Multi-lamp-post cooperative intelligent dimming energy-saving method based on cloud LSTM prediction

Abstract

The invention discloses a multi-lamp-post cooperative intelligent dimming and energy saving method based on cloud LSTM prediction, and belongs to the field of intelligent dimming and energy saving. The method comprises the following steps: collecting traffic flow, illuminance, weather information, time information and adjacent lamp post brightness data by using a local lamp post, and uploading the data to a cloud through 4G / 5G; the cloud end utilizes an LSTM deep learning model to predict the traffic flow trend in the future 5-30 minutes; generating a brightness interval and a dimming strategy according to the predicted traffic flow and the road lighting standard in combination with an edge cloud brightness fusion strategy; constructing an adjacent matrix by using the positions of the lamp posts, and performing multi-lamp-post collaborative brightness calculation on the lamp posts according to the states of the adjacent lamp posts; and according to the actual lighting use condition, the cloud end executes a global energy-saving optimization algorithm to obtain an optimal brightness distribution strategy of the whole road. According to the invention, the technical system of edge real-time perception, cloud trend prediction, multi-lamp-post collaborative dimming, energy-saving optimization and lifelong learning is adopted, so that the road illumination efficiency, safety and energy-saving capability are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to intelligent dimming energy saving, and in particular to a multi-lamp-pole collaborative intelligent dimming energy saving method based on cloud LSTM prediction. BACKGROUND

[0002] Traditional road lighting mostly relies on fixed brightness or timing control mode, and cannot adaptively dim according to real-time traffic, environmental lighting, weather changes and other information of the road. With the increase of traffic complexity, the fixed dimming scheme gradually cannot meet the dual needs of energy saving and safety.

[0003] The existing intelligent street light system usually adopts local sensor data for simple dimming, but has the following disadvantages: 1. Poor prediction ability, unable to dim in advance; 2. Lack of collaborative control between lamp poles, poor lighting continuity; 3. Lack of global energy saving optimization strategy; 4. Low utilization rate of cloud data, and low overall intelligence level of the system.

[0004] Therefore, there is an urgent need for an intelligent dimming method based on cloud prediction, global scheduling, multi-lamp-pole collaboration and long-term evolutionary optimization. SUMMARY

[0005] The present application aims to provide an intelligent road lighting method based on cloud LSTM prediction, multi-lamp-pole collaborative dimming and cloud energy saving optimization control, to solve the problems of existing system dimming lag, high energy consumption, lighting discontinuity, etc., and realize efficient, intelligent and energy saving control of road lighting.

[0006] Technical scheme: A multi-lamp-pole collaborative intelligent dimming energy saving method based on cloud LSTM prediction, comprising the following steps: S1: The local lamp pole collects traffic flow, illuminance, weather information, time information and adjacent lamp pole brightness data, and uploads them to the cloud through 4G / 5G; S2: The cloud uses an LSTM deep learning model to predict the traffic trend for the next 5-30 minutes; S3: According to the predicted traffic and road lighting standards, combined with the edge cloud brightness fusion strategy, generate brightness interval and dimming strategy; S4: Use the lamp pole position to construct an adjacency matrix, and the lamp pole performs multi-lamp-pole collaborative brightness calculation according to the adjacent lamp pole state; S5: According to the actual lighting usage, the cloud executes a global energy saving optimization algorithm to obtain the optimal brightness distribution strategy for the entire road; S6: The lamp pole outputs the light source according to the target brightness, and uploads the running data for continuous learning and model optimization of the cloud; S7: Use bidirectional certificate authentication, encrypted channel and containerized deployment technology to build a high-security, scalable edge cloud collaborative platform that can support a large number of lamp pole devices.

[0007] Furthermore, the predicted data in step S2 is calculated using the following formula: in, For time The input feature vector includes traffic flow, pedestrian count, weather parameters, and holiday features. For the hidden state vector, , These are the weight matrices, This is a bias term.

[0008] Furthermore, the brightness requirement value in step S2 is calculated using the following formula: in, For the future The brightness requirement value predicted by the time step. Also a weight matrix, Also a bias term, the cloud-based prediction results are used to generate suggested trend brightness values. They participate in subsequent collaborative dimming and energy-saving optimization decisions.

[0009] Furthermore, the brightness setting value in step S3 is obtained through the following calculation formula: in The brightness suggestion value is calculated based on real-time perception for the edge nodes. The brightness recommendation value output by the aforementioned cloud-based time-series prediction model. Let be the fusion coefficient, satisfying It can be dynamically adjusted according to channel delay, prediction confidence and environmental changes.

[0010] Furthermore, the multi-lamp pole coordinated brightness in step S4 is calculated using the following formula: in, For the target objects (vehicles, pedestrians) in time Location, The distance between each light pole. For data decay factor, For the first A lamppost Collaborative weights Set the brightness value for this road section. For the first The final brightness value of each lamp post.

[0011] Furthermore, the optimization objective function in step S5 can be expressed as: Its constraints are: in, For the number of light poles, For the first The brightness setting value for each light pole. For the first Each lamppost has varying brightness The energy consumption function under the following conditions To meet the minimum brightness requirements of road lighting standards, , These are the allowable brightness ranges, This represents the maximum permissible brightness variation between adjacent light poles.

[0012] Beneficial effects: (1) The LSTM / Transformer prediction model was introduced into road lighting for the first time, enabling the system to have the ability to "advance dimming", which solved the problem of response lag in traditional systems from a mechanism perspective; (2) Real-time trajectory analysis is used to realize the dynamic linkage of the light pole group, so that the light pole in front lights up in advance and the light pole behind lights down with a delay, forming a continuous light strip lighting without discontinuity; (3) Use cloud computing to plan the overall energy consumption of the entire road network, and use mathematical programming / reinforcement learning to achieve the optimal energy consumption under the premise of meeting the illuminance standard, breaking through the limitations of the existing single lamp independent dimming; (4) Supports automatic reporting of difficult cases by edge nodes, incremental training of models in the cloud and distribution of new versions, so that the lighting system can adapt to changes in seasons, weather, events and so on, and achieve continuous evolution; (5) Adopt two-way certificate authentication, encrypted channels and containerized deployment technology to build a highly secure and scalable edge-cloud collaboration platform that can support a large number of light pole devices; Attached Figure Description

[0013] Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is the overall system architecture diagram of the present invention. Detailed Implementation

[0014] To make the technical solution of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0015] Example: A cloud-based LSTM-predicted multi-lamp collaborative intelligent dimming and energy-saving method for a city I. System Initialization Configuration 1. Deploy 100 smart light poles, with a spacing of 30 meters between adjacent light poles, forming a 3-kilometer road lighting network. 2. Each light pole is equipped with: Millimeter-wave radar traffic flow detector; Illuminance sensor; Temperature and humidity weather sensors; 4G / 5G communication module; 3. The cloud platform is deployed using Kubernetes containerization and establishes a secure communication channel using two-way certificate authentication.

[0016] Specific implementation steps Step S1: Data Collection and Upload Real-time data collection from light poles: Traffic flow: The number of vehicles passing through per minute is detected by radar (e.g., 25 vehicles / minute). Illuminance: Real-time ambient light intensity (e.g., 50 lux); Weather information: temperature, humidity, rainfall; Time information: current time, holiday information; Neighboring light pole brightness: The current brightness value of the three adjacent light poles; Data is encrypted and uploaded to the cloud via 4G / 5G network, with a transmission interval of 1 minute; Step S2: Cloud-based LSTM prediction Predicting traffic flow trends for the next 15 minutes using an LSTM model: in The information includes a traffic volume of 85, pedestrian count of 12, temperature of 28 degrees Celsius, humidity of 65%, and a holiday indicator of 0. , The weight matrix obtained by linking is called the weight matrix. As a bias term, the output prediction result is: traffic flow will increase by 40% in the next 15 minutes.

[0017] Step S3: Brightness Strategy Generation Calculate the brightness requirement value based on the prediction results: in Also a weight matrix, Similarly, as an offset term, the calculated brightness requirement value is 300 lux.

[0018] Perform edge-cloud brightness fusion: Take the fusion coefficient =0.3 (when channel delay is low): =280 lux (real-time edge calculation value) =300 lux (cloud-predicted value), calculated to determine the final brightness setpoint. =0.3×280 + 0.7×300= 294 lux.

[0019] Step S4: Multi-lamp pole collaborative calculation Perform coordinated brightness calculation on the target vehicle's position at time t: Target vehicle location (150,0) (150 meters from light pole #1), data attenuation factor =0.02, light pole #1 is 150 meters away, with a coordination weight of 0.85; light pole #2 is 120 meters away, with a coordination weight of 0.78. The calculated result is... =217 lux.

[0020] Step S5: Global Energy Saving Optimization Execute the optimization objective function: Its constraints are: Specific parameters: Number of light poles N=100, and simultaneously , The maximum brightness difference between adjacent light poles is 150 lux and 500 lux respectively. =100 lux, and the optimal brightness distribution scheme for each lamp post is obtained by solving linear programming.

[0021] Step S6: Dimming Execution and Data Feedback The light pole receives the target brightness value transmitted from the cloud. Adjust the LED driver power supply to output the corresponding brightness; Simultaneously upload actual operating data (actual energy consumption, equipment status, etc.); Difficult example data is collected in the cloud, and the LSTM model is updated weekly through incremental training.

[0022] Step S7: Security Operation and Maintenance Assurance Digital certificates are rotated monthly. Monitor the encryption status of communication channels; Containerized services automatically scale up and down to handle traffic peaks.

[0023] III. Implementation Results This embodiment fully implements: LSTM-based 15-minute advance predictive dimming; Multiple light poles work together to form a continuous light strip; Global energy-saving optimization reduces energy consumption by 23%; The system adapts to seasonal changes through continuous learning; Ensure the safe and stable operation of equipment with a scale of thousands.

[0024] 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 multi-lamp pole collaborative intelligent dimming and energy-saving method based on cloud-based LSTM prediction, characterized in that, Includes the following steps: S1: Local light poles collect traffic flow, illuminance, weather information, time information, and brightness data of nearby light poles, and upload them to the cloud via 4G / 5G; S2: The cloud uses an LSTM deep learning model to predict traffic flow trends for the next 5-30 minutes; S3: Based on predicted traffic flow and road lighting standards, combined with edge-cloud brightness fusion strategy, generate brightness range and dimming strategy; S4: Construct an adjacency matrix using the position of the light poles, and perform multi-light pole collaborative brightness calculation based on the status of neighboring light poles; S5: Based on actual lighting usage, the cloud executes a global energy-saving optimization algorithm to obtain the optimal brightness distribution strategy for the entire road; S6: The light pole outputs light source according to the target brightness and uploads operation data for continuous learning and model optimization in the cloud; S7: Employs two-way certificate authentication, encrypted channels, and containerized deployment technologies to build a highly secure and scalable edge-cloud collaboration platform that supports large-scale light pole devices.

2. The multi-lamp pole collaborative intelligent dimming and energy-saving method based on cloud-based LSTM prediction according to claim 1, characterized in that, The predicted data in step S2 is calculated using the following formula: in, For time The input feature vector includes traffic flow, pedestrian count, weather parameters, and holiday features. For the hidden state vector, , These are the weight matrices, This is a bias term.

3. The multi-lamp pole collaborative intelligent dimming and energy-saving method based on cloud-based LSTM prediction according to claim 1, characterized in that, The brightness requirement value in step S2 is calculated using the following formula: in, For the future The brightness requirement value predicted by the time step. Also a weight matrix Also a bias term, the cloud-based prediction results are used to generate suggested trend brightness values. They participate in subsequent collaborative dimming and energy-saving optimization decisions.

4. The multi-lamp pole collaborative intelligent dimming and energy-saving method based on cloud-based LSTM prediction according to claim 3, characterized in that, The brightness setting value in step S3 is obtained through the following calculation formula: in The brightness suggestion value is calculated based on real-time perception for the edge nodes. The brightness recommendation value output by the aforementioned cloud-based time-series prediction model. Let be the fusion coefficient, satisfying It can be dynamically adjusted according to channel delay, prediction confidence and environmental changes.

5. The multi-lamp pole collaborative intelligent dimming and energy-saving method based on cloud-based LSTM prediction according to claim 1, characterized in that, The multi-lamp pole coordinated brightness in step S4 is calculated using the following formula: in, For the target objects (vehicles, pedestrians) in time Location, The distance between each light pole. For data decay factor, For the first A lamppost Collaborative weights, Set the brightness value for this road section. For the first The final brightness value of each lamp post.

6. The multi-lamp pole collaborative intelligent dimming and energy-saving method based on cloud-based LSTM prediction according to claim 1, characterized in that, The optimization objective function in step S5 can be expressed as: Its constraints are: in, For the number of light poles, For the first The brightness setting value for each light pole. For the first Each lamppost has varying brightness The energy consumption function under the following conditions To meet the minimum brightness requirements of road lighting standards, , These are the allowable brightness ranges, This represents the maximum permissible brightness variation between adjacent light poles.

Images