Wireless networking type intelligent dimming control system for solar street lamps

Abstract

The invention relates to the technical field of intelligent control, in particular to a solar street lamp wireless networking type intelligent dimming control system which comprises a data acquisition module, an intelligent cloud management module and a dimming control module. The data acquisition module acquires real-time operation data, environment illumination conditions and road section demand data of a plurality of solar street lamps through wireless networking, and communicates with the intelligent cloud management module. The intelligent cloud management module determines the use state of the street lamp according to the real-time data, calculates the brightness adjustment value of each street lamp based on a reinforcement learning algorithm in combination with the environment illumination, the road section demand and the street lamp state, and outputs an adjustment signal to the dimming control module. The dimming control module adjusts the brightness value of the corresponding street lamp according to the signal. The system realizes high-efficiency energy-saving and intelligent management by cooperatively controlling the brightness value of the street lamp, and improves the overall lighting effect and the energy utilization rate.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent control, and particularly relates to a solar street lamp wireless networking type intelligent dimming control system. BACKGROUND

[0002] With the rapid development of solar street lamp technology, its application in urban lighting, rural roads and remote areas is becoming more and more widespread. However, the existing solar street lamp system mostly adopts an independent light induction control mechanism, that is, each street lamp only performs a simple opening or closing operation according to its own environmental light conditions. This independent control method can meet the basic lighting needs, but has obvious limitations in actual application. First, there is a lack of coordinated control of light intensity between multiple street lamps, resulting in the light intensity state of different street lamps not being able to dynamically change with the environment when the light conditions change, affecting the overall lighting effect. Second, this independent control method cannot be optimized and adjusted according to the global light conditions, which may lead to energy waste or insufficient lighting problems. SUMMARY

[0003] The main purpose of the present application is to provide a solar street lamp wireless networking type intelligent dimming control system, which aims to solve the technical problem that the existing solar street lamp system lacks coordinated control of light intensity between multiple street lamps, resulting in the light intensity state of different street lamps not being able to dynamically change with the environment when the light conditions change, affecting the overall lighting effect.

[0004] To achieve the above purpose, the present application provides a solar street lamp wireless networking type intelligent dimming control system, which comprises a data acquisition module, an intelligent cloud management module and a dimming control module. The data acquisition module is in communication connection with the intelligent cloud management module, and the intelligent cloud management module is in communication connection with the dimming control module. The data acquisition module is used to collect real-time running data, real-time environmental light condition data and road segment demand data of multiple solar street lamps, wherein the multiple solar street lamps realize data interaction through wireless networking. The intelligent cloud management module is used to determine the use state of each solar street lamp according to the real-time running data, wherein the use state includes opening, closing, dimming and failure, and based on a reinforcement learning algorithm, determines the brightness adjustment value of each solar street lamp according to the real-time environmental light condition data, the road segment demand data and the use state of each solar street lamp, and outputs the adjustment signal corresponding to the brightness adjustment value to the dimming control module. The dimming control module is used to adjust the brightness value of the corresponding solar street lamp according to the adjustment signal.

[0005] Optionally, the reinforcement learning algorithm is based on the real-time ambient light condition data, the road segment demand data, and the usage state of each solar street lamp to determine a brightness adjustment value of each solar street lamp, and the method comprises the following steps of: collecting historical light data of each road segment, including daily ambient light condition data, road segment demand data, and usage state of each solar street lamp; predicting lighting demand of each road segment in a future preset time according to the historical light data, and calculating actual light adjustment demand of each road segment according to the lighting demand of each road segment in the future preset time and real-time ambient light condition data between road segments; taking the actual light adjustment demand of each road segment, the current position of each street lamp, and the real-time ambient light condition between road segments as a state, taking adjustment of brightness of street lamps of different road segments as an action, defining a reward function to meet the degree of actual light adjustment demand of each road segment and to meet the actual light adjustment demand of each road segment, and training a DQN network; generating the brightness adjustment value of each solar street lamp in real time based on the trained DQN network.

[0006] Optionally, the formula of the DQN network is as follows: ; In the formula, is a current state parameter set, including actual light adjustment demand of each road segment, current position of each street lamp, and real-time ambient light condition between road segments; is a current action parameter set, including a strategy of adjusting brightness of street lamps between road segments; is a state parameter set is an action parameter set ; is a value of the action parameter set is a learning rate; is an immediate reward of the current action; is a discount factor; is a next state parameter set; is an action parameter set in the next state, is an updated value of the action parameter set .

[0007] Optionally, the formula of the reward function is as follows:

[0008] In the formula, R is a reward value; and β are weight coefficients; is actual satisfied light adjustment demand of a road segment; is total light adjustment demand of the road segment; is an actual light adjustment response time; is a maximum allowed light adjustment response time.

[0009] Optionally, the system further comprises a feature solving module, a parameter analysis module, a control correction module and an instruction output module; The feature solving module is configured to perform multi-dimensional dynamic feature extraction on real-time operation data, real-time environmental lighting condition data and road segment demand data of the plurality of solar street lamps, obtain a system operation feature vector, and perform state feedback equation solving to obtain a mapping relationship matrix of the master street lamp and the plurality of slave street lamps; The parameter analysis module is configured to input the mapping relationship matrix and the brightness adjustment value of each solar street lamp into a target control model for control parameter analysis to obtain an initial control sequence; The control correction module is configured to input the system operation feature vector and the initial control sequence into a preset correction model for master-slave street lamp control correction to obtain a control correction amount; The instruction output module is configured to compensate and correct the initial control sequence based on the control correction amount to obtain a target control sequence, and output the target control sequence to the dimming driver of each street lamp through wireless networking to output a cooperative dimming control result.

[0010] Optionally, the multi-dimensional dynamic feature extraction on real-time operation data, real-time environmental lighting condition data and road segment demand data of the plurality of solar street lamps to obtain a system operation feature vector, and the state feedback equation solving to obtain a mapping relationship matrix of the master street lamp and the plurality of slave street lamps comprise: The real-time operation data, real-time environmental lighting condition data and road segment demand data of the plurality of solar street lamps are grouped according to master-slave street lamps, the data matrix of the master street lamp is marked as a master control matrix, and the data matrix of the plurality of slave street lamps is marked as a slave control matrix; The master control matrix and the slave control matrix are standardized to obtain standardized time sequence data of the master-slave street lamps, the standardized time sequence data is subjected to frequency spectrum analysis and time domain feature calculation to obtain a system operation feature vector; A state space equation of the master-slave street lamps is established according to the system operation feature vector, a control equation containing state feedback items and following feedback items is constructed, a master-slave coupled state equation is obtained, stability analysis is performed on the master-slave coupled state equation, feedback gain parameters are calculated, and cooperative control parameters of the master-slave street lamps are obtained; A master-slave mapping function is constructed according to the cooperative control parameters, a state transition matrix is calculated, a following control amount of the slave control street lamp to the master control street lamp is obtained, the following control amount is subjected to symmetric matrix transformation, a master-slave state corresponding relationship is established, and a state conversion matrix is obtained; The state conversion matrix is subjected to parameter calibration to obtain the mapping relationship matrix of the master street lamp and the plurality of slave street lamps.

[0011] Optionally, the mapping relationship matrix and the brightness adjustment value of each solar street lamp are input into a target control model for control parameter analysis to obtain an initial control sequence, including: A master-slave mapping vector group is extracted based on the mapping relationship matrix, and a master-slave mapping vector group is generated; The master-slave mapping vector group is input into a target control model to establish a state transition cost function of the master-slave street lamps, obtain a state cost matrix, and perform weighting processing on the state cost matrix according to the brightness adjustment value of each solar street lamp, construct a collaborative control objective function of the master-slave street lamps, and obtain a control cost function; The control cost function is solved based on dynamic programming to obtain an optimal control strategy, and the optimal control strategy is subjected to time discretization processing to extract the control amount of the master street lamp and the slave street lamp at each time point, respectively, to obtain a discrete control sequence; The discrete control sequence is subjected to master-slave synchronization analysis to construct a brightness synchronization constraint condition between the master-slave street lamps, obtain a synchronization control constraint, and modify the discrete control sequence according to the synchronization control constraint to establish a collaborative control instruction set of the master-slave street lamps, and obtain an initial control sequence.

[0012] Optionally, the system operation feature vector and the initial control sequence are input into a preset correction model for master-slave street lamp control correction to obtain a control correction amount, including: The master-slave components of the system operation feature vector are extracted, and the feature components of the master street lamp and the feature components of each slave street lamp are classified respectively to obtain a master-slave feature classification matrix; A master-slave collaborative state space is established according to the master-slave feature classification matrix and the initial control sequence, the master street lamp is set as a decision maker, and the slave street lamp is set as an execution subject to obtain a master-slave collaborative learning environment; The master-slave collaborative learning environment is input into a preset correction model to construct a hierarchical structure of the master control network and the slave control network, and a decision unit driven by light is embedded in the master control network to obtain a hierarchical control architecture; The control strategy of the master street lamp is evaluated online based on the hierarchical control architecture to generate a master control correction instruction, and a control signal is passed down through the master control network to obtain a master control adjustment sequence, and the control parameters of the multiple slave street lamps are collaboratively optimized according to the master control adjustment sequence, the response deviation of each slave street lamp is calculated through the slave control network to obtain a slave control deviation sequence; The master control adjustment sequence and the slave control deviation sequence are subjected to time alignment and phase correction to establish a synchronous response relationship between the master-slave street lamps, obtain a synchronous control parameter, and overall correct the control strategy of the master-slave street lamps according to the synchronous control parameter to construct a control correction amount containing a brightness correction amount and a phase correction amount.

[0013] Optionally, the initial control sequence is compensated and corrected based on the control correction amount to obtain a target control sequence, and the target control sequence is respectively issued to the dimming driver of each street lamp through wireless networking to output a cooperative dimming control result, including: The control correction amount is subjected to master-slave decomposition operation to extract the brightness correction parameter of the master street lamp and the phase following parameter of each slave street lamp to obtain a master-slave correction parameter matrix; The master-slave correction parameter matrix is subjected to compensation superposition operation with the initial control sequence, and the control instruction of the master street lamp is marked with priority to obtain a target control sequence; The synchronous following instruction of the slave street lamp is generated according to the master street lamp control instruction in the target control sequence, and the control timing relationship between the master and slave street lamps is established to obtain a master-slave cooperative control table; The control instruction in the master-slave cooperative control table is subjected to wireless network communication format conversion to generate the control data frame of the master street lamp and the response data frame of the slave street lamp to obtain a communication data frame sequence; The real-time communication scheduling strategy is constructed based on the communication data frame sequence, the data of the master street lamp is set to be transmitted preferentially, and the slave street lamp is transmitted in turn according to the phase difference order to obtain a communication scheduling sequence; The communication scheduling sequence is packaged according to the address coding rule of the dimming driver, and a synchronous trigger signal is added to obtain a driver control instruction; The network bandwidth is allocated to the driver control instruction to ensure that the control instruction of the master street lamp is issued preferentially and the control instruction of the slave street lamp is executed in sequence to obtain a time-sharing control sequence; The control instruction is respectively issued to the dimming driver of the master street lamp and each slave street lamp based on the time-sharing control sequence through wireless network, and the execution feedback of each street lamp is collected to output a cooperative dimming control result.

[0014] Optionally, the system further comprises an abnormal alarm module; The abnormal alarm module is configured to determine that the street lamp is faulty when the real-time running data of the street lamp deviates from the preset normal running parameter by more than a threshold value, generate corresponding alarm information according to the fault type, send the alarm information to the monitoring center through wireless networking, so that the monitoring center generates a maintenance task and distributes it to the maintenance personnel according to the received alarm information, and updates the state of the street lamp to fault.

[0015] The application discloses a solar street lamp wireless networking intelligent dimming control system. BRIEF DESCRIPTION OF DRAWINGS BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 is a structural schematic diagram of a first embodiment of the solar street lamp wireless networking intelligent dimming control system of the application; Figure 2 is a structural schematic diagram of a second embodiment of the solar street lamp wireless networking intelligent dimming control system of the application; Figure 3 is a specific step flow chart of the solar street lamp wireless networking intelligent dimming control system of the application, in which, based on a reinforcement learning algorithm, brightness adjustment values of each solar street lamp are determined according to real-time environmental illumination condition data, road segment demand data and the use state of each solar street lamp.

[0017] BRIEF DESCRIPTION OF DRAWINGS

[0018] The implementation, functional features and advantages of the application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the application will be clearly and completely described with reference to the accompanying drawings in the embodiments of the application. Obviously, the described embodiments are only a part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the application.

[0020] It should be noted that if the embodiments of the present invention involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0021] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0022] This invention provides a wireless networking intelligent dimming control system for solar streetlights, referring to... Figure 1 As shown, Figure 1 This is a structural block diagram of the first embodiment of the wireless networking intelligent dimming control system for solar streetlights of the present invention. The wireless networking intelligent dimming control system for solar streetlights of the present invention includes: a data acquisition module 10, an intelligent cloud management module 20, and a dimming control module 30; The data acquisition module 10 is communicatively connected to the intelligent cloud management module 20, and the intelligent cloud management module 20 is communicatively connected to the dimming control module 30. It should be noted that the data acquisition module 10 is the core component of the solar street lamp wireless networking intelligent dimming control system, mainly responsible for collecting multi-dimensional data to provide decision basis for the intelligent cloud management module 20. Specifically, collecting real-time operation data means obtaining the working state of each solar street lamp (such as voltage, current, power, etc.), monitoring the health status of the street lamp, and discovering faults in time. Collecting environmental light data means monitoring the environmental light intensity in real time through the light sensor to provide the basis for brightness adjustment. Collecting road segment demand data means dynamically adjusting the lighting demand according to the road segment characteristics (such as traffic flow, passenger flow, road grade, etc.). In addition, the data acquisition module 10 transmits the collected data to the intelligent cloud management module 20 through wireless networking technology, and realizes data sharing and collaboration between multiple street lamps. In specific implementation, light sensors, voltage sensors, current sensors, etc. can be installed on each street lamp to collect environmental light and operation data, secondly, ZigBee, LoRa or NB-IoT low-power wireless communication technology can be used to realize data interaction between street lamps and communication with the cloud, and at the same time, a microcontroller (such as STM32) can be integrated for data preprocessing and transmission control. For example, the light sensor monitors the environmental light intensity in real time, sends the data to the microcontroller, the voltage and current sensors collect the street lamp operation data, judge the working state of the street lamp, the microcontroller preliminarily processes the data, and uploads it to the intelligent cloud management module 20 through the wireless communication module.

[0023] It should be understood that the intelligent cloud management module 20 is configured to receive real-time operation data, environmental lighting data, and road segment demand data uploaded by the data acquisition module 10, clean, classify, and standardize heterogeneous data, and form a unified decision dataset. The real-time operation data can include lamp voltage, current, and fault status; the environmental lighting data can include lighting intensity and diurnal variation; and the road segment demand data can include traffic flow, passenger flow, and time period lighting demand. Second, the intelligent cloud management module 20 is also configured to dynamically determine the use state of each street lamp according to the real-time operation data, establish a street lamp state account, support fault early warning and remote state query, and provide real-time basis for operation and maintenance. The use state of the street lamp includes on, off, dimming, and fault state; and the fault early warning includes battery undervoltage and lamp abnormality. In addition, the intelligent cloud management module 20 is also configured to generate a globally optimal brightness adjustment strategy based on a reinforcement learning algorithm, in combination with environmental lighting variation, road segment real-time demand, and street lamp state. Through algorithm optimization, the lighting effect and energy consumption are balanced, and fine-tuned dimming is achieved. The road segment real-time demand can be to reduce brightness for a low-traffic road segment at night. In addition, the intelligent cloud management module 20 constructs a street lamp network topology through wireless networking, breaks the limitation of single-lamp independent control, supports coordinated adjustment of the lighting intensity of adjacent street lamps, and improves overall lighting uniformity and energy efficiency. The coordinated adjustment can be that when the lighting of a certain road segment is insufficient, the brightness of adjacent street lamps is automatically increased to compensate.

[0024] In specific implementation, a distributed cloud computing platform can be adopted to support high-concurrency data processing and elastic expansion, to ensure real-time access and calculation of massive street lamp data; an edge server is deployed in areas with weak network coverage to realize local data preprocessing, reduce the computing pressure of the cloud, and improve response speed; a street lamp dedicated database is constructed to store real-time data, historical records, device archives, etc., support mixed SQL and NoSQL queries, and provide a data visualization interface, such as a street lamp status map and an energy consumption curve; a reinforcement learning framework (such as TensorFlow or PyTorch) is integrated to construct a dimming strategy model. The core algorithm logic includes state space definition, action space definition, and reward function design. Among them, the state space definition includes ambient light intensity (0-100%), current brightness of the street lamp (0-100%), battery remaining capacity, and real-time traffic volume of the road section (predicted through Internet of Things sensors or historical data); the action space definition includes brightness adjustment value (such as ±10%, ±20% step, or 0-100% continuous value), and supports single lamp or regional batch adjustment. The reward function design includes comprehensive energy saving target (such as reducing energy consumption weight 0.6), lighting compliance rate (such as road section average illumination ≥ 15 lux weight 0.3), and device life (such as avoiding frequent dimming loss weight 0.1), to find the optimal strategy through algorithm iteration. For example, real-time data uploaded by the data collection module 10 is received to construct a current system state matrix, such as 100 street lamps x state parameter dimension; the reinforcement learning model outputs the brightness adjustment action of each street lamp, such as “street lamp A is adjusted to 60% brightness”, according to the current state and in combination with historical experience (stored in the experience replay buffer); after the adjustment action is executed, the next time data is collected, the reward value is calculated, such as reward +1 for energy consumption reduction and penalty -1 for non-compliance of illumination, and the model parameters are updated to form a closed-loop optimization.

[0025] In addition, the dimming control module 30 is configured to receive the brightness adjustment signal issued by the intelligent cloud management module 20, such as adjusting to 60% brightness, to realize stepless or stepwise adjustment of the street lamp brightness through a hardware driving circuit, to ensure dimming accuracy and response speed; the dimming control module 30 is compatible with different types of street lamp fixtures and driving methods, to realize unified control of different hardware through a standardized interface. Among them, the street lamp fixture can be a LED or a high-pressure sodium lamp, and the driving method can be PWM dimming or constant current dimming; in addition, the dimming control module 30 is also configured to feed back the dimming execution result to the intelligent cloud management module 20 in real time to form a closed-loop control of “strategy generation-execution-feedback”, to ensure the reliability of system adjustment; among them, the dimming execution result can be the actual brightness value and whether the adjustment is successful.

[0026] In the embodiment, a solar street lamp wireless networking intelligent dimming control system, the system comprises: a data acquisition module 10, an intelligent cloud management module 20 and a dimming control module 30; wherein the data acquisition module 10 is in communication connection with the intelligent cloud management module 20, the intelligent cloud management module 20 is in communication connection with the dimming control module 30; the data acquisition module 10 is used for collecting real-time running data, real-time environmental illumination condition data and road segment demand data of a plurality of solar street lamps, wherein the plurality of solar street lamps realize data interaction through wireless networking; the intelligent cloud management module 20 is used for determining the use state of each solar street lamp according to the real-time running data, the use state including opening, closing, dimming and failure, and determining the brightness adjustment value of each solar street lamp according to the real-time environmental illumination condition data, the road segment demand data and the use state of each solar street lamp based on a reinforcement learning algorithm, and outputting an adjustment signal corresponding to the brightness adjustment value to the dimming control module 30; the dimming control module 30 is used for adjusting the brightness value of the corresponding solar street lamp according to the adjustment signal. The system realizes efficient energy saving and intelligent management through cooperative control, improves the lighting consistency and energy utilization rate.

[0027] Further, with reference to Figure 3 , the brightness adjustment value of each solar street lamp is determined according to the real-time environmental illumination condition data, the road segment demand data and the use state of each solar street lamp based on a reinforcement learning algorithm, comprising: Collecting historical illumination data of each road segment, including daily environmental illumination condition data, road segment demand data and the use state of each solar street lamp; According to the historical illumination data, predicting the lighting demand of each road segment in the future preset time, and calculating the actual dimming demand of each road segment according to the lighting demand of each road segment in the future preset time and the real-time environmental illumination condition data of each road segment.

[0028] It should be noted that ambient light intensity can be collected every minute by a light sensor and stored by timestamp; real-time traffic flow, pedestrian flow, and road grade for each road segment can be obtained through inductive loops, camera AI recognition, or manual input. Streetlight status data comes from real-time reporting by data acquisition module 10, including the current brightness, battery level, and fault status of each streetlight. A time-series database is used to store high-frequency light intensity data, and a relational database is used to store road segment attributes and historical dimming records. Data is partitioned by "road segment ID + date," and at least one year of historical data is retained for seasonal pattern learning. Secondly, LSTM + fully connected network is used for future lighting demand prediction, which is suitable for time series and multi-feature coupled prediction. The input layer is the historical 24-hour light intensity sequence, corresponding traffic flow, weekday features, and month. The output layer is the lighting demand level for each time period in the next 4 hours (levels 0-5, corresponding to the minimum to maximum illuminance requirements, such as level 1 corresponding to 10 lux, level 5 corresponding to 30 lux, with preset thresholds based on road grade). The actual dimming requirements of each road segment are determined based on the lighting needs of each road segment and the real-time ambient light conditions of each road segment. For example, if the predicted demand level at 2 a.m. is level 2 (15 lux), but the real-time light is only 10 lux, then the actual dimming requirement is 15 lux - 10 lux = 5 lux compensation, which can be converted into a brightness percentage adjustment value.

[0029] The DQN network is trained by taking the actual dimming needs of each road segment, the current position of each street light, and the real-time ambient light conditions between road segments as the state, and adjusting the brightness of street lights in different road segments as the action. The reward function is defined to satisfy the actual dimming needs of each road segment and to achieve the desired dimming needs. The trained DQN network generates brightness adjustment values ​​for each solar street light in real time.

[0030] The formula for the DQN network is as follows: ; In the formula, This is the current state parameter set, including the actual dimming requirements of each road segment, the current position of each street light, and the real-time ambient light conditions between each road segment; The current action parameter set includes strategies for adjusting street light brightness between road segments; For state parameter set Take action parameter set of value; The learning rate; An immediate reward for the current action; Discount factor; For the next set of state parameters; For the action parameter set of the next state, The updated value.

[0031] It should be noted that the state s includes the actual dimming demand, the street light position, and the neighborhood lighting state, which is the input basis for the agent decision and directly reflects the current system running environment. Among them, the actual dimming demand refers to the difference between the target illumination of the road section and the real-time illumination, the street light position refers to the coordinates or neighborhood relationship, and the neighborhood lighting state refers to the adjacent street light brightness distribution. The action a specifically refers to the street light brightness adjustment strategy between road sections, such as "adjusting the brightness of a certain road section by 20%", which is a discrete or continuous adjustment instruction, and needs to be compatible with the hardware dimming module. The immediate reward r is calculated by the reward function in real time, which measures the short-term benefit of the current action, such as whether the dimming meets the demand and whether the response is timely. The learning rate a is used to control the update amplitude of the new and old Q values (0 < a < 1), a = 1 completely trusts the new sample, and the smaller a is, the more it depends on historical experience, which needs to be dynamically adjusted according to the stability of the data, such as setting 0.3 at the beginning and reducing to 0.1 after convergence. The discount factor γ is used to weigh the weight of future rewards and immediate rewards (0 ≤ γ < 1 ≤ γ < 1), γ = 0 only pays attention to the immediate reward (greedy strategy), and γ close to 1 pays more attention to long-term benefits, such as the long-term improvement of battery life brought by energy saving.

[0032] The above formula is essentially the iterative update formula of the TD(0) algorithm, which uses the maximum Q value of the subsequent state to update the current Q value, avoiding direct dependence on the terminal state reward, and is suitable for continuous control scenarios without explicit termination state, such as street light dimming. The TD error in the parentheses is positive, indicating that the actual return of the current action is higher than expected, and the Q value is increased; if it is negative, it means that the return is lower than expected, and the Q value is reduced, thereby guiding the strategy to converge to the high return action.

[0033] Due to the large scale of the street light network, which may have hundreds or even thousands of nodes, the state s needs to be reduced in dimension, such as clustering by road section and ignoring secondary neighborhood nodes, to avoid the curse of dimensionality. The actual dimming action (such as brightness adjustment step) needs to be discretized into a finite set (such as ±10%, ±20%, etc.), so that can be calculated, while reducing the control complexity.

[0034] The formula of the reward function is as follows:

[0035] In the formula, R is the reward value; and β are weight coefficients; is the actual satisfied road dimming demand; is the total dimming demand of the road section; is the actual dimming response time; Maximal allowed dimming response time.

[0036] It should be noted that, The number of road segments that actually meet the dimming demand, such as uniformity and compliance with the standard of illuminance, needs to be clearly defined as "satisfactory" (e.g. illuminance error < 5%). The weight coefficient of demand satisfaction, reflecting the system's priority for "lighting quality", such as setting = 0.8 for the main road scene and = 0.6 for the branch road. The actual time from the cloud instruction to the completion of the dimming by the street lamp, including communication delay and hardware response time. The weight coefficient of response time, balancing "fast response" and "demand satisfaction", such as increasing the weight coefficient to reduce delay. By rewarding "more demand satisfaction" and punishing "delayed response", the agent is guided to prefer actions that can meet more lighting demands and execute quickly, avoiding excessive dimming delay due to complex strategies, such as communication congestion in cross-region collaborative dimming.

[0037] Referring to Figure 2 , Figure 2 is a structural schematic diagram of the second embodiment of the solar street lamp wireless networking intelligent dimming control system of the present application; based on the above-mentioned first embodiment, the second embodiment of the solar street lamp wireless networking intelligent dimming control system of the present application is proposed.

[0038] Further, in order to realize the collaborative control of multiple solar street lamps in the dimming process and avoid response deviation, in the present embodiment, the system further comprises a feature solving module 40, a parameter analysis module 50, a control correction module 60 and an instruction output module 70; The feature solving module 40 is used for multi-dimensional dynamic feature extraction of real-time running data, real-time environmental light condition data and road segment demand data of multiple solar street lamps, obtaining a system running feature vector, and solving a state feedback equation to obtain a mapping relationship matrix of the master street lamp and multiple slave street lamps; The parameter analysis module 50 is used for inputting the mapping relationship matrix and the brightness adjustment value of each solar street lamp into a target control model for control parameter analysis to obtain an initial control sequence; The control correction module 60 is used for inputting the system running feature vector and the initial control sequence into a preset correction model for master-slave street lamp control correction to obtain a control correction amount; The instruction output module 70 is configured to compensate and correct the initial control sequence based on the control correction amount to obtain a target control sequence, and output a cooperative dimming control result by issuing the target control sequence to the dimming driver of each street lamp through wireless networking.

[0039] It should be noted that the feature solving module 40 is configured to perform outlier processing, missing value filling and feature engineering on the real-time operation data, real-time environment lighting condition data and road segment demand data of the plurality of solar street lamps, and select key features such as light intensity, traffic volume, passenger flow, battery power, current brightness, etc.; normalize or standardize the features to ensure that the numerical values are within a reasonable range; generate composite features such as light change rate and time period features; and establish a state feedback equation to describe the mapping relationship between the master street lamp and the slave street lamp. For example, the relationship between the state of the master street lamp and the state of the slave street lamp may be represented as:

[0040] wherein, A, B and C are mapping relationship matrices, and u is a control input; the mapping relationship matrices A, B and C can be solved by using a least square method, a Kalman filter or other optimization algorithms.

[0041] The parameter analysis module 50 is configured to input the mapping relationship matrix and the brightness adjustment value of each solar street lamp into a target control model to perform control parameter analysis, and obtain an initial control sequence. The mapping relationship matrix is output by the feature solving module 40; the brightness adjustment value is the brightness adjustment value of each street lamp generated by the DQN network. The target control model can select a suitable control model such as a PID controller or an LQR controller; the model configuration configures the model parameters such as the proportional coefficient, the integral coefficient and the differential coefficient according to the system requirements. The input mapping relationship matrix is to input the mapping relationship matrix as part of the model. The input brightness adjustment value is to input the brightness adjustment value generated by the DQN as a control input; the calculation of the initial control sequence is to calculate the initial control sequence, i.e. the preliminary brightness adjustment instruction of each street lamp, by the target control model. The output result initial control sequence is the preliminary brightness adjustment instruction of each street lamp, which is used as the input of the subsequent modules.

[0042] The control correction module 60 is configured to input the system operation feature vector and the initial control sequence into a preset correction model to correct the master-slave street lamp control, so as to obtain a control correction amount. The system operation feature vector is output by the feature solving module 40; and the initial control sequence is output by the parameter analysis module 50. The preset correction model can be a neural network, a support vector machine (SVM), or the like. The correction model is trained using historical data, so that the correction model can learn the relationship between the system operation feature and the control correction amount. Inputting the system operation feature vector means that the system operation feature vector is input as part of the model. Inputting the initial control sequence means that the initial control sequence is input as another part. Calculating the control correction amount means that the control correction amount, that is, the adjustment amount of the initial control sequence, is calculated by the preset correction model. The control correction amount includes the brightness adjustment correction amount of each street lamp, which is used for input of a subsequent module.

[0043] The instruction output module 70 compensates and corrects the initial control sequence based on the control correction amount, so as to obtain a target control sequence, and transmits the target control sequence to the dimming drivers of the street lamps through wireless networking, so as to output a cooperative dimming control result. The initial control sequence is output by the parameter analysis module 50; and the control correction amount is output by the control correction module 60. Compensation calculation means that the initial control sequence and the control correction amount are added to obtain the target control sequence. For example, the target control sequence = the initial control sequence + the control correction amount. Generating the dimming instruction means that the target control sequence is converted into a specific dimming instruction, such as “road section ID = 123, brightness adjustment to 80%”. The wireless networking transmission can use an Internet of Things communication protocol such as MQTT or CoAP to transmit the dimming instruction to the dimming drivers of the street lamps. After receiving the instruction, the dimming driver performs the corresponding brightness adjustment operation. After the dimming is completed, the driver feeds back the actual brightness value, and the system records and updates the state. The cooperative dimming control result means that each street lamp adjusts the brightness according to the target control sequence, so as to realize the cooperative dimming control.

[0044] Further, the real-time operation data, the real-time environmental light condition data, and the road section demand data of the plurality of solar street lamps are subjected to multi-dimensional dynamic feature extraction, so as to obtain a system operation feature vector, and a state feedback equation is solved, so as to obtain a mapping relationship matrix of the master street lamp and the plurality of slave street lamps, including: The real-time operation data, the real-time environmental light condition data, and the road section demand data of the plurality of solar street lamps are grouped according to the master-slave street lamps, the data matrix of the master street lamp is marked as a master control matrix, and the data matrix of the plurality of slave street lamps is marked as a slave control matrix; The master control matrix and the slave control matrix are subjected to standardization processing, so as to obtain standardized time sequence data of the master-slave street lamps, the standardized time sequence data is subjected to frequency spectrum analysis and time domain feature calculation, so as to obtain a system operation feature vector; According to the system operation feature vector, a state space equation of the master-slave street lamp is established, a control equation containing a state feedback term and a following feedback term is constructed, a master-slave coupled state equation is obtained, and stability analysis is performed on the master-slave coupled state equation to calculate a feedback gain parameter, thereby obtaining a cooperative control parameter of the master-slave street lamp; According to the cooperative control parameter, a master-slave mapping function is constructed, a state transition matrix is calculated, a following control amount of the slave street lamp to the master street lamp is obtained, and the following control amount is subjected to a symmetric matrix transformation to establish a master-slave state correspondence relationship, thereby obtaining a state conversion matrix. The state conversion matrix is subjected to parameter calibration to obtain a mapping relationship matrix of the master street lamp and the plurality of slave street lamps.

[0045] It should be noted that the data of the master street lamp is grouped and labeled as a master control matrix, and the data of each slave street lamp is grouped and labeled as a slave control matrix; the master control matrix and the slave control matrix are subjected to standardization processing to make the data mean value 0 and variance 1; the time series data after standardization is subjected to frequency spectrum analysis to extract frequency domain features; time domain features such as mean value, variance, and change rate are calculated; the frequency domain features and the time domain features are spliced into a system operation feature vector; and a state space equation of the master street lamp and the slave street lamp is established:

[0046] Wherein, x is a state vector, u is a control input, A and B are state transition matrices and control matrices.

[0047] A control equation containing a state feedback term and a following feedback term is constructed:

[0048] Wherein, is a feedback gain, is a feedforward gain.

[0049] The state space equation and the control equation of the master street lamp and the slave street lamp are combined to obtain a master-slave coupled state equation:

[0050] Wherein, Describes the joint state of the master street lamp and the slave street lamp, usually including brightness, battery power, fault state, etc. Describes the control instructions for the master street lamp and the slave street lamp, such as brightness adjustment value, switch instruction, etc. Describes the change relationship of the system state with time, reflecting the dynamic coupling characteristics between the master street lamp and the slave street lamp. Describes the influence of the control input on the system state, reflecting the adjustment effect of the control instruction on the master street lamp and the slave street lamp.

[0051] Stability analysis is performed on the master-slave coupled state equation, and the feedback gain parameters are calculated to obtain the cooperative control parameters of the master-slave street lamps. Specifically, the Lyapunov method or eigenvalue analysis method is used to judge the stability of the master-slave coupled state equation; the feedback gain and feedforward gain are calculated by pole placement or LQR optimization method and feedforward gain ; the feedback gain and feedforward gain are used as cooperative control parameters to construct the master-slave mapping function. According to the cooperative control parameters, the master-slave mapping function is constructed, and the state transition matrix is calculated to obtain the following control amount of the slave street lamp to the master street lamp. Specifically, the master-slave mapping function is constructed to describe the relationship between the state of the slave street lamp and the state of the master street lamp:

[0052] The state transition matrix is calculated to describe the response of the state of the slave street lamp to the state of the master street lamp:

[0053] Symmetric matrix transformation is performed on the state transition matrix to ensure the stability of the master-slave state correspondence relationship; the state conversion matrix is established to describe the state correspondence relationship between the master street lamp and the slave street lamp; the state conversion matrix is parameterized to obtain the mapping relationship matrix of the master street lamp and multiple slave street lamps; Specifically, the state conversion matrix is parameterized using historical data to ensure its accuracy; according to the parameterized state conversion matrix, the mapping relationship matrix M of the master street lamp and the slave street lamp is generated, for example, assuming that the state conversion relationship of the master street lamp and the slave street lamp is:

[0054] wherein, describes the mapping relationship between the state i of the master street lamp and the state j of the slave street lamp.

[0055] Further, the mapping relationship matrix and the brightness adjustment value of each solar street lamp are input into the target control model for control parameter analysis to obtain an initial control sequence, including: The mapping vector of the master street lamp and the mapping vector of multiple slave street lamps are extracted based on the mapping relationship matrix, and a master-slave mapping vector group is generated; The master-slave mapping vector group is input into the target control model to establish a state transition cost function of the master-slave street lamps, obtain a state cost matrix, and perform weighting processing on the state cost matrix according to the brightness adjustment value of each solar street lamp to construct a cooperative control objective function of the master-slave street lamps, and obtain a control cost function; Based on the control cost function, dynamic programming is solved to obtain an optimal control strategy, and the optimal control strategy is time-discretized to extract the control amount of the master street lamp and the slave street lamp at each time point, respectively, to obtain a discrete control sequence; The master-slave synchronism of the discrete control sequence is analyzed, a brightness synchronization constraint condition between the master and slave street lamps is constructed, a synchronization control constraint is obtained, the discrete control sequence is revised according to the synchronization control constraint, a collaborative control instruction set of the master and slave street lamps is established, and an initial control sequence is obtained.

[0056] It should be noted that the mapping vector of the master street lamp is extracted from the mapping relationship matrix M and the mapping vector of each slave street lamp . The mapping vector of the master street lamp and the mapping vector of each slave street lamp are combined into a master-slave mapping vector group V=[ , , ,…]. Assuming that the mapping relationship matrix M is:

[0057] The mapping vector of the master street lamp =[ , ], and the mapping vector of the slave street lamp =[ , ]. Based on the master-slave mapping vector group V, a state transition cost function of the master and slave street lamps is established:

[0058] wherein, is a state vector of the i-th street lamp, is an expected state vector, is a state cost matrix. The state cost matrix is weighted processed according to the brightness adjustment value of each street lamp, and a collaborative control objective function of the master and slave street lamps is constructed:

[0059] wherein, is a control input of the i-th street lamp, is a control cost matrix, is a weight coefficient. The state transition cost function and the control objective function are combined to obtain a control cost function:

[0060] Based on the control cost function , a dynamic programming algorithm is used to solve an optimal control strategy ; the optimal control strategy is time-discretized to be decomposed into control amounts at multiple time points. The control amounts of the master street lamp and the slave street lamp at each time point are extracted respectively to obtain a discrete control sequence . For example, assuming that the optimal control strategy is:

[0061] The discrete control sequence is:

[0062] The master-slave synchronization analysis is performed on the discrete control sequence, and the brightness synchronization constraint condition between the master and slave street lamps is constructed:

[0063] wherein, and are the brightness of the master street lamp and the slave street lamp at time t, is a synchronization error threshold; the discrete control sequence is corrected according to the synchronization control constraint to ensure that the brightness of the master street lamp and the slave street lamp meets the synchronization requirement; and the corrected discrete control sequence is taken as the initial control sequence .

[0064] Further, the system operation feature vector and the initial control sequence are input into a preset correction model to correct the master-slave street lamp control, and a control correction amount is obtained, including: The master-slave component extraction is performed on the system operation feature vector, and the feature components of the master street lamp and the feature components of each slave street lamp are classified respectively to obtain a master-slave feature classification matrix; According to the master-slave feature classification matrix and the initial control sequence, a master-slave collaborative state space is established, the master street lamp is set as a decision maker, and the slave street lamp is set as an execution subject, and a master-slave collaborative learning environment is obtained; The master-slave collaborative learning environment is input into a preset correction model, a hierarchical structure of a master control network and a slave control network is constructed, a decision unit driven by light is embedded in the master control network, and a hierarchical control architecture is obtained; Based on the hierarchical control architecture, the control strategy of the master street lamp is evaluated online, a master control correction instruction is generated, a control signal is transmitted downward through the master control network, a master control adjustment sequence is obtained, the control parameters of the multiple slave street lamps are cooperatively optimized according to the master control adjustment sequence, the response deviation of each slave street lamp is calculated through the slave control network, and a slave control deviation sequence is obtained; The master control adjustment sequence and the slave control deviation sequence are time-aligned and phase-corrected, a synchronization response relationship between the master and slave street lamps is established, synchronization control parameters are obtained, the control strategy of the master and slave street lamps is overall corrected according to the synchronization control parameters, and a control correction amount containing a brightness correction amount and a phase correction amount is constructed.

[0065] It is necessary to explain that the system operation feature vector is a multi-dimensional vector used to describe the operation state and characteristics of the solar street light system. It includes parameters such as light intensity, battery power, working temperature, brightness adjustment value, etc. Master-slave component extraction refers to extracting the feature components of the master street light and the slave street light from the system operation feature vector. The master street light is the core of the control system, while the slave street light is controlled by the master street light. The master-slave feature classification matrix is a matrix that arranges and arranges the feature components of the master street light and the slave street light according to categories. It is used to describe the feature distribution of the master and slave street lights. The master-slave collaborative state space is a mathematical model used to describe the state transition relationship and control relationship between the master street light and the slave street light. It is usually composed of state equations and control equations. The master-slave collaborative learning environment is a virtual simulation environment used to simulate the collaborative control process between the master street light and the slave street light. It is based on the master-slave collaborative state space and role setting. The hierarchical control architecture is an organizational structure of the control system, including the master control network and the slave control network. The master control network is responsible for decision-making, and the slave control network is responsible for execution. The illumination-driven decision unit is a module embedded in the master control network, which adjusts the control strategy of the master street light according to the real-time environmental light conditions. The master control correction instruction is a control instruction generated based on the hierarchical control architecture for online evaluation of the master street light control strategy, used to optimize the dimming effect of the master street light. The master control adjustment sequence is the control signal sequence passed down by the master control network, used to adjust the brightness of the master street light. The slave control deviation sequence is the sequence of response deviations calculated by the slave control network for each slave street light, used to describe the difference between the actual brightness and the expected brightness of the slave street light. Time alignment and phase correction is the process of time synchronization and phase matching of the master control adjustment sequence and the slave control deviation sequence to ensure the synchronization of the brightness changes of the master and slave street lights. The synchronization control parameter is a parameter used to describe the synchronization response relationship between the master and slave street lights, including time offset and phase difference. The control correction amount is the result of the overall correction of the master and slave street light control strategies based on the synchronization control parameter, including brightness correction amount and phase correction amount. The brightness correction amount is the correction amount of the brightness adjustment value of the master and slave street lights, used to optimize the brightness output of the street light. The phase correction amount is the correction amount of the phase difference of the brightness change of the master and slave street lights, used to ensure the synchronization of the brightness change of the street light.

[0066] Further, the initial control sequence is compensated and corrected based on the control correction amount to obtain a target control sequence, and the target control sequence is respectively issued to the dimming driver of each street light through wireless networking to output a collaborative dimming control result, including: The control correction amount is subjected to master-slave decomposition operation to extract the brightness correction parameter of the master street light and the phase following parameter of each slave street light to obtain a master-slave correction parameter matrix; The master-slave correction parameter matrix is subjected to compensation superposition operation with the initial control sequence, and the control instruction of the master street light is subjected to priority marking to obtain a target control sequence; Synchronization following instructions of the slave street lamp are generated according to the main street lamp control instructions in the target control sequence, and a control timing relationship between the master and slave street lamps is established to obtain a master-slave cooperative control table; The control instructions in the master-slave cooperative control table are converted into a wireless network communication format to generate control data frames of the master street lamp and response data frames of the slave street lamp, and a communication data frame sequence is obtained; A real-time communication scheduling strategy is constructed based on the communication data frame sequence, the master street lamp data is set to be transmitted preferentially, and the slave street lamps are sequentially transmitted according to a phase difference order to obtain a communication scheduling sequence; The communication scheduling sequence is packaged according to an address coding rule of the light driving driver, and a synchronization trigger signal is added to obtain a driver control instruction; The driver control instruction is allocated network bandwidth to ensure that the control instruction of the master street lamp is preferentially issued, and to ensure that the control instructions of the slave street lamps are sequentially executed to obtain a time-sharing control sequence; Based on the time-sharing control sequence, control instructions are respectively issued to the light driving driver of the master street lamp and each slave street lamp through a wireless network, and execution feedback of each street lamp is collected to output a cooperative light control result.

[0067] It should be noted that the control correction amount is the result of overall correction of the master-slave street lamp control strategy based on the synchronization control parameter, including the brightness correction amount and the phase correction amount. The initial control sequence is the control instruction sequence of the system before correction, used to describe the initial brightness adjustment value of the master-slave street lamp. The target control sequence is the control instruction sequence obtained by compensating and correcting the initial control sequence based on the control correction amount, used to realize the collaborative dimming of the master-slave street lamp. The master-slave decomposition operation refers to the process of extracting the brightness correction parameter of the master street lamp and the phase following parameter of the slave street lamp from the control correction amount. The master-slave correction parameter matrix is a matrix that arranges and arranges the brightness correction parameter of the master street lamp and the phase following parameter of the slave street lamp according to the category. The compensation superposition operation refers to the process of superimposing the master-slave correction parameter matrix and the initial control sequence to generate the target control sequence. The priority mark is a mark for the control instruction of the master street lamp to ensure that it has the highest execution priority in the target control sequence. The synchronization following instruction is the slave street lamp control instruction generated based on the master street lamp control instruction in the target control sequence, used to ensure that the brightness change of the slave street lamp is synchronized with the master street lamp. The master-slave collaborative control table is a table used to describe the control timing relationship between the master-slave street lamp, including the control instruction and its execution time of the master street lamp and the slave street lamp. The wireless network communication format conversion is the process of converting the control instruction in the master-slave collaborative control table into a data format suitable for wireless network transmission. The communication data frame sequence is the sequence generated after the control instruction is converted into a data frame, used for transmission through the wireless network. The real-time communication scheduling strategy is a strategy for managing the transmission of the communication data frame sequence, usually including master street lamp data priority transmission and slave street lamp sequential transmission. The communication scheduling sequence is a communication data frame transmission order list generated based on the real-time communication scheduling strategy. The driver control instruction is the instruction generated by packaging the communication scheduling sequence according to the address coding rules of the dimming driver, used to control the dimming driver of the street lamp. The time-sharing control sequence is a driver control instruction transmission order list generated based on network bandwidth allocation, ensuring that the control instruction of the master street lamp is preferentially issued. The collaborative dimming control result is the actual brightness adjustment result of the master-slave street lamp after the control instruction is issued through the wireless network based on the time-sharing control sequence.

[0068] Further, the system further comprises an abnormal alarm module 80. The abnormal alarm module 80 is configured to determine that the street lamp has failed when the real-time operation data of the street lamp deviates from the preset normal operation parameter by more than a threshold value, generate corresponding alarm information according to the fault type, send the alarm information to the monitoring center through wireless networking, so that the monitoring center generates a maintenance task and distributes it to a maintenance personnel according to the received alarm information, and updates the state of the street lamp to fault.

[0069] It should be noted that the abnormal alarm module 80 is a functional module in the system, which is used to monitor the real-time running data of the street lamp, generate an alarm information when an abnormality is found and send it to the monitoring center. The real-time running data is the current state data collected by the street lamp during operation, including brightness, battery power, working temperature and other parameters. The normal running parameter is the parameter range of the street lamp in the normal working state preset by the system, which is used to judge whether the street lamp is in normal state. The deviation threshold is the maximum value of the deviation between the real-time running data and the normal running parameter allowed by the system, and if it exceeds the value, it is determined as abnormal. The fault type is the fault category classified according to the deviation between the real-time running data and the normal running parameter, such as brightness abnormality, insufficient battery power and the like. The alarm information is the notification generated by the abnormal alarm module 80, which contains the fault type, the position of the street lamp, the fault time and the like, and is used to prompt the monitoring center to process. The wireless networking is the network architecture of the system connecting the street lamp and the monitoring center through wireless communication technology, which is used for data transmission and instruction issuing. The monitoring center is the core management platform of the system, which is used to receive the alarm information, generate the maintenance task, distribute the maintenance personnel and update the state of the street lamp. The maintenance task is the specific maintenance instruction generated by the monitoring center according to the alarm information, including the position of the fault street lamp, the fault type and the maintenance priority.

[0070] The above is only the preferred embodiment of the present application, and does not limit the patent scope of the present application, and any equivalent structure or equivalent process transformation using the content of the specification and drawings of the present application, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A solar street lamp wireless networking intelligent dimming control system, characterized in that, The system comprises a data acquisition module, an intelligent cloud management module and a dimming control module; The data acquisition module is in communication connection with the intelligent cloud management module, and the intelligent cloud management module is in communication connection with the dimming control module; The data acquisition module is configured to acquire real-time operation data, real-time environmental illumination condition data and road segment demand data of a plurality of solar street lamps, wherein the plurality of solar street lamps realize data interaction through wireless networking. The intelligent cloud management module is configured to determine a use state of each solar street lamp according to the real-time operation data, wherein the use state comprises turning on, turning off, dimming and failure, and determine a brightness adjustment value of each solar street lamp according to the real-time environmental illumination condition data, the road segment demand data and the use state of each solar street lamp based on a reinforcement learning algorithm, and output an adjustment signal corresponding to the brightness adjustment value to the dimming control module. The dimming control module is configured to adjust a corresponding solar street lamp brightness value according to the adjustment signal.

2. The solar street light wireless networking intelligent dimming control system according to claim 1, wherein, The reinforcement learning algorithm comprises the following steps: Collecting historical illumination data of each road segment, including daily environmental illumination condition data, road segment demand data and use state of each solar street lamp; Predicting lighting demand of each road segment in a future preset time according to the historical illumination data, and calculating actual dimming demand of each road segment according to the lighting demand of each road segment in the future preset time and real-time environmental illumination condition data of each road segment; Taking the actual dimming demand of each road segment, the current position of each street lamp and the real-time environmental illumination condition between road segments as a state, taking adjustment of street lamp brightness between different road segments as an action, defining a reward function in terms of the degree of meeting the actual dimming demand of each road segment and the actual dimming demand of each road segment, and training a DQN network; Generating a brightness adjustment value of each solar street lamp in real time based on the trained DQN network.

3. The solar street light wireless networking intelligent dimming control system according to claim 2, characterized in that, The formula of the DQN network is as follows: In the formula, s is a current state parameter set, including actual dimming demand of each road segment, current position of each street lamp and real-time environmental illumination condition between road segments; a is a current action parameter set, including a strategy of adjusting street lamp brightness between road segments; Q(s, a) is a Q value of the action parameter set a under the state parameter set s; α is a learning rate; r is an immediate reward of the current action; γ is a discount factor; s' is a next state parameter set; a' is an action parameter set under the next state; and Q'(s, a) is an updated Q value.

4. The solar street light wireless networking intelligent dimming control system according to claim 3, characterized in that, The formula of the reward function is as follows: In the formula, R is a reward value; α and β are weight coefficients; N s is the actual satisfied road section dimming demand; N demand is the total dimming demand of the road section; T t is the actual dimming response time; T max is the maximum allowed dimming response time.

5. The solar street light wireless networking intelligent dimming control system according to claim 1, wherein, The system further comprises a feature solving module, a parameter analysis module, a control correction module and an instruction output module; The feature solving module is configured to perform multi-dimensional dynamic feature extraction on the real-time operation data, the real-time environmental illumination condition data and the road segment demand data of the plurality of solar street lamps, obtain a system operation feature vector, and solve a state feedback equation to obtain a mapping relationship matrix of a master street lamp and a plurality of slave street lamps. The parameter analysis module is configured to input the mapping relationship matrix and the brightness adjustment value of each solar street lamp into a target control model to perform control parameter analysis and obtain an initial control sequence. The control correction module is configured to input the system operation feature vector and the initial control sequence into a preset correction model to perform master-slave street lamp control correction and obtain a control correction amount. The instruction output module is configured to compensate and correct the initial control sequence based on the control correction amount, obtain a target control sequence, and output a cooperative dimming control result by respectively issuing the target control sequence to the dimming drivers of each street lamp through wireless networking.

6. The solar street light wireless networking intelligent dimming control system according to claim 5, wherein, The system operation feature vector is obtained by performing multi-dimensional dynamic feature extraction on the real-time operation data, real-time environmental lighting condition data, and road segment demand data of the plurality of solar street lamps, and a state feedback equation is solved to obtain a mapping relationship matrix of the master street lamp and the plurality of slave street lamps, including: The real-time operation data, real-time environmental lighting condition data, and road segment demand data of the plurality of solar street lamps are grouped according to master-slave street lamps, the data matrix of the master street lamp is marked as a master control matrix, and the data matrix of the plurality of slave street lamps is marked as a slave control matrix; The master control matrix and the slave control matrix are standardized to obtain standardized time sequence data of the master-slave street lamps, the standardized time sequence data is subjected to frequency spectrum analysis and time domain feature calculation to obtain a system operation feature vector; A state space equation of the master-slave street lamps is established according to the system operation feature vector, a control equation containing a state feedback term and a following feedback term is constructed, a master-slave coupled state equation is obtained, and stability analysis is performed on the master-slave coupled state equation to calculate a feedback gain parameter and obtain cooperative control parameters of the master-slave street lamps; A master-slave mapping function is constructed according to the cooperative control parameters, a state transition matrix is calculated to obtain a following control amount of the slave control street lamp to the master control street lamp, and the following control amount is subjected to symmetric matrix transformation to establish a master-slave state correspondence relationship and obtain a state conversion matrix; The state conversion matrix is subjected to parameter calibration to obtain the mapping relationship matrix of the master street lamp and the plurality of slave street lamps.

7. The solar street light wireless networking intelligent dimming control system according to claim 6, wherein, The mapping relationship matrix and the brightness adjustment value of each solar street lamp are input into a target control model to perform control parameter analysis and obtain an initial control sequence, including: A master street lamp mapping vector and a plurality of slave street lamp mapping vectors are extracted based on the mapping relationship matrix, and a master-slave mapping vector group is generated; The master-slave mapping vector group is input into a target control model, a state transition cost function of the master-slave street lamps is established to obtain a state cost matrix, and the state cost matrix is subjected to weighting processing according to the brightness adjustment value of each solar street lamp to construct a cooperative control target function of the master-slave street lamps and obtain a control cost function; An optimal control strategy is obtained by performing dynamic programming solving based on the control cost function, and the optimal control strategy is subjected to time discretization processing to respectively extract the control amount of the master street lamp and the slave street lamp at each time point and obtain a discrete control sequence; The master-slave synchronism analysis is performed on the discrete control sequence, a brightness synchronization constraint condition between master and slave street lamps is constructed, a synchronization control constraint is obtained, the discrete control sequence is modified according to the synchronization control constraint, a collaborative control instruction set of the master and slave street lamps is established, and an initial control sequence is obtained.

8. The solar street light wireless networking intelligent dimming control system according to claim 7, characterized in that, The system operation feature vector and the initial control sequence are input into a preset modification model to perform master-slave street lamp control modification, and a control modification amount is obtained. The system operation feature vector is subjected to master-slave component extraction, the feature components of the master street lamp and the feature components of each slave street lamp are classified respectively, and a master-slave feature classification matrix is obtained. A master-slave collaborative state space is established according to the master-slave feature classification matrix and the initial control sequence, the master street lamp is set as a decision maker, and the slave street lamp is set as an execution subject, and a master-slave collaborative learning environment is obtained. The master-slave collaborative learning environment is input into a preset modification model, a hierarchical structure of the master control network and the slave control network is constructed, a decision unit of light driving is embedded in the master control network, and a hierarchical control architecture is obtained. Based on the hierarchical control architecture, the control strategy of the master street lamp is subjected to online evaluation, a master control modification instruction is generated, a control signal is transmitted downward through the master control network, a master control adjustment sequence is obtained, the control parameters of the multiple slave street lamps are collaboratively optimized according to the master control adjustment sequence, the response deviation of each slave street lamp is calculated through the slave control network, and a slave control deviation sequence is obtained. The master control adjustment sequence and the slave control deviation sequence are subjected to timing alignment and phase correction, a synchronization response relationship between the master and slave street lamps is established, a synchronization control parameter is obtained, and the control strategy of the master and slave street lamps is integrally modified according to the synchronization control parameter, and a control modification amount containing a brightness modification amount and a phase modification amount is constructed.

9. The solar street light wireless networking intelligent dimming control system according to claim 8, wherein, The initial control sequence is compensated and modified based on the control modification amount, a target control sequence is obtained, the target control sequence is respectively distributed to the dimming drivers of each street lamp through wireless networking, and a collaborative dimming control result is output. The control modification amount is subjected to master-slave decomposition operation, the brightness modification parameter of the master street lamp and the phase following parameter of each slave street lamp are extracted, and a master-slave modification parameter matrix is obtained. The master-slave modification parameter matrix and the initial control sequence are subjected to compensation superposition operation, and the control instruction of the master street lamp is subjected to priority marking, and a target control sequence is obtained. The master street lamp control instruction in the target control sequence is used to generate a synchronous following instruction of the slave street lamp, a control timing relationship between the master and slave street lamps is established, and a master-slave collaborative control table is obtained. The control instruction in the master-slave collaborative control table is subjected to wireless network communication format conversion, a control data frame of the master street lamp and a response data frame of the slave street lamp are generated, and a communication data frame sequence is obtained. A real-time communication scheduling strategy is constructed based on the communication data frame sequence, the data of the master street lamp is set to be transmitted preferentially, and the slave street lamps are sequentially transmitted according to the phase difference, and a communication scheduling sequence is obtained. The communication scheduling sequence is packaged according to the address coding rule of the dimming driver, and a synchronous trigger signal is added, and a driver control instruction is obtained. The driver control instructions are allocated network bandwidth, control instructions of the master street lamp are preferentially issued, and control instructions of the slave street lamps are sequentially executed to obtain a time-sharing control sequence; Control instructions are issued to the dimming drivers of the master street lamp and the slave street lamps based on the time-sharing control sequence through a wireless network, execution feedback of the street lamps is collected, and a cooperative dimming control result is output.

10. The solar street light wireless networking intelligent dimming control system according to claim 1, wherein, The system further comprises an abnormality alarm module; When deviation of real-time operation data of the street lamp from preset normal operation parameters exceeds a threshold value, the abnormality alarm module determines that the street lamp has a fault, generates corresponding alarm information according to a fault type, sends the alarm information to the monitoring center through wireless networking, so that the monitoring center generates a maintenance task and distributes it to a maintenance personnel according to the received alarm information, and simultaneously updates a state of the street lamp to fault.

Images