Road lighting control method based on solar street lamp and intelligent control system

Abstract

The invention provides a road illumination control method based on solar street lamps and an intelligent control system. The method comprises the following steps: determining illumination attribute information of each solar street lamp in a preset time period according to historical working condition data and current working condition data of each solar street lamp and historical flow data and current flow data near a road; determining target required electric energy of each solar street lamp based on the illumination attribute information, and marking the solar street lamp of which the current electric energy of the power supply does not meet the target required electric energy as a to-be-powered street lamp; selecting other solar street lamps of the street lamp to be powered within a preset power supply threshold range as street lamps to be powered; the target required electric energy of the to-be-powered street lamp, the current working condition data of the to-be-powered street lamp and the current flow data near the road serve as input of a preset power supply model, and a power supply strategy for supplying power to the to-be-powered street lamp is obtained; power can be supplied to the street lamp with insufficient power supply through the support of other street lamp power supplies so as to meet the illumination brightness and duration of the street lamp.

Description

Technical Field

[0001] The present invention relates to the field of solar street lamps, and in particular to a road lighting control method and an intelligent control system based on solar street lamps. Background Art

[0002] Solar street lights can use solar panels to convert light energy into electrical energy during the day and store the electrical energy in batteries. At night, when street lighting is needed, solar street lights can release the electrical energy in the batteries through intelligent controllers to provide power for the solar street light fixtures.

[0003] At present, solar street lights are widely used in road lighting due to their energy-saving and environmentally friendly features. However, most existing solar street lights work independently and lack an effective collaborative control mechanism. When the power supply of a street light is insufficient due to weather or other reasons, it cannot get timely support from the power supply of other street lights, which may cause the brightness of the street light to decrease or even go out, affecting the road lighting effect. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a road lighting control method and intelligent control system based on solar street lamps, which solves the problem of the lack of effective coordinated control mechanism of solar street lamps in the prior art.

[0005] In a first aspect, the present application provides a road lighting control method based on solar street lamps, which is used for an intelligent control system of solar street lamps. The intelligent control system includes a plurality of solar street lamps, and the power sources of the plurality of solar street lamps are electrically connected in sequence along the extension direction of the road path. The method includes:

[0006] Determine the lighting attribute information of each solar street lamp within a preset time period according to the historical working condition data, current working condition data of each solar street lamp and the historical traffic data and current traffic data near the road;

[0007] Determine the target power requirement of each solar street lamp based on the lighting attribute information, and mark the solar street lamp whose current power supply does not meet the target power requirement as a street lamp to be received;

[0008] Selecting other solar street lamps within a preset power supply threshold range among the street lamps to be supplied as street lamps to be supplied;

[0009] The target power demand of the waiting street lamp, the current working condition data of the waiting street lamp and the current traffic data near the road are used as inputs of a preset power supply model to obtain a power supply strategy for supplying power to the waiting street lamp.

[0010] In one embodiment, the historical operating condition data includes the power status, lighting brightness status, charging and discharging status and weather status of the street lamp in a preset time period; the lighting attribute information of each solar street lamp in the preset time period is determined based on the historical operating condition data, current operating condition data of each solar street lamp and the historical flow data and current flow data near the road, specifically including:

[0011] Determine the time series of illumination variables of the solar street lamp based on the power state, lighting state, charge and discharge state and the historical flow data; the time series of the time series of illumination variables includes a time series of power change, a time series of lighting brightness change, a time series of charge and discharge change and a time series of flow change;

[0012] Inputting the time series of the illumination variable into a pre-trained long short-term memory network to determine an illumination feature vector corresponding to the illumination variable;

[0013] The illumination feature vector and each illumination variable in the historical working condition data and the historical traffic data are respectively used as topological nodes of the Bayesian network to establish a lighting prediction model of the Bayesian network structure;

[0014] The current working condition data of each solar street lamp and the current traffic data near the road are used as inputs of the lighting prediction model, and the lighting attribute information of each solar street lamp within a preset time period is output.

[0015] In one embodiment, inputting the time series of the illumination variables into a pre-trained long short-term memory network to determine the illumination feature vector corresponding to the illumination variable specifically includes:

[0016] Dividing the time series into a first training set and a first test set according to a preset ratio;

[0017] Training the initial long short-term memory network based on the first training sample set to obtain a trained long short-term memory network;

[0018] The first test set is input into the long short-term memory network for training to obtain predicted values ​​of illumination variables, and based on the predicted values ​​of illumination variables and the long short-term memory network, an illumination feature vector corresponding to the illumination variable is obtained.

[0019] In one embodiment, the feature vector and each illumination variable in the historical operating condition data and the historical traffic data are respectively used as topological nodes of the Bayesian network to establish a lighting prediction model with a Bayesian network structure, specifically including:

[0020] Perform edge addition operation on the preset initial Bayesian network according to the node order, and select the Bayesian network structure with the highest score value after the edge modification operation as the current Bayesian network structure based on the preset scoring function;

[0021] The conditional probability of each node in the current Bayesian network structure is estimated by an expectation maximization algorithm;

[0022] The edge adding operation is repeated on the current Bayesian network structure, and the current Bayesian network structure whose conditional probability satisfies a preset iteration termination condition is used as the final Bayesian network structure for constructing the lighting prediction model.

[0023] In one embodiment, the edge adding operation is performed on the preset initial Bayesian network in the order of nodes, and the Bayesian network structure with the highest score value after the edge modification operation is selected as the current Bayesian network structure based on a preset scoring function, specifically including:

[0024] Determine the node order between the illumination feature vector and each illumination variable according to a preset order rule;

[0025] An empty network structure is used as an initial Bayesian network structure, and an edge adding operation of child nodes is performed on the initial Bayesian network structure according to the node order, so as to obtain a Bayesian network structure with different parent nodes added;

[0026] The scores of the Bayesian network structures with different parent nodes added to the edges are evaluated based on a preset scoring function, and the parent node set of the Bayesian network structure with the highest scoring value is selected as the final parent node of the child node to establish the current Bayesian network structure.

[0027] In one embodiment, the preset power supply model is trained by the following method:

[0028] Determine the lamps to be supplied and the lamps to be supplied in the training data set based on the historical operating condition data, and obtain the target required electric energy of the lamps to be supplied;

[0029] A time node at which the power change of the waiting-for-received lamp exceeds a threshold value under historical working condition data is defined as a state change node;

[0030] Determine the state transition probability between two adjacent state change nodes based on the target required power of the waiting-for-received lamp;

[0031] Determine the value function corresponding to the execution of the initialization power supply strategy for two adjacent states according to the state transition probability;

[0032] The initialized power supply strategy is updated according to the value function and a new power supply strategy is generated, until the value function corresponding to the new power supply strategy meets the preset termination condition, and the corresponding new power supply strategy is used as the optimal power supply strategy.

[0033] In one embodiment, the expression for iterating the value function is:

[0034]

[0035] Among them, V π (s) is the value function at state s, V π (s^) is the value function at the state s^; a represents the execution action corresponding to the power supply strategy, and its numerical expression is a=(e 1 ,e 2 ,……,e n ), where e 1 is the amount of electricity supplied by the first waiting circuit lamp to the waiting circuit lamp, e 2 is the amount of electricity that the second waiting circuit lamp supplies to the waiting circuit lamp, e n is the amount of power supplied by the nth waiting-to-supply lamp to the waiting-to-receive lamp; π represents the power supply strategy corresponding to the switch from state s to state s^; π(a|s) represents the probability of taking action a at state s under power supply strategy v; P(s^|s, a) is the probability of transferring to state s^ after adopting power supply strategy a at state s; R(s,a,s) is the reward function, R(s,a,s^)=α·ΔE+β·ΔL+σ·ΔS, where ΔE is the change in the remaining power of the waiting-to-supply lamp, ΔL is the ratio between the actual brightness of the waiting-to-receive lamp and the target brightness, ΔS is the number of power supplies, and α, β and σ are the corresponding weight coefficients respectively; γ is the discount factor, 0≤γ≤1.

[0036] In a second aspect, the present application provides a road lighting control device based on a solar street lamp, comprising:

[0037] A lighting attribute determination module, used to determine the lighting attribute information of each solar street lamp within a preset time period based on the historical working condition data, current working condition data of each solar street lamp and the historical traffic data and current traffic data near the road;

[0038] A waiting-for-receiving street lamp determination module is used to determine the target required electric energy of each solar street lamp based on the lighting attribute information, and mark the solar street lamp whose current electric energy of the power source does not meet the target required electric energy as a waiting-for-receiving street lamp;

[0039] A module for determining a street lamp to be powered, used for selecting other solar street lamps within a preset power supply threshold range among the street lamps to be powered as street lamps to be powered;

[0040] The power supply strategy generation module is used to take the target power demand of the waiting street lamp, the current working condition data of the waiting street lamp and the current traffic data near the road as inputs of a preset power supply model to obtain a power supply strategy for powering the waiting street lamp.

[0041] In a third aspect, the present application provides an intelligent control system, comprising a plurality of solar street lights, the power sources of the plurality of solar street lights being electrically connected in sequence along the extension direction of a road path, and further comprising a processor and a memory; wherein the memory stores a computer program, and the computer program is used to be loaded by the processor and execute a road lighting control method based on solar street lights as described in any one of the first aspects.

[0042] In a fourth aspect, the present application provides a computer-readable storage medium storing instructions for a processor to load and execute a road lighting control method based on solar street lamps as described in any one of the first aspects.

[0043] In the road lighting control method and intelligent control system based on solar street lamps of the present embodiment, by connecting the power sources of the solar street lamps in sequence, when the solar power source of a street lamp is insufficiently supplied due to weather or power failure, it can be powered by the support of the power sources of other street lamps to meet the lighting brightness and duration of the street lamp, thereby ensuring that the road has a good lighting effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0045] Figure 1 Schematic diagram of the process of controlling road lighting based on solar street lamps in an embodiment of the present application.

[0046] Figure 2 This is a schematic diagram of the structure of a road lighting control device based on a solar street lamp in an embodiment of the present application. DETAILED DESCRIPTION

[0047] The specific embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the description of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] In the description of the present invention, unless otherwise clearly specified and limited, the terms "set", "install", "connection" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0049] The directions or positional relationships indicated by terms such as “upper”, “lower”, “left”, “right”, “front”, “back”, “top”, “bottom”, “inside” and “outside” are based on the directions or positional relationships shown in the accompanying drawings, or are the directions or positional relationships in which the inventive product is usually placed when used. They are only for the convenience of description and simplified description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as a limitation on the present invention.

[0050] The terms "first", "second", "third" and the like are merely used to distinguish elements of similar nature, and do not indicate or imply relative importance or a particular order.

[0051] The terms "comprises," "comprising," or any other variations thereof, are intended to cover a non-exclusive inclusion of the elements listed and may also include additional elements not expressly listed.

[0052] like Figure 1 As shown, this embodiment provides a road lighting control method based on solar street lamps, which is used for an intelligent control system of solar street lamps. The intelligent control system includes multiple solar street lamps, and the power sources of the multiple solar street lamps are electrically connected in sequence along the extension direction of the road path. The method includes:

[0053] Step S10: determining the lighting attribute information of each solar street lamp within a preset time period according to the historical working condition data, current working condition data of each solar street lamp and the historical traffic data and current traffic data near the road;

[0054] Step S20: determining the target power requirement of each solar street lamp based on the lighting attribute information, and marking the solar street lamp whose current power supply does not meet the target power requirement as a street lamp to be received;

[0055] Step S30: selecting other solar street lamps within a preset power supply threshold range as the street lamps to be powered;

[0056] Step S40: taking the target power demand of the circuit lamps to be supplied, the current operating condition data of the circuit lamps to be supplied and the current traffic data near the road as inputs of a preset power supply model to obtain a power supply strategy for supplying power to the circuit lamps to be supplied; and controlling each of the circuit lamps to be supplied to execute the power supply strategy to supply power to the circuit lamps to be supplied.

[0057] In the road lighting control method based on solar street lamps of the present embodiment, by connecting the power sources of the solar street lamps in sequence, when the solar power source of a street lamp is insufficiently supplied due to weather or power failure, it can be powered by the support of the power sources of other street lamps to meet the lighting brightness and duration of the street lamp, thereby ensuring that the road has a good lighting effect.

[0058] Step S10: determining the lighting attribute information of each solar street lamp within a preset time period according to the historical working condition data, current working condition data of each solar street lamp and the historical traffic data and current traffic data near the road.

[0059] The historical operating condition data includes the power state, lighting brightness state, charging and discharging state and the weather state of the street lamp in a preset time period. The current operating condition data is the power state, lighting brightness state, charging and discharging state and the weather state of the street lamp up to the current moment.

[0060] A power meter chip is installed in the power module of the solar street light, which can monitor the remaining power, voltage, current and other parameters of the power supply in real time. The controller of the solar street light can receive the remaining power, voltage, current and other parameter data transmitted by the power meter chip at a certain time period. In addition, the solar street light is also provided with a memory for storing the above parameter data, which can upload the stored parameter data to a remote server through wireless communication modules such as ZigBee and GPRS, so as to process and analyze the historical working condition data.

[0061] A light sensor such as a photodiode (or photoresistor) is installed near the solar street light. The light sensor is used to collect the lighting brightness of the installed solar street light and transmit the collected lighting brightness data to the controller of the solar street light. Similarly, the controller can also store and transmit the collected lighting brightness data to a remote server.

[0062] Current sensors and voltage sensors are installed between the solar panel and the power supply, and between the power supply and the load. The current sensors and voltage sensors are used to monitor the current data and voltage data during the charging and discharging process, so that the charging and discharging state (charging, discharging, idle) can be determined according to the changes in the current data and voltage data. Similarly, the controller can also store and transmit the collected current data and voltage data to a remote server.

[0063] Each solar street light has a specific code number when it is installed, and its geographical location coordinates are bound to its code number in the remote server. Through the API interface provided by the meteorological data platform, the weather data corresponding to the location of the solar street light can be queried according to the predetermined time period and geographical location coordinates, and the collected weather data can be stored and transmitted to the remote server through the controller.

[0064] Historical traffic data includes vehicle traffic data and pedestrian traffic data within a preset time period. Current traffic data includes vehicle traffic data and pedestrian traffic data at the current moment.

[0065] Each street light is equipped with a camera with AI recognition function, which can collect the vehicle flow data and pedestrian flow data of the solar street light in real time, and upload the collected vehicle flow data and pedestrian flow data to the remote server through the vehicle flow data and pedestrian flow data. It is understandable that the vehicle flow data or pedestrian flow data can also be obtained through radar sensors or geomagnetic induction coils.

[0066] The working condition data of solar street lights (such as light intensity, battery power) and road traffic data have time series characteristics. The long short-term memory network can capture the long-term dependency of data in the time dimension (such as the alternation of day and night, the traffic pattern of weekdays / weekends), and can predict the lighting demand in the future period based on the time series pattern of historical data (such as seasonality and periodicity), for example, it can judge in advance that the lighting level needs to be increased during the high traffic period at night. However, the battery charging and discharging status and weather status do not have the characteristics of time series, so this embodiment uses the long short-term memory network combined with the Bayesian network structure to construct a lighting prediction model, and outputs the lighting attribute information within the preset time period.

[0067] The method of determining the lighting attribute information of each solar street lamp within a preset time period according to the historical working condition data, current working condition data of each solar street lamp and the historical flow data and current flow data near the road specifically includes:

[0068] Step S101: determining the time series of illumination variables of the solar street lamp based on the power state, lighting state, charge-discharge state and the historical flow data; the time series of the time series of illumination variables includes a power change time series, a lighting brightness change time series, a charge-discharge change time series and a flow change time series;

[0069] Step S102: inputting the time series of the illumination variables into a pre-trained long short-term memory network to determine an illumination feature vector corresponding to the illumination variable;

[0070] Step S103: using the illumination feature vector and each illumination variable in the historical working condition data and the historical traffic data as topological nodes of the Bayesian network, and establishing a lighting prediction model of the Bayesian network structure;

[0071] Step S104: taking the current working condition data of each solar street lamp and the current traffic data near the road as inputs of the lighting prediction model, and outputting the lighting attribute information of each solar street lamp within a preset time period.

[0072] In step S101, the power change time series records the power change between two adjacent time points between multiple time points. The lighting brightness change time series records the lighting brightness change between two adjacent time points between multiple time points. The charge and discharge state can be represented by a binary value, for example, charging is 1 and discharging is 0. The traffic flow change time series records the changes in vehicle flow and pedestrian flow between two adjacent time points between multiple time points.

[0073] In step S102, the time series of the illumination variables is input into a pre-trained long short-term memory network to determine the illumination feature vector corresponding to the illumination variable, specifically including:

[0074] Step S1021: Divide the time series into a first training set and a first test set according to a preset ratio;

[0075] Step S1022: training the initial long short-term memory network based on the first training sample set, and obtaining a trained long short-term memory network;

[0076] Step S1023: input the first test set into the long short-term memory network for training to obtain predicted values ​​of illumination variables, and obtain illumination feature vectors corresponding to the illumination variables based on the predicted values ​​of the illumination variables and the long short-term memory network.

[0077] In step S1021, the time series can be divided into a first training set and a first test set in a ratio of 8:2 (or other ratios according to actual needs), wherein the first training set accounts for 80% and the first test set accounts for 20%. The first training sample set is used to train the long short-term memory network, and the first test set is used to test the long short-term memory network.

[0078] Before dividing the time series into the first training set and the first test set, the test work data needs to be cleaned and normalized. When cleaning the data, interpolation methods (such as linear interpolation and polynomial interpolation) can be used to fill missing values, and statistical methods (such as the 3σ principle) can be used to identify and process outliers. When normalizing the data, Min-Max normalization and Z-Score normalization can be used to normalize the time series to a specific range (such as [-1,1]), and the time series data can be divided into windows of fixed length, with each window as a sample.

[0079] In step S1022, the preset initial long short-term memory network is defined by the following method: determining the number of neurons in the input layer, the number of neurons and input shape parameters of the LSTM layer, the number of hidden layers and the activation function of the initial long short-term memory network according to the characteristic dimension of the time series.

[0080] The structural parameters of the initial recurrent neural network are customized based on the characteristics of time series data. Specifically. Specifically, the number of neurons in the input layer is mainly set according to the characteristic dimension of the time series. In this case, there are four dimensions: the time series of power changes, the time series of lighting brightness changes, the time series of charging and discharging changes, and the time series of flow changes. Therefore, the input layer is set to 4 neurons. The number of neurons in the LSTM core layer usually needs to be larger than the input layer scale to capture complex patterns. Here, a bidirectional LSTM layer (128 units) + Dropout (0.2) is used. The input dimension parameter consists of three elements: sample batch size, time step (customizable according to the acquisition cycle), and feature dimension (set to 2 in this scenario). Considering the nonlinear coupling characteristics between multiple variables, the model is configured with 3 hidden layers. In terms of activation function selection, the gating mechanism uses the Sigmoid function to realize information regulation in the 0-1 value range, and the state update uses the tanh function to retain positive and negative bidirectional features. The output layer uses a fully connected layer to generate a 64-dimensional feature vector.

[0081] During the model training phase, time series data is processed time-step by time step through the LSTM layer to generate hidden states containing historical information. The synergy of the gating mechanism (input gate / forget gate / output gate) enables the model to effectively capture the dynamic dependencies between multiple variables, such as the weight adjustment of the battery status to the power supply strategy when the light is insufficient. The training process uses the back propagation algorithm combined with the optimizer to iteratively optimize the weights, and finally forms a long short-term memory network with time series prediction capabilities.

[0082] The lighting demand of solar street lamps is affected by multiple factors such as light, battery power, and traffic flow, and there may be complex nonlinear relationships between these factors (for example, when light is insufficient, battery power has a higher weight on the power supply strategy). The long short-term memory network can automatically learn these implicit association patterns through multiple layers of neurons and gating mechanisms. It can also process real-time data (such as current light intensity, real-time traffic) step by step, and dynamically update the internal state to support real-time control of the solar street lamp system. By adjusting the input sequence length and prediction step size, the long short-term memory network can predict the lighting demand in the next few hours or days, providing a basis for early decision-making for power supply strategy optimization.

[0083] In step S1023, the target feature vector can be obtained by weighted fusion according to the predicted value of the target variable and the output of the hidden layer of the long short-term memory network. Since the hidden layer generally outputs a hidden layer vector for predicting the trend of change, this hidden layer vector can be used as a reference for predicting the value of the next time step. In order to make the target feature vector have excellent prediction ability in the time dimension, the weight will be corrected in combination with the predicted value output after training the first test set.

[0084] Due to the potential causal logic of "weather → light → traffic → electricity consumption", this embodiment uses the K2 algorithm to add edges, gradually adding edges through greedy search, reducing the search space and improving efficiency. In addition, when an accident occurs, the user is allowed to specify the node order and the upper limit of the number of parent nodes, combining domain knowledge to avoid blind search.

[0085] In step S103, the feature vector and each illumination variable in the historical operating condition data and the historical traffic data are respectively used as topological nodes of the Bayesian network to establish a lighting prediction model of the Bayesian network structure, which specifically includes:

[0086] Step S1031: performing an edge adding operation on the preset initial Bayesian network according to the node order, and selecting the Bayesian network structure with the highest score value after the edge modification operation as the current Bayesian network structure based on the preset scoring function;

[0087] Step S1032: estimating the conditional probability of each node in the current Bayesian network structure by using an expectation maximization algorithm;

[0088] Step S1033: Repeat the edge adding operation on the current Bayesian network structure, and use the current Bayesian network structure whose conditional probability satisfies a preset iteration termination condition as the final Bayesian network structure for constructing the lighting prediction model.

[0089] In steps S1031-S1033, the K2 algorithm adopts a greedy search strategy, starting from an empty network structure (i.e., the initial Bayesian network), and gradually adds edges to build a network. Specifically, it considers the parent node set of each node in turn according to the predefined node order. For each node, the algorithm tries to add different parent node combinations and calculates the score of the network structure after adding these parent nodes. Then, the parent node set that can maximize the score improvement is selected as the final parent node of the node and the final Bayesian network structure is established.

[0090] In step S1032, since the "weather status" data cannot be directly measured, the weather data obtained through the meteorological data platform may have errors or delays, so it is necessary to further judge the weather status data based on hidden variables. In addition, sensor data such as light intensity and battery power may be missing. Therefore, the expectation maximization algorithm is used to estimate the conditional probability of each node in the current Bayesian network structure.

[0091] When the expectation maximization algorithm is used to estimate the conditional probability of each node, the parameters are initialized first: the conditional probability CPT of each node is randomly initialized according to the uniform distribution or Gaussian distribution (for example, if the "lighting status" node has the parent nodes "battery power" and "traffic data", the initial conditional probability CPT may be set to P (brightness | power, traffic) = uniform distribution). By combining historical experience, the Dirichlet prior is added to set the prior parameter α (for example: on sunny days, the solar panel charging efficiency is high, and the probability of the power state being high is 80%, at this time α = 8; on rainy days, the charging efficiency is low, and the probability of the power state being high is 10%, at this time α = 1).

[0092] Then the expectation calculation is performed: using the belief propagation algorithm, the current conditional probability CPT is used to infer the posterior distribution of the latent variable for each training sample.

[0093] Secondly, parameter maximization is performed: for each node X and its parent node combination U, the number of co-occurrences of X and U in the training set is counted using the likelihood function.

[0094] The formula for the likelihood function is:

[0095]

[0096] in: It represents the conditional probability parameter of the child node X taking the value of x under the condition that the parent node U takes the value of u after the t+1th iteration update; the x and x` nodes are different values ​​of the child node X; u is a specific value of the parent node U; D i Represents the observed data of the i-th sample; α x The corresponding child node X takes the value of x as the prior parameter, α x`The prior parameter corresponding to X takes the value x`.

[0097] Repeat the steps of expectation calculation and parameter maximization to update the state of the node until the logarithm of the likelihood function is less than the preset threshold. At this time, the conditional probability value of the node is the final value.

[0098] Step S20: determining the target power requirement of each solar street lamp based on the lighting attribute information, and marking the solar street lamp whose current power supply does not meet the target power requirement as a street lamp to be received.

[0099] By comparing the target power demand of each solar street lamp with the current power stored in the current state, if the current power is less than the target power demand, the corresponding solar street lamp can be marked as a waiting-for-power lamp.

[0100] Step S30: Select other solar street lamps within a preset power supply threshold range as the street lamps to be powered.

[0101] Usually, two other solar street lamps on both sides of the waiting lamp and electrically connected to it are used as the waiting lamps. If the power of the two adjacent solar street lamps does not meet the target power demand of the waiting lamp, the query range of the waiting lamps is further expanded (two or three on each side. Due to the large loss of power during long-distance transmission, the maximum number of waiting lamps is usually set to six).

[0102] Step S40: taking the target power demand of the circuit lamps to be supplied, the current operating condition data of the circuit lamps to be supplied and the current traffic data near the road as inputs of a preset power supply model to obtain a power supply strategy for supplying power to the circuit lamps to be supplied; and controlling each of the circuit lamps to be supplied to execute the power supply strategy to supply power to the circuit lamps to be supplied.

[0103] Since there are many uncertain factors in the power supply, such as weather changes and load fluctuations. This embodiment uses the Markov model to quantify these uncertainties through state transition probabilities, converting continuous power supply states (such as power, light intensity, load demand) into limited discrete states (such as "high power-strong light", "low power-weak light"), and uses state transition probabilities (such as the probability of changing from "cloudy" to "rainy") to track environmental changes in real time and flexibly adjust the power supply strategy.

[0104] The preset power supply model is trained by the following method:

[0105] Step S401: determining the lamps to be supplied and the lamps to be supplied in the training data set based on the historical operating condition data, and obtaining the target required electric energy of the lamps to be supplied;

[0106] Step S402: defining a time node at which the power change of the waiting lamp exceeds a threshold value under historical working condition data as a state change node;

[0107] Step S403: determining the state transition probability between two adjacent state change nodes based on the target required power of the lamp to be received;

[0108] Step S404: determining the cost function corresponding to executing the initialization power supply strategy in two adjacent states according to the state transition probability;

[0109] Step S405: updating the initialized power supply strategy according to the value function and generating a new power supply strategy, until the value function corresponding to the new power supply strategy meets the preset termination condition, and taking the corresponding new power supply strategy as the optimal power supply strategy.

[0110] In step S401, according to the charging and discharging relationship between the solar street lamps under the historical working condition data, the waiting street lamps and the waiting street lamps for supply in different time periods can be determined, and the power value of the waiting street lamps that just has enough power within the specified time is used as the target required power.

[0111] In step S402, when the waiting lamp is charging, the power consumption rate of its power source will be significantly reduced, or the charging speed will be significantly increased. By comparing the power change data of three connected time nodes with obvious abnormal changes, if the change rates of the three time nodes all exceed the threshold, the middle time node can be defined as the state change node. It can be understood that the time node whose change rate exceeds the threshold can be directly used as the state change node.

[0112] In step S403, the variables of the historical operating condition data are first discretized. For example, the power state is discretized into four levels (<20% critical, 20-40% low, 40-70% medium, >70% high). The lighting state is divided into shielding (<50lux), weak light (50-300lux), and strong light (>300lux). The charging state is divided into autonomous charging (sufficient light), passive discharge (lighting power consumption), receiving power (neighboring lights power supply), and power supply (transmitting power to neighboring lights). The flow data is divided into more, more, general, and less. The power supply mutual assistance threshold is set to (30% / 70% / 50%). When the power of street lamp A is less than 30%, the neighbor lamp search is triggered: the neighbor lamp B with the closest distance and power >70% is given priority, and the transfer probability increases by 30% (historical power supply success experience). If the power of B drops to 50%, it will automatically switch to the next nearest neighbor lamp C, and the transfer probability will be attenuated by 50% (to avoid excessive power consumption).

[0113] Then, based on the statistics of the frequency of the system transferring to different states under the same or similar states and actions, this is used as the state transition probability P(s^|s, a) to estimate the probability of transferring to the next state s^ after taking action a in state s.

[0114] The state s may be composed of the working condition data of the waiting lamp and the waiting lamp at a specified time (such as battery power, light intensity, etc.), the traffic data near the road, and the target power demand of the waiting lamp. For example, the state s may be expressed as s=(B r ,L r ,T r ,B si ,L si ,T si ,…,B sn ,L sn ,T sn ,F,E r ), where B r is the battery level of the waiting light, L r is the light intensity at the waiting lamp, T r is the lighting time of the waiting circuit lamp, B si , L si , T si are the battery capacity, light intensity and lighting duration of the i-th street lamp to be supplied, F is the traffic flow near the road, and 2 is the target power demand of the street lamp to be supplied.

[0115] In step S404, the initially defined value function V(s)=0 can be set, and the value function is updated based on the greedy strategy. The greedy strategy selects actions based on the state value function V(s) or the action value function Q(s,a). In the greedy strategy, the action that maximizes the value function in the current state is selected, and the probability of this action is 1, and the probability of other actions is 0.

[0116] In step S405, the expression for iterating the value function is:

[0117]

[0118] Among them, V π (s) is the value function at state s, V π (s^) is the value function at the state s^; a represents the execution action corresponding to the power supply strategy, and its numerical expression is a=(e 1 ,e 2 ,……,e n ), where e 1 is the amount of electricity supplied by the first waiting circuit lamp to the waiting circuit lamp, e 2 is the amount of electricity that the second waiting circuit lamp supplies to the waiting circuit lamp, en is the amount of power supplied by the nth waiting-to-supply lamp to the waiting-to-receive lamp; π represents the power supply strategy corresponding to the switch from state s to state s^; π(a|s) represents the probability of taking action a in state s under power supply strategy π; P(s^|s, a) is the probability of transferring to state s^ after adopting power supply strategy a in state s; R(s,a,s^) is the reward function, R(s,a,s^)=α·ΔE+β·ΔL+σ·ΔS, where ΔE is the change in the remaining power of the waiting-to-supply lamp, ΔL is the ratio between the actual brightness of the waiting-to-receive lamp and the target brightness, ΔS is the number of power supplies, and α, β and σ are the corresponding weight coefficients respectively; γ is the discount factor, 0≤γ≤1.

[0119] Each time a new power supply strategy is generated, the value function is iterated once, and the above steps are repeated until the value function converges. The optimal power supply strategy is determined based on the converged value function.

[0120] like Figure 2 As shown, based on the same inventive concept as the above embodiment, this embodiment also provides a road lighting control device based on a solar street lamp, including:

[0121] The lighting attribute determination module M10 is used to determine the lighting attribute information of each solar street lamp within a preset time period according to the historical working condition data, current working condition data of each solar street lamp and the historical flow data and current flow data near the road;

[0122] A waiting-for-receiving street lamp determination module M20 is used to determine the target required electric energy of each solar street lamp based on the lighting attribute information, and mark the solar street lamp whose current electric energy of the power source does not meet the target required electric energy as a waiting-for-receiving street lamp;

[0123] A module M30 for determining a street lamp to be powered is used to select other solar street lamps within a preset power supply threshold range among the street lamps to be powered as street lamps to be powered;

[0124] The power supply strategy generation module M40 is used to take the target power demand of the waiting street lamp, the current working condition data of the waiting street lamp and the current traffic data near the road as inputs of a preset power supply model to obtain a power supply strategy for supplying power to the waiting street lamp.

[0125] Based on the same inventive concept as the above embodiments, this embodiment also provides an intelligent control system, including multiple solar street lights, the power supplies of the multiple solar street lights are electrically connected in sequence along the extension direction of the road path, and also includes a processor and a memory; wherein the memory stores a computer program, and the computer program is used to be loaded by the processor and execute the road lighting control method based on solar street lights as mentioned above.

[0126] Based on the same inventive concept as the above embodiments, this embodiment further provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, and the instructions are used for a processor to load and execute the above-mentioned road lighting control method based on solar street lamps.

[0127] In the embodiments of the mobile terminal and computer-readable storage medium provided in the present application, all technical features of the above-mentioned control method embodiments are included, and the expansion and explanation content of the specification are basically the same as those of the above-mentioned method embodiments, which will not be repeated here.

[0128] The embodiment of the present application further provides a computer program product, which includes a computer program code. When the computer program code runs on a computer, the computer executes the methods in the above various possible implementation modes.

[0129] An embodiment of the present application also provides a chip, including a memory and a processor, the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that a device equipped with the chip executes the methods in various possible implementation modes as described above.

[0130] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0131] In the present application, the same or similar terminology concepts, technical solutions and / or application scenario descriptions are generally described in detail only the first time they appear. When they appear again later, they are generally not repeated for the sake of brevity. When understanding the technical solutions and other contents of the present application, for the same or similar terminology concepts, technical solutions and / or application scenario descriptions that are not described in detail later, reference can be made to the previous related detailed descriptions.

[0132] In the present application, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0133] The various technical features of the technical solution of the present application can be arbitrarily combined. In order to make the description concise, not all possible combinations of the various technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present application.

[0134] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product is stored in a storage medium as above, including a number of instructions for a terminal device to execute the method of each embodiment of the present application. The above is only a preferred embodiment of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of this application, or directly or indirectly used in other related technical fields, is similarly included in the patent protection scope of the present application.

[0135] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0136] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the attached claims.

Claims

1. A road lighting control method based on solar street lamps, used for an intelligent control system of solar street lamps, the intelligent control system comprises a plurality of solar street lamps, the power sources of the plurality of solar street lamps are electrically connected in sequence along the extension direction of the road path, characterized in that: The method comprises: Determine the lighting attribute information of each solar street lamp within a preset time period according to the historical working condition data, current working condition data of each solar street lamp and the historical traffic data and current traffic data near the road; Determine the target power requirement of each solar street lamp based on the lighting attribute information, and mark the solar street lamp whose current power supply does not meet the target power requirement as a street lamp to be received; Selecting other solar street lamps within a preset power supply threshold range among the street lamps to be supplied as street lamps to be supplied; The target power demand of the waiting street lamp, the current working condition data of the waiting street lamp and the current traffic data near the road are used as inputs of a preset power supply model to obtain a power supply strategy for supplying power to the waiting street lamp.

2. The road lighting control method based on solar street lamps according to claim 1 is characterized in that: The historical working condition data includes the power status, lighting brightness status, charging and discharging status and weather status of the street lamp in a preset time period; the lighting attribute information of each solar street lamp in a preset time period is determined based on the historical working condition data, current working condition data and historical flow data and current flow data near the road, specifically including: Determine the time series of illumination variables of the solar street lamp based on the power state, lighting state, charge and discharge state and the historical flow data; the time series of the time series of illumination variables includes a time series of power change, a time series of lighting brightness change, a time series of charge and discharge change and a time series of flow change; Inputting the time series of the illumination variable into a pre-trained long short-term memory network to determine an illumination feature vector corresponding to the illumination variable; The illumination feature vector and each illumination variable in the historical working condition data and the historical traffic data are respectively used as topological nodes of the Bayesian network to establish a lighting prediction model of the Bayesian network structure; The current working condition data of each solar street lamp and the current traffic data near the road are used as inputs of the lighting prediction model, and the lighting attribute information of each solar street lamp within a preset time period is output.

3. The road lighting control method based on solar street lamps according to claim 2 is characterized in that: Inputting the time series of the illumination variables into a pre-trained long short-term memory network to determine the illumination feature vector corresponding to the illumination variable specifically includes: Dividing the time series into a first training set and a first test set according to a preset ratio; Training the initial long short-term memory network based on the first training sample set to obtain a trained long short-term memory network; The first test set is input into the long short-term memory network for training to obtain predicted values ​​of illumination variables, and based on the predicted values ​​of illumination variables and the long short-term memory network, an illumination feature vector corresponding to the illumination variable is obtained.

4. The road lighting control method based on solar street lamps according to claim 2 is characterized in that: The characteristic vector and each illumination variable in the historical working condition data and the historical traffic data are respectively used as the topological nodes of the Bayesian network to establish a lighting prediction model with a Bayesian network structure, specifically including: Perform edge addition operation on the preset initial Bayesian network according to the node order, and select the Bayesian network structure with the highest score value after the edge modification operation as the current Bayesian network structure based on the preset scoring function; The conditional probability of each node in the current Bayesian network structure is estimated by an expectation maximization algorithm; The edge adding operation is repeated on the current Bayesian network structure, and the current Bayesian network structure whose conditional probability satisfies a preset iteration termination condition is used as the final Bayesian network structure for constructing the lighting prediction model.

5. The road lighting control method based on solar street lamps according to claim 4 is characterized in that: The step of performing an edge adding operation on the preset initial Bayesian network in the order of nodes, and selecting the Bayesian network structure with the highest score value after the edge modification operation as the current Bayesian network structure based on a preset scoring function, specifically includes: Determine the node order between the illumination feature vector and each illumination variable according to a preset order rule; An empty network structure is used as an initial Bayesian network structure, and an edge adding operation of child nodes is performed on the initial Bayesian network structure according to the node order, so as to obtain a Bayesian network structure with different parent nodes added; The scores of the Bayesian network structures with different parent nodes added to the edges are evaluated based on a preset scoring function, and the parent node set of the Bayesian network structure with the highest scoring value is selected as the final parent node of the child node to establish the current Bayesian network structure.

6. The road lighting control method based on solar street lamps according to claim 1 is characterized in that: The preset power supply model is trained by the following method: Determine the lamps to be supplied and the lamps to be supplied in the training data set based on the historical working condition data, and obtain the target required electric energy of the lamps to be supplied; A time node at which the power change of the waiting-for-received lamp exceeds a threshold value under historical working condition data is defined as a state change node; Determine the state transition probability between two adjacent state change nodes based on the target required power of the waiting-for-received lamp; Determine the value function corresponding to the execution of the initialization power supply strategy for two adjacent states according to the state transition probability; The initialized power supply strategy is updated according to the value function and a new power supply strategy is generated, until the value function corresponding to the new power supply strategy meets the preset termination condition, and the corresponding new power supply strategy is used as the optimal power supply strategy.

7. The road lighting control method based on solar street lamps according to claim 6 is characterized in that: The expression for iterating the value function is: Among them, V π (s) is the value function at state s, V π (s^) is the value function at the state s^; a represents the execution action corresponding to the power supply strategy, and its numerical expression is a=(e 1 ,e 2 ,……,e n ), where e 1 is the amount of electricity supplied by the first waiting circuit lamp to the waiting circuit lamp, e 2 is the amount of electricity that the second waiting circuit lamp supplies to the waiting circuit lamp, e n is the amount of power supplied by the nth waiting-to-supply lamp to the waiting-to-receive lamp; π represents the power supply strategy corresponding to the switch from state s to state s^; π(a|s) represents the probability of taking action a in state s under power supply strategy π; P(s^|s, a) is the probability of transferring to state s^ after adopting power supply strategy a in state s; R(s,a,s^) is the reward function, R(s,a,s^)=α·ΔE+β·ΔL+σ·ΔS, where ΔE is the change in the remaining power of the waiting-to-supply lamp, ΔL is the ratio between the actual brightness of the waiting-to-receive lamp and the target brightness, ΔS is the number of power supplies, and α, β and σ are the corresponding weight coefficients respectively; γ is the discount factor, 0≤γ≤1.

8. A road lighting control device based on solar street lamps, characterized in that: include: A lighting attribute determination module, used to determine the lighting attribute information of each solar street lamp within a preset time period based on the historical working condition data, current working condition data of each solar street lamp and the historical traffic data and current traffic data near the road; A waiting-for-receiving street lamp determination module is used to determine the target required electric energy of each solar street lamp based on the lighting attribute information, and mark the solar street lamp whose current electric energy of the power source does not meet the target required electric energy as a waiting-for-receiving street lamp; A module for determining a street lamp to be powered, used for selecting other solar street lamps within a preset power supply threshold range among the street lamps to be powered as street lamps to be powered; The power supply strategy generation module is used to take the target power demand of the waiting street lamp, the current working condition data of the waiting street lamp and the current traffic data near the road as inputs of a preset power supply model to obtain a power supply strategy for powering the waiting street lamp.

9. An intelligent control system, comprising a plurality of solar street lamps, the power sources of the plurality of solar street lamps being electrically connected in sequence along the extension direction of a road path, characterized in that: It also includes a processor and a memory; wherein the memory stores a computer program, and the computer program is used for the processor to load and execute the road lighting control method based on solar street lamps as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and the instructions are used by the processor to load and execute the road lighting control method based on solar street lamps as described in any one of claims 1 to 7.

Images