermined distance.

[0090] Specifically, if no passing vehicle is detected at the predetermined position, that is, there is no vehicle passing at the current moment, testing and dynamically regulating the street lamp system of the target road according to the predetermined minimum brightness helps to avoid energy waste, especially during periods when no vehicle passes. The predetermined minimum brightness is a set minimum lighting standard, usually to ensure that the road still has the lowest safe lighting when there is no one passing. For example, on a night road, although there is no vehicle passing, a certain brightness will still be maintained to ensure the safety of pedestrians or other unexpected situations.

[0091] The predetermined position is usually set in front of the target road, and may be a place at a certain distance from the target road, such as 300 meters away from the target road, leaving a buffer time for analysis and control, so as to monitor the traffic conditions of the road in advance and take measures in advance. By automatically adjusting to the minimum brightness when no vehicle is detected at the predetermined position, energy is effectively saved and unnecessary lighting is reduced.

[0092] In summary, the AIoT device control method under the cloud-edge collaboration architecture provided by this application has the following technical effects:

[0093] By clustering and partitioning the target road according to the night brightness monitoring records of the target road, determining multiple independent sections, and arranging a brightness perception layer for each independent section; monitoring and obtaining the real-time ambient brightness of the multiple independent sections through the brightness perception layer; monitoring and obtaining the road traffic flow and vehicle images through the data perception layer at a predetermined position, wherein an edge computing layer is also embedded at the predetermined position; in the edge computing layer, analyzing and obtaining the predicted headlight brightness according to the road traffic flow and vehicle images; optimizing the standard night driving brightness according to the multiple real-time ambient brightness and the predicted headlight brightness to obtain multiple adaptive driving brightnesses, and dynamically regulating the street lamp system of the target road. That is to say, the target road is divided into multiple independent sections according to the night brightness monitoring records, a brightness perception layer is arranged respectively, the real-time ambient brightness of each section is obtained, the headlight brightness is predicted according to the road traffic flow and vehicle images, and the street lamp system is dynamically regulated through the real-time ambient brightness and the predicted headlight brightness, which not only avoids over-illumination but also ensures sufficient illumination, enhancing driving safety while improving energy efficiency.

[0094] Embodiment 2. Based on the same inventive concept as the AIoT device control method in the cloud-edge collaborative architecture in the foregoing Embodiment 1, the present application also provides an AIoT device control system in the cloud-edge collaborative architecture. Please refer to the attached Figure 2 , the AIoT device control system in the cloud-edge collaborative architecture includes:

[0095] A clustering and partitioning module 11, configured to cluster and partition the target road according to the night brightness monitoring records of the target road, determine multiple independent sections, and arrange a brightness perception layer for each independent section; a brightness monitoring module 12, configured to monitor and obtain the real-time ambient brightness of the multiple independent sections through the brightness perception layer; a road condition monitoring module 13, configured to monitor and obtain the road traffic flow and vehicle images through the data perception layer at a predetermined position, wherein an edge computing layer is also embedded at the predetermined position; a headlight brightness prediction module 14, configured to analyze and obtain multiple predicted headlight brightnesses in the edge computing layer according to the road traffic flow and vehicle images; a brightness regulation module 15, configured to optimize the standard night driving brightness according to the multiple real-time ambient brightness and the multiple predicted headlight brightnesses to obtain multiple adaptive driving brightnesses, and dynamically regulate the street lamp system of the target road.

[0096] Further, the clustering and partitioning module 11 in the AIoT device control system in the cloud-edge collaborative architecture is further configured to:

[0097] Obtain the night brightness monitoring records of the target road, where the night brightness monitoring records include multiple pieces of night brightness monitoring data, and each piece of night brightness monitoring data includes environmental brightness data at multiple predetermined monitoring coordinates, with a predetermined length interval between adjacent predetermined monitoring coordinates; divide the multiple pieces of night brightness monitoring data according to multiple predetermined moonlight intensity intervals to determine multiple sample brightness monitoring data sets; perform clustering division on the target road according to the multiple sample brightness monitoring data sets respectively to determine multiple road section division results; perform fusion clustering on the multiple road section division results in sequence to obtain multiple independent road sections.

[0098] Further, the clustering division module 11 in the AIoT device control system under the cloud-edge collaborative architecture is further configured to:

[0099] Divide the target road according to the predetermined length to determine multiple sub-road sections, and sequentially select the first sub-road section and the second sub-road section; randomly select the first sample brightness monitoring data set, and respectively extract the first brightness data set and the second brightness data set of the first sub-road section and the second sub-road section, and calculate the deviation to obtain the first brightness deviation set; if all the first brightness deviation sets are less than the predetermined deviation scalar, then merge the first sub-road section and the second sub-road section and set it as the first aggregated road section, where the brightness data of the aggregated road section is the mean value of the brightness data of several sub-road sections; if there is data greater than or equal to the predetermined deviation scalar in the first brightness deviation set, then set the first sub-road section as the first aggregated road section, and start iterative clustering with the second sub-road section as the starting point until all sub-road sections are traversed to obtain the first road section division result, and add it to the multiple road section division results.

[0100] Further, the clustering division module 11 in the AIoT device control system under the cloud-edge collaborative architecture is further configured to:

[0101] Sequentially obtain multiple first aggregated road sections in the multiple road section division results, where each first aggregated road section includes several first sub-road sections; respectively count the frequencies of the first sub-road sections appearing in the multiple first aggregated road sections, and add those with frequencies greater than the predetermined threshold to the first independent road section, and add those with frequencies less than or equal to the predetermined threshold to the second independent road section; start continuous fusion clustering with the second independent road section as the starting point to obtain multiple independent road sections.

[0102] Further, the AIoT device control system under the cloud-edge collaborative architecture further includes a second dynamic regulation module, and the second dynamic regulation module is further configured to:

[0103] If no passing vehicle is detected at the predetermined position, dynamically regulate the street lamp system of the target road according to the predetermined minimum brightness, where the predetermined position is in front of the target road and is separated from the target road by a predetermined distance.

[0104] Further, the road condition monitoring module 13 in the AIoT device control system under the cloud-edge collaboration architecture is further configured to:

[0105] The brightness perception layer, the data perception layer are communicatively connected to the edge computing layer. Among them, the brightness perception layer is embedded with a brightness sensor, and the data perception layer is embedded with a monitoring camera.

[0106] Further, the headlight brightness prediction module 14 in the AIoT device control system under the cloud-edge collaboration architecture is further configured to:

[0107] In the edge computing layer, using the vehicle recognition plugin, perform model recognition based on the vehicle image to obtain the vehicle model. The vehicle recognition plugin is constructed based on a convolutional neural network, trained in the cloud computing layer, and deployed to the edge computing layer after convergence; match the headlight information in the database based on the vehicle model to obtain the vehicle high beam brightness and the vehicle low beam brightness; in the edge computing layer, use the headlight state prediction plugin to predict and output the headlight state probabilities of multiple independent road segments according to the current time node and the road traffic flow respectively. The headlight state probabilities include the low beam probability and the high beam probability; calculate and obtain multiple predicted headlight brightnesses according to the vehicle high beam brightness, the vehicle low beam brightness, multiple low beam probabilities and multiple high beam probabilities.

[0108] Further, the headlight brightness prediction module 14 in the AIoT device control system under the cloud-edge collaboration architecture is further configured to:

[0109] In the cloud computing layer, according to the historical monitoring data of the first independent road segment, collect the sample time node set and the sample traffic flow set, and mark the proportion of the vehicle using high and low beams under different sample time nodes and sample traffic flows to obtain the sample low beam probability set and the sample high beam probability set; use the sample time node set, the sample traffic flow set, the sample low beam probability set and the sample high beam probability set to train the feedforward neural network until convergence to obtain the first headlight state prediction branch; sequentially analyze and obtain the headlight state prediction branches of multiple independent road segments, map and construct the headlight state prediction plugin, and deploy it to the edge computing layer.

[0110] Further, the headlight brightness prediction module 14 in the AIoT device control system under the cloud-edge collaboration architecture is further configured to:

[0111] If the low beam probability or the high beam probability is greater than the predetermined probability threshold, mark the low beam probability or the high beam probability as 1, where the predetermined probability threshold is 75%; perform weighted calculation according to the vehicle high beam brightness, the vehicle low beam brightness, multiple low beam probabilities and multiple high beam probabilities, and output multiple predicted headlight brightnesses.

[0112] The various embodiments in this specification are described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The aforesaid Figure 1 The AIoT device control method and specific examples under the cloud-edge collaboration architecture in the first embodiment are equally applicable to the AIoT device control system under the cloud-edge collaboration architecture of this embodiment. Through the detailed description of the AIoT device control method under the cloud-edge collaboration architecture above, those skilled in the art can clearly understand the AIoT device control system under the cloud-edge collaboration architecture in this embodiment. Therefore, for the sake of brevity of the specification, it will not be elaborated herein.

[0113] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0114] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is also intended to include these changes and modifications.

Claims

1. AIoT device control method under cloud-edge collaborative architecture, characterized in that: include: According to the nighttime brightness monitoring record of the target road, the target road is clustered and divided to determine a plurality of independent road sections, and a brightness perception layer is arranged for each independent road section; Acquire multiple real-time ambient brightness of the multiple independent road sections through brightness perception layer monitoring; The road traffic volume and vehicle images are monitored by a data perception layer at a predetermined location, wherein an edge computing layer is also embedded in the predetermined location; At the edge computing layer, multiple predicted headlight brightness are obtained based on the road traffic volume and vehicle image analysis, including: At the edge computing layer, a vehicle identification plug-in is used to identify the vehicle model according to the vehicle image to obtain the vehicle model, wherein the vehicle identification plug-in is constructed based on a convolutional neural network, trained at the cloud computing layer, and transferred to the edge computing layer after convergence; Matching the vehicle light information in the database based on the vehicle model to obtain the vehicle high beam brightness and the vehicle low beam brightness; At the edge computing layer, a vehicle light state prediction plug-in is used to predict and output multiple vehicle light state probabilities of the multiple independent road sections according to the current time node and the road traffic volume, wherein the vehicle light state probability includes a low beam probability and a high beam probability; Calculate and obtain multiple predicted vehicle light brightnesses according to the vehicle high beam brightness, the vehicle low beam brightness, the multiple low beam probabilities, and the multiple high beam probabilities; According to the multiple real-time ambient brightnesses and the multiple predicted headlight brightnesses, the standard driving brightness at night is optimized to obtain multiple adaptive driving brightnesses, and the streetlight system of the target road is dynamically adjusted.

2. According to the AIoT device control method under the cloud-edge collaborative architecture of claim 1, it is characterized in that: If no passing vehicle is detected at the predetermined position, the street light system of the target road is dynamically adjusted according to the predetermined minimum brightness, wherein the predetermined position is in front of the target road and is separated from the target road by a predetermined distance.

3. The AIoT device control method under the cloud-edge collaborative architecture according to claim 2 is characterized in that: The brightness perception layer and the data perception layer are communicatively connected with the edge computing layer, wherein the brightness perception layer is embedded with a brightness sensor, and the data perception layer is embedded with a surveillance camera.

4. The AIoT device control method under the cloud-edge collaborative architecture according to claim 1 is characterized in that: According to the nighttime brightness monitoring records of the target road, the target road is clustered and divided to determine multiple independent road sections, including: Acquire a nighttime brightness monitoring record of a target road, wherein the nighttime brightness monitoring record includes a plurality of nighttime brightness monitoring data, each of which includes ambient brightness data at a plurality of predetermined monitoring coordinates, and adjacent predetermined monitoring coordinates are spaced at a predetermined length; Dividing the plurality of nighttime brightness monitoring data according to a plurality of predetermined moonlight intensity intervals to determine a plurality of sample brightness monitoring data sets; According to the multiple sample brightness monitoring data sets, clustering the target roads respectively to determine multiple road segment division results; The multiple road segment division results are sequentially fused and clustered to obtain multiple independent road segments.

5. The AIoT device control method under the cloud-edge collaborative architecture according to claim 4 is characterized in that: According to the multiple sample brightness monitoring data sets, the target roads are clustered and divided respectively to determine multiple road segment division results, including: Dividing the target road according to the predetermined length to determine a plurality of sub-segments, and selecting a first sub-segment and a second sub-segment in sequence; Randomly selecting a first sample brightness monitoring data set, and extracting a first brightness data set and a second brightness data set of the first sub-section and the second sub-section respectively, and performing deviation calculation to obtain a first brightness deviation set; If the first brightness deviation set is smaller than a predetermined deviation scalar, the first sub-segment and the second sub-segment are combined to form a first aggregated segment, wherein the brightness data of the aggregated segment is the average brightness data of the plurality of sub-segments; If there is data greater than or equal to a predetermined deviation scalar in the first brightness deviation set, the first sub-segment is set as the first aggregation segment, and iterative clustering is performed starting from the second sub-segment until all sub-segments are traversed, and the first segment division result is obtained and added to the multiple segment division results.

6. The AIoT device control method under the cloud-edge collaborative architecture according to claim 4 is characterized in that: The plurality of road segment division results are sequentially fused and clustered to obtain a plurality of independent road segments, including: Sequentially acquiring a plurality of first aggregated road segments from the plurality of road segment division results, wherein each first aggregated road segment includes a plurality of first sub-road segments; Counting the frequencies of occurrence of the first sub-segment in the plurality of first aggregated segments respectively, and adding the sub-segments whose frequencies are greater than a predetermined threshold to the first independent segment, and adding the sub-segments whose frequencies are less than or equal to the predetermined threshold to the second independent segment; Taking the second independent road segment as a starting point, continue to perform fusion clustering to obtain multiple independent road segments.

7. The AIoT device control method under the cloud-edge collaborative architecture according to claim 1 is characterized in that: The construction process of the vehicle light state prediction plug-in includes: At the cloud computing layer, based on the historical monitoring data of the first independent road section, a sample time node set and a sample traffic flow set are collected, and the proportion of vehicles using high and low beams at different sample time nodes and sample traffic flows is marked to obtain a sample low beam probability set and a sample high beam probability set; Using the sample time node set, the sample traffic volume set, the sample low beam probability set and the sample high beam probability set to train a feedforward neural network until convergence, to obtain a first vehicle light state prediction branch; Analyze and obtain multiple vehicle light state prediction branches of multiple independent road sections in sequence, map and construct the vehicle light state prediction plug-in, and delegate it to the edge computing layer.

8. The AIoT device control method under the cloud-edge collaborative architecture according to claim 1 is characterized in that: According to the vehicle high beam brightness, the vehicle low beam brightness, the multiple low beam probabilities and the multiple high beam probabilities, multiple predicted vehicle light brightnesses are calculated, including: If the low beam probability or the high beam probability is greater than a predetermined probability threshold, the low beam probability or the high beam probability is marked as 1, wherein the predetermined probability threshold is 75%; A weighted calculation is performed based on the vehicle high beam brightness, the vehicle low beam brightness, a plurality of low beam probabilities and a plurality of high beam probabilities, and a plurality of predicted vehicle light brightnesses are output.

9. The AIoT device control system under the cloud-edge collaborative architecture is characterized by: Steps for implementing the AIoT device control method under the cloud-edge collaborative architecture as described in any one of claims 1 to 8, wherein the AIoT device control system under the cloud-edge collaborative architecture includes: A clustering module is used to cluster the target road according to the nighttime brightness monitoring record of the target road, determine multiple independent road sections, and lay out a brightness perception layer for each independent road section; A brightness monitoring module, used for acquiring a plurality of real-time ambient brightness of the plurality of independent road sections through a brightness perception layer monitoring; A road condition monitoring module is used to monitor and obtain road traffic volume and vehicle images through a data perception layer at a predetermined location, wherein an edge computing layer is also embedded in the predetermined location; A headlight brightness prediction module, used to obtain multiple predicted headlight brightnesses based on the road traffic flow and vehicle image analysis at the edge computing layer; The brightness control module is used to optimize the standard driving brightness at night according to the multiple real-time ambient brightnesses and the multiple predicted headlight brightnesses, obtain multiple adaptive driving brightnesses, and dynamically control the streetlight system of the target road.

Images