Solar street lamp brightness adjusting method based on Bayesian network

Abstract

The invention provides a solar street lamp brightness adjusting method based on a Bayesian network. The method comprises the steps that solar street lamps are clustered according to coordinate data, historical illumination intensity data and historical electric quantity data of all the solar street lamps, and a plurality of solar street lamp power supply sets are obtained; taking the current environment illumination data, the current traffic data and the current electric quantity data as input, and determining the illumination brightness priority of each solar street lamp in the solar street lamp power supply group based on a preset Bayesian network; determining a power receiving solar street lamp and a power supply solar street lamp of each solar street lamp power supply group based on the illumination brightness priority; and adjusting the brightness of the powered solar street lamp to the first target brightness, and adjusting the brightness of the power supply solar street lamp to the second target brightness. The power supply solar street lamps in the same solar street lamp power supply set can supply power to the power receiving solar street lamps, electric energy waste caused by long-distance power supply can be avoided, the local brightness of a necessary area is accurately adjusted, and on-demand power supply is achieved.

Description

Technical Field

[0001] The present invention relates to the field of solar street lamps, and in particular to a solar street lamp brightness adjustment method based on a Bayesian network. Background Art

[0002] Solar street lights use solar panels to convert sunlight into electricity during the day and store it in batteries. At night, when street lighting is needed, the smart controller releases the energy from the batteries to power the lights.

[0003] The relevant technology provides an intelligent lighting system for solar street lights. By electrically connecting the power sources of multiple solar street lights in sequence, when a solar street light lacks power or is insufficiently supplied, it can be powered by other connected solar street lights to ensure that the solar street light has a good lighting effect in a specific time period.

[0004] Due to weather or other factors (for example, several consecutive days of cloudy weather), although adjacent solar street lights can power each other, the overall power storage level is not high, and there will be a certain amount of power loss when solar street lights power each other, making it difficult to maintain high brightness lighting for a long time. In particular, for solar street lights installed in dangerous sections of road (such as forks, intersections, and construction sites), when pedestrians or vehicles pass under the street lights, in order to ensure that the adjacent street lights have good lighting effects, the solar street lights with insufficient power need to rely on several solar street lights with relatively high power to supply power, further resulting in excessive power loss of the solar street lights. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide a solar street lamp brightness adjustment method based on Bayesian network, which solves the problem of excessive power loss of solar street lamps in intelligent lighting systems in the prior art.

[0006] In a first aspect, the present application provides a solar street lamp brightness adjustment method based on a Bayesian network, which is used in an intelligent control system for solar street lamps. The intelligent control system includes a plurality of solar street lamps electrically connected in sequence along a road path. The method includes:

[0007] Clustering the solar street lamps according to coordinate data, historical light intensity data, and historical power data of each solar street lamp to obtain multiple solar street lamp power supply groups;

[0008] Using the current ambient light data, current traffic data, and current power data as input, the preset Bayesian network is used to determine the brightness priority of each solar street lamp in the solar street lamp power supply group;

[0009] Determine the powered solar street lamps and powered solar street lamps of each solar street lamp power supply group based on the light brightness priority;

[0010] Control the power supply solar street lamps of each solar street lamp power supply group to power the powered solar street lamps, adjust the brightness of the powered solar street lamps to a first target brightness, and adjust the brightness of the power supply solar street lamps to a second target brightness; the first target brightness is greater than the second target brightness.

[0011] In one embodiment, clustering the solar street lamps according to the coordinate data, historical light intensity data, and historical power data of each solar street lamp to obtain multiple solar street lamp power supply groups specifically includes:

[0012] Randomly assigning an initial cluster label to each of the solar street lamps; the initial cluster label has multiple categories;

[0013] According to the coordinate data, historical light intensity data and historical power data of each solar street lamp, the category of the initial cluster label is updated based on the Dirichlet discriminant method to obtain a temporary cluster label;

[0014] The temporary clustering label that meets the preset termination update condition is used as the final clustering label, the final clustering label is used as the category of the clustering label of the corresponding solar street lamp, and the solar street lamps are grouped according to the category of the clustering label to obtain multiple solar street lamp power supply groups.

[0015] In one embodiment, the updating of the category of the initial cluster label based on the Dirichlet discriminant method and obtaining the temporary cluster label specifically includes:

[0016] Determine the probability that each solar street lamp belongs to each category of the initial cluster label according to a preset Dirichlet discriminant function, and use the initial cluster label with the largest probability as a temporary cluster label of the solar street lamp;

[0017] Based on the temporary clustering label, the parameter variables of the corresponding solar street lamp in the Dirichlet discriminant function are updated and a new temporary clustering label is generated.

[0018] In one embodiment, the expression of the Dirichlet discriminant function is:

[0019]

[0020] Among them, x i is a vector consisting of the coordinate data, historical light intensity data, and historical power data of the i-th solar street light; -iis the clustering label of all solar street lights except the i-th solar street light; α is the concentration parameter, ranging from 0.1 to 1; n k,-i is the number of solar street lamps in cluster k after excluding the i-th solar street lamp; K is the number of categories of the initial cluster label; n j,-i is the number of solar street lamps in cluster j after excluding the i-th solar street lamp; The mean is μ k , covariance is ∑ k Multivariate Gaussian distribution with mean μ k and covariance∑ k is the parameter variable of the Dirichlet discriminant function to be updated; G0(x i ) is the vector x i The base distribution of .

[0021] In one embodiment, the current ambient light data, current traffic data, and current power data are used as inputs, and the light intensity priority of each solar street lamp in the solar street lamp power supply group is determined based on a preset Bayesian network, specifically including:

[0022] Constructing a joint tree corresponding to the Bayesian network; the joint tree includes a plurality of connected clique nodes;

[0023] Determining a potential function of each clique node according to a conditional probability table of a Bayesian network;

[0024] Distributing the current ambient light data, current traffic data, and current power data as evidence information in the joint tree, and transmitting the evidence information according to the connection relationship between each cluster node;

[0025] The potential function of each clique node is updated according to the evidence information, and the light brightness priority of each solar street lamp in its corresponding solar street lamp power supply group is determined based on the updated potential function.

[0026] In one embodiment, constructing the joint tree corresponding to the Bayesian network specifically includes:

[0027] Selecting a node with multiple parent nodes in the Bayesian network as a target node, connecting the parent nodes of the target node in pairs and adding undirected edges;

[0028] Convert all directed edges of the Bayesian network into undirected edges, and obtain a moral graph consisting of all undirected edges;

[0029] Selecting a ring in the moral graph whose length is greater than a preset length threshold as a target ring, and adding a chord to the target ring to triangulate the moral graph; the chord is used to connect any two non-adjacent nodes in the target ring with an edge;

[0030] From the moral graph after triangulation, a clique that cannot add any more nodes and remains fully connected is determined as a maximal clique, and a joint tree is constructed using the maximal clique as a clique node.

[0031] In one embodiment, before determining the light intensity priority of each solar street lamp in the solar street lamp power supply group based on the preset Bayesian network, the method further includes:

[0032] Constructing a basic Bayesian network based on historical ambient light data, historical traffic data, and historical power data; the nodes of the basic Bayesian network include ambient light nodes, traffic nodes, and power nodes;

[0033] A distance loss probability table is constructed based on the coordinate data, the distance loss probability table is associated with the power node, and the basic Bayesian network structure is updated to obtain a preset Bayesian network structure.

[0034] In one embodiment, determining the powered solar street lamps and the powered solar street lamps of each solar street lamp power supply group based on the light brightness priority specifically includes:

[0035] Sorting the solar street lamps of the solar street lamp power supply group according to the priority of light brightness;

[0036] When the brightness priority ranking of the solar street lamp is within a preset ranking range and the current power of the solar street lamp is lower than a preset power threshold, the solar street lamp is selected as the powered solar street lamp, and other solar street lamps in the solar street lamp power supply group are selected as the power supply solar street lamps.

[0037] In one embodiment, before adjusting the brightness of the powered solar street lamp to the first target brightness and adjusting the brightness of the power supply solar street lamp to the second target brightness, the method further includes:

[0038] Determining the first target brightness according to the current power of the powered solar street lamp;

[0039] The second target brightness is determined according to the current power of the powered solar street lamp.

[0040] In a second aspect, the present application provides an intelligent control system comprising a processor and a memory; wherein the memory stores a computer program, and the computer program is used by the processor to load and execute the solar street lamp brightness adjustment method based on the Bayesian network as described in any one of the first aspects.

[0041] In this embodiment of the Bayesian network-based solar street light brightness adjustment method, power-supplying solar street lights within the same solar street light power supply group can power receiving solar street lights, thus avoiding energy waste caused by long-distance power supply. Furthermore, the receiving and power-supplying solar street lights within the same solar street light power supply group can flexibly adjust their brightness based on current ambient light data, traffic data, and power consumption data. When pedestrians or vehicles pass by, the brightness of necessary areas is precisely adjusted to brighten, while non-essential areas remain dimmed. This achieves on-demand power supply, resulting in excellent energy-saving and emission-reduction effects for the solar street lights. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0043] Figure 1 Schematic diagram of the process of adjusting the brightness of a solar street lamp in an embodiment of the present application.

[0044] Figure 2 Schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0045] Specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the described embodiments are merely some, and not all, of the embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the description of the present invention without inventive effort are intended to fall within the scope of protection of the present invention.

[0046] In the description of the present invention, unless otherwise specified or limited, the terms "disposed," "installed," and "connected" should be understood broadly. For example, they may refer to fixed, detachable, or integral connections; mechanical or electrical connections; and direct or indirect connections through an intermediary. Those skilled in the art will understand the specific meanings of these terms based on the specific circumstances.

[0047] 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 in use. They are only for the convenience and simplification of description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operate in a specific direction. Therefore, they should not be understood as limiting the present invention.

[0048] The terms "first," "second," "third," etc. are merely used to distinguish elements of similar nature and do not indicate or imply relative importance or a particular order.

[0049] The terms "comprises," "comprising," or any other variations thereof, are intended to cover a non-exclusive inclusion of elements other than the listed elements and may also include additional elements not specifically listed.

[0050] The Bayesian network-based solar street light brightness adjustment method of this embodiment is applied to the intelligent control system of solar street lights. The intelligent control system includes multiple solar street lights, and the power supplies of the multiple solar street lights are electrically connected in sequence along the extension direction of the road path. When the solar power supply of a street light is insufficient due to weather or power failure, it can be powered by the support of other street light power supplies to meet the lighting brightness and duration of the street light.

[0051] like Figure 1 As shown, this embodiment provides a solar street light brightness adjustment method based on a Bayesian network, which is used in an intelligent control system for solar street lights. The intelligent control system includes a plurality of solar street lights electrically connected in sequence along a road path. The method includes:

[0052] Step S10: clustering the solar street lamps according to the coordinate data, historical light intensity data, and historical power data of each solar street lamp to obtain a plurality of solar street lamp power supply groups;

[0053] Step S20: using the current ambient light data, current traffic data, and current power data as input, determining the light intensity priority of each solar street lamp in the solar street lamp power supply group based on a preset Bayesian network;

[0054] Step S30: determining the powered solar street lamps and the powered solar street lamps of each solar street lamp power supply group based on the light brightness priority;

[0055] Step S40: Control the power supply solar street lights of each solar street light power supply group to power the powered solar street lights, adjust the brightness of the powered solar street lights to a first target brightness, and adjust the brightness of the power supply solar street lights to a second target brightness; the first target brightness is greater than the second target brightness.

[0056] In this embodiment of the Bayesian network-based solar street light brightness adjustment method, power-supplying solar street lights within the same solar street light power supply group can power receiving solar street lights, thus avoiding energy waste caused by long-distance power supply. Furthermore, the receiving and power-supplying solar street lights within the same solar street light power supply group can flexibly adjust their brightness based on current ambient light data, traffic data, and power consumption data. When pedestrians or vehicles pass by, the brightness of necessary areas is precisely adjusted to brighten, while non-essential areas remain dimmed. This achieves on-demand power supply, resulting in excellent energy-saving and emission-reduction effects for the solar street lights.

[0057] Step S10: clustering the solar street lamps according to the coordinate data, historical light intensity data and historical power data of each solar street lamp to obtain a plurality of solar street lamp power supply groups.

[0058] The coordinate data of a solar street light specifically refers to the longitude and latitude coordinates of the solar street light. Each solar street light has a unique code number when it is installed. The geographical coordinates of each solar street light are bound to its code number on the remote server, and the coordinate data of each solar street light can be determined by querying the code number.

[0059] Solar street lights use LED panels to generate light, and the light intensity of the LED panels can be adjusted by adjusting the forward current of the LED panels. The LED panels are connected in series with a current sensing resistor. The voltage across the current sensing resistor is input to the inverting terminal of the op amp, and the non-inverting terminal is connected to a reference voltage. The op amp output controls the gate of the switching tube, forming a closed-loop constant current control. By changing the reference voltage or controlling the on-time of the switching tube through a PWM signal, the average current through the LED is adjusted, thereby changing the light intensity of the LED panel. Therefore, the light intensity of the LED panel can be determined by measuring the current passing through the current sensing resistor. A mapping relationship between the light intensity and time is established to obtain light intensity data for each time period. The light intensity data can then be stored on a local storage device or uploaded to a cloud server. Historical light intensity data for historical time periods can be queried by time.

[0060] Each solar street light's power supply is equipped with a power monitoring module that monitors the power status in real time. It records power data at regular intervals and stores it on the street light's local storage device or uploads it to a cloud server. Historical power data can be retrieved and exported from the storage device or cloud server, and can be filtered and downloaded based on criteria such as street light number and time range.

[0061] Because the power consumption of each solar street light varies flexibly, and because the time spent by traffic participants such as pedestrians and vehicles under each solar street light is random, the area required to maintain high brightness also varies randomly. This embodiment dynamically groups solar street lights based on their spatiotemporal and power characteristics by establishing a Dirichlet mixture probability model.

[0062] Clustering the solar street lamps according to the coordinate data, historical light intensity data, and historical power data of each solar street lamp to obtain multiple solar street lamp power supply groups specifically includes:

[0063] Step S101: randomly assigning an initial cluster label to each of the solar street lamps; the initial cluster label has multiple categories;

[0064] Step S102: updating the category of the initial cluster label based on the Dirichlet discriminant method according to the coordinate data, historical light intensity data, and historical power data of each solar street lamp and obtaining a temporary cluster label;

[0065] Step S103: taking the temporary clustering label that meets the preset termination update condition as the final clustering label, taking the final clustering label as the category of the corresponding solar street lamp clustering label, and grouping the solar street lamps according to the category of the clustering label to obtain multiple solar street lamp power supply groups.

[0066] In step S101, three clusters are assigned based on power demand: low power demand, medium power demand, and high power demand. The total number of clusters, k, is 3. For each solar street light, a random number generator generates an integer between 1 and K, which serves as the initial cluster label for the street light. For example, assuming there are five solar street lights, L1, L2, L3, L4, and L5, the randomly generated initial cluster labels might be L1 for category 2, L2 for category 1, L3 for category 3, L4 for category 2, and L5 for category 1.

[0067] In step S102, the flexibly changing power consumption of each solar streetlight and the random dwell times of pedestrians or vehicles cause the area required to maintain high brightness to vary randomly. The Dirichlet discriminant does not simply perform a deterministic classification on each streetlight. Instead, it assigns a probability of belonging to each cluster category, achieving relatively stable and accurate clustering within a relatively small number of iterations, saving computational resources and time. For example, at a given moment, based on current data, the probability that a streetlight belongs to the "high power demand group" is 0.6, and the probability that it belongs to the "medium power demand group" is 0.4. This probabilistic representation can better reflect the actual situation of streetlights in a dynamic environment. When environmental factors change, the probabilities are adjusted accordingly, allowing the clustering results to adapt to changes in a timely manner.

[0068] Furthermore, the power consumption of solar streetlights is influenced by a complex interaction of factors, including light intensity, geographic location, and traffic flow. The Dirichlet discriminant can uncover these underlying complex patterns, discovering hidden laws in the data. It can also capture unique patterns in power consumption under certain specific light intensities and geographic locations, and use these patterns to rationally cluster solar streetlights.

[0069] In step S102, the updating of the category of the initial cluster label based on the Dirichlet discriminant method and obtaining a temporary cluster label specifically includes:

[0070] Step S1021: determining the probability that each solar street light belongs to each type of the initial cluster label according to a preset Dirichlet discriminant function, and using the initial cluster label with the highest probability as a temporary cluster label for the solar street light;

[0071] Step S1022: Based on the temporary clustering label, the parameter variables of the corresponding solar street lamp in the Dirichlet discriminant function are updated and a new temporary clustering label is generated.

[0072] Among them, the expression of Dirichlet discriminant function is:

[0073]

[0074] Among them, x i is a vector consisting of the coordinate data, historical light intensity data, and historical power data of the i-th solar street light; -i is the clustering label of all solar street lights except the i-th solar street light; α is the concentration parameter, ranging from 0.1 to 1; n k,-i is the number of solar street lamps in cluster k after excluding the i-th solar street lamp; K is the number of categories of the initial cluster label; n j,-i is the number of solar street lamps in cluster j after excluding the i-th solar street lamp; The mean is μk , covariance is ∑ k Multivariate Gaussian distribution with mean μ k and covariance∑ k is the parameter variable of the Dirichlet discriminant function to be updated; G0(x i ) is the vector x i The base distribution of .

[0075] For example, if the probability that solar street lamp L1 belongs to category 1 is P(1|X L1 ), it is necessary to estimate the coordinates of solar street light L1, light intensity and power characteristic vector X in category 1 based on historical data L1 The likelihood probability is P(X L1 ∣1). If category 1 represents solar street lamps with sufficient light, stable power consumption and relatively concentrated geographical locations, and solar street lamp L1 has a high average light intensity, a stable power consumption change rate, and its coordinates are close to the center of the category representative area, then P(X L1 |1) is relatively large, category 1 is used as the temporary clustering label of solar street lamp L1.

[0076] When updating the parameter variables of the Dirichlet discriminant function, the main update is the mean μ of the Dirichlet discriminant function k and covariance∑ k Mean μ k The update formula is: Covariance Σ k The update formula is: Among them, n k is the number of solar street lights in cluster k, x i is the data vector of street lamp i belonging to cluster k.

[0077] When updating the mean μ k and covariance Σ k After that, the probability value of the temporary cluster label obtained by the Dirichlet discriminant function will also change.

[0078] In step S103, when the update iteration number reaches the maximum iteration number, or when the ratio of changes in all solar street lamp cluster labels in two adjacent iterations is less than a certain threshold (such as 0.01, i.e. 1%), the clustering result is considered to have converged and the update is stopped.

[0079] When the preset termination conditions are met, the temporary cluster labels obtained at this time become the final cluster labels. The solar street lights are then grouped according to the categories of the final cluster labels. For example, the final cluster labels are divided into three categories: all solar street lights labeled "category 1" are grouped into one group, those labeled "category 2" into another group, and those labeled "category 3" into a third group. This results in multiple solar street light power supply groups.

[0080] Considering the uncertainty surrounding the power status, light availability, and actual power demand of each solar street light, even within the same solar street light power supply group, the light intensity of each solar street light varies due to its specific location, resulting in different power availability. Furthermore, the random duration of time pedestrians or vehicles spend under different street lights also makes the actual high-brightness demand of the solar street light uncertain, leading to fluctuations in power consumption. Therefore, this embodiment uses a Bayesian network to determine the causal relationship and probabilistic dependency between each solar street light. This allows for a more efficient way to determine power supply relationships and predict future power changes, thereby achieving more accurate and efficient power supply configuration, reducing unnecessary power transmission losses, and improving the energy efficiency of the entire solar street light power supply group.

[0081] Step S20: using the current ambient light data, the current traffic data and the current power data as input, the brightness priority of each solar street lamp in the solar street lamp power supply group is determined based on a preset Bayesian network.

[0082] Each solar street light is equipped with a light intensity sensor, which can sense the ambient light conditions in real time and measure and record the light intensity value at a set time interval (such as every minute).

[0083] Each solar street light is equipped with a camera with AI recognition capabilities. The camera can collect real-time traffic data (pedestrians, vehicles, or other traffic participants) of the solar street light and store the collected traffic data in a local storage device or upload it to a remote server. It is understood that current traffic data can also be obtained through radar sensors or geomagnetic induction coils.

[0084] The power monitoring module of each solar street light can be directly called to obtain its corresponding current power data.

[0085] The default Bayesian network contains four nodes: ambient light (E), traffic data (T), power (P), and light intensity (L). Ambient light and traffic flow both affect the light intensity priority, and power also affects the light intensity priority. Therefore, the structure of the Bayesian network can be simplified as: E->L<-TP.

[0086] Since a Bayesian network is a probability-based directed acyclic graph model, its nodes represent random variables, and its edges represent conditional dependencies between variables. These dependencies are quantified using a conditional probability table. Therefore, it is necessary to discretize the current ambient light data, traffic data, and power data so that each variable has a finite number of distinct states or values ​​that can directly correspond to each item in the conditional probability table.

[0087] During discretization processing, the current ambient light data is discretized into two levels: low ambient light (less than or equal to 50 Lux) and high ambient light (greater than 50 Lux); the current traffic data is discretized into three levels: low flow (less than 2 traffic participants), medium flow (2 to 5 traffic participants) and high flow (greater than 5 traffic participants); the current power data is divided into three levels: low power (less than or equal to 30% of the total power), medium power (30%-70% of the total power) and high power (greater than 70% of the total power).

[0088] Because the current ambient light data, traffic data, and power data for solar streetlights come from different sources, the Bayesian network constructed using the joint tree algorithm in this embodiment can clearly demonstrate how changes in ambient light affect streetlight brightness requirements, and how traffic flow and power status causally influence light intensity priority decisions. Furthermore, the joint tree algorithm transforms the Bayesian network into a tree structure and utilizes a message-passing mechanism for inference calculations, completing calculations within a reasonable timeframe. Using real-time ambient light, traffic, and power data, the algorithm quickly and accurately determines the brightness priority of each streetlight, enabling timely brightness adjustment and control.

[0089] The method uses the current ambient light data, the current traffic data, and the current power data as inputs and determines the light intensity priority of each solar street lamp in the solar street lamp power supply group based on a preset Bayesian network, specifically including:

[0090] Step S201: constructing a joint tree corresponding to the Bayesian network; the joint tree includes a plurality of connected clique nodes;

[0091] Step S202: determining the potential function of each clique node according to the conditional probability table of the Bayesian network;

[0092] Step S203: distributing the current ambient light data, current traffic data, and current power data as evidence information in the joint tree, and transmitting the evidence information according to the connection relationship between each cluster node;

[0093] Step S204: updating the potential function of each clique node according to the evidence information, and determining the light brightness priority of each solar street lamp in its corresponding solar street lamp power supply group based on the updated potential function.

[0094] In step S201 , the joint tree is an undirected tree structure that converts the complex dependency relationship of the Bayesian network into a hierarchical structure of clusters and efficiently calculates the probability through a message passing algorithm.

[0095] The constructing of the joint tree corresponding to the Bayesian network specifically includes:

[0096] Step S2011: selecting a node with multiple parent nodes in the Bayesian network as a target node, connecting the parent nodes of the target node in pairs and adding undirected edges;

[0097] Step S2012: converting all directed edges of the Bayesian network into undirected edges, and obtaining a moral graph consisting entirely of the undirected edges;

[0098] Step S2013: selecting a ring in the moral graph whose length is greater than a preset length threshold as a target ring, and adding a chord to the target ring to triangulate the moral graph; the chord is used to connect any two non-adjacent nodes in the target ring with an edge;

[0099] Step S2014: Determine from the triangulated moral graph a clique that cannot add any more nodes and remains fully connected as a maximal clique, use the maximal clique as a clique node and construct a joint tree.

[0100] For example, the light intensity (L) node has three parent nodes: ambient light (E), traffic data (T), and power (P). Therefore, by adding undirected edges between E and T, E and P, and T and P, and converting the directed edges between L and E, T, and P to undirected edges, we can obtain an undirected moral graph. When triangulating the moral graph, we first search for a target ring (e.g., ELTPE) with a length greater than 3. Then, we add the chord ET to the target ring, splitting it into smaller subgraphs. This ensures that there are no unchorded rings greater than 3 in the graph, completing the triangulation of the moral graph. Then, we obtain the clique nodes {E, L}, {T, L}, and {P, L} from the moral graph. Since these cliques share the element L, the interface connects these clique nodes based on the element L to obtain the corresponding union tree.

[0101] In steps S202 and S203, for each clique node, its potential function is calculated according to the conditional probability table in the Bayesian network. For example, for the clique {E, L}, assuming P(E = low ambient light) = 0.3, P(E = high ambient light) = 0.7, and P(L = high priority | E = low ambient light) = 0.1, P(L = high priority | E = high ambient light) = 0.9, then its potential function φ E,L(E = low ambient light, L = high priority) = P(L = high priority | E = low ambient light) x P(E = low ambient light) = 0.1 x 0.3 = 0.03. The potential function values ​​for other state combinations can be calculated in this way.

[0102] Distributing evidence information is divided into two stages: collecting evidence and distributing evidence:

[0103] In the evidence collection phase, each node calculates and sends a message to its neighboring nodes. The message content is the product of the node's potential function and the messages received from other neighboring nodes, and then sums all variables except the variable receiving the message. For example, node {E, L} has two neighboring nodes {T, L} and {P, L}. In the evidence collection phase, node {E, L} first receives messages m from {T, L} and {P, L}. T,L→E,L and m P,L→E,L Assume that the potential function of {E,L} is φ E,L (E,L), then the message m transmitted by node {E,L} to its parent node E,L→parent (L) is: m E,L→parent (L)=∑ E φ E,L (E,L)×m T,L→E,L (L)×m P,L→E,L (L), the message is carried along the joint tree from the leaf node to the root node in sequence, and each node calculates and transmits a new message based on the received message and its own potential function.

[0104] In the evidence distribution phase, each node calculates and sends messages to its neighboring nodes. The calculation method is similar to the evidence collection phase. For example, after the root node R receives the messages from all child nodes, it sends the message m to its child node C1. R→C1 :m R→C1 (variablesC1)=∑ variablesR-variablesC1 φR(variables R )×∏ child≠C1 m child→R (variables R ), where is the set of variables contained in the root node R, variables C1 It is the set of variables contained in child node C1.

[0105] In step S204, after the information transfer of the distribution evidence information, the potential function of each clique node contains the information of the light brightness priority L, and its probability distribution can be calculated in the clique node containing L. For example, in the clique node {E, L}, the probability distribution P(L) of L is calculated as follows:

[0106]

[0107] It is understandable that the probability distribution of other clique nodes can be determined similarly according to the above logic.

[0108] Before determining the light intensity priority of each solar street lamp in the solar street lamp power supply group based on the preset Bayesian network, the method further includes:

[0109] Step S205: constructing a basic Bayesian network based on historical ambient light data, historical traffic data, and historical power data; the nodes of the basic Bayesian network include ambient light nodes, traffic nodes, and power nodes;

[0110] Step S206: constructing a distance loss probability table based on the coordinate data, associating the distance loss probability table with the power node, and updating the basic Bayesian network structure to obtain a preset Bayesian network structure.

[0111] In step S205, the historical ambient light data, historical traffic data, and historical power data are first discretized. Then, industry expertise is combined to determine the dependencies between these nodes and construct a basic Bayesian network. For example, ambient light intensity can affect traffic flow (traffic flow at night, when light levels are low, is generally lower than during daytime, when light levels are high), so the ambient light node has an impact on the traffic node. Furthermore, ambient light intensity directly affects the charging status of solar streetlights, which in turn affects the power node. This means that the ambient light node also has an impact on the power node.

[0112] Statistical methods can be used to learn the conditional probability table for each node. For example, for a traffic node, its conditional probability table describes the probability of traffic flow being in different states under different ambient light intensities. Assume that the ambient light node has two states: "low ambient light" and "high ambient light", and the traffic node has three states: "low flow", "medium flow", and "high flow". By counting the number of times each traffic flow state occurs under different lighting conditions, calculating the corresponding probability, and filling it into the conditional probability table of the traffic node, the conditional probability table of the traffic node can be obtained. The same method can also be used to determine the conditional probability table of the power node based on the ambient light node or the conditional probability table of other nodes.

[0113] In step S206, a distance loss probability table is introduced into the Bayesian network to accurately assess power loss during transmission and provide a probabilistic basis for calculating power transmission loss. For example, when the distance between solar street lights is within 5 meters, the loss probability is set to 5% per meter; when the distance between solar street lights is within 5-10 meters, the loss probability is set to 10% per meter; and when the distance between solar street lights is greater than 10 meters, the loss probability is set to 15% per meter.

[0114] In a Bayesian network, when it comes to updating the power status of street lights, the impact of power transmission loss is taken into account. When calculating the probability distribution of the current power node of a solar street light, the connection distance information between the solar street light and other related solar street lights is combined, and the loss probability during the power transmission process is calculated using a distance loss probability table, thereby more accurately evaluating the actual remaining power of the street light.

[0115] Assume that there are two adjacent solar street lights A and B, and solar street light A transmits electricity to solar street light B. In the Bayesian network, the power state of solar street light A (E A ) is a parent node, the power status of solar street light B (E B ) is a child node. When the current power of solar street light A and the distance between solar street lights A and B are known based on sensor data, the corresponding loss probability P is obtained through the distance loss probability table. loss , the amount of electricity E after solar street light B receives electricity B The amount of electricity before transmission to solar street light A is E A And the loss probability P loss Related, then E B =E A ×(1-P loss ).

[0116] In this way, the Bayesian network can more accurately reflect the power status of street lamps during actual operation, providing a more accurate basis for street lamp power supply decisions. When deciding whether to adjust the brightness of a solar street lamp or switch the power supply mode, considering the power information after transmission loss can make the decision more reasonable and avoid insufficient or excessive power supply caused by ignoring transmission loss.

[0117] To reflect the impact of distance on the power node, a directed edge is added to the basic Bayesian network structure from a virtual node representing distance (which can be named "distance node") to the power node. This new directed edge represents the direct impact of distance on the power status. At the same time, the conditional probability table for the power node is reviewed and adjusted to comprehensively consider the combined effects of ambient light and distance on the power status. The basic Bayesian network structure is then updated to the pre-set Bayesian network structure, enabling it to more accurately reflect the impact of various factors on the power status of solar street lights in real-world scenarios.

[0118] Step S30: Determine the powered solar street lamps and the powered solar street lamps of each solar street lamp power supply group based on the light brightness priority.

[0119] Usually, the solar street light with the highest priority in terms of brightness is selected as the receiving solar street light, while the other solar street lights are used as the power supply solar street lights. However, if the solar street light with the highest priority in terms of brightness has sufficient power, and there are other solar street lights with relatively low power in the solar street light power supply group, the solar street light with low power can be selected as the receiving solar street light.

[0120] The method of determining the powered solar street lights and the power supplying solar street lights of each solar street light power supply group based on the light brightness priority specifically includes: sorting the solar street lights of the solar street light power supply group according to the light brightness priority; when the brightness priority sorting of the solar street lights is within a preset sorting range and the current power of the solar street lights is lower than a preset power threshold, selecting the solar street lights as the powered solar street lights, and selecting other solar street lights of the solar street light power supply group as the power supplying solar street lights.

[0121] After determining the receiving and supplying solar street lights, the system doesn't simply transfer power from the supplying solar street light to the receiving one. Instead, it develops a refined power supply strategy based on the actual power demand of the receiving solar street light and the remaining power of the supplying solar street light. For example, if the receiving solar street light's current power level is 20%, significantly below its preset full power level of 100%, while some of the supplying solar street lights are also running low (e.g., with 50% remaining), power can be allocated to the receiving solar street light from the supplying solar street light with relatively sufficient power (e.g., with more than 70% remaining). Furthermore, to ensure the stability and reliability of the power supply process, the amount of power supplied at each time should be limited (no more than 30% of the remaining power of the supplying solar street light) to prevent excessive power from affecting the normal operation of the supplying solar street light.

[0122] Step S40: Control the power supply solar street lamps of each solar street lamp power supply group to power the powered solar street lamps, adjust the brightness of the powered solar street lamps to a first target brightness, and adjust the brightness of the power supply solar street lamps to a second target brightness; the first target brightness is greater than the said.

[0123] The value of the first target brightness is usually significantly higher than the second target brightness, and in order to have a good display effect, the brightness of the power supplying solar street lamp gradually decreases as the distance between the power supplying solar street lamp and the powered solar street lamp becomes farther.

[0124] Before adjusting the brightness of the powered solar street lamp to the first target brightness and adjusting the brightness of the power supplying solar street lamp to the second target brightness, it also includes: determining the first target brightness according to the current power of the powered solar street lamp; determining the second target brightness according to the current power of the power supplying solar street lamp.

[0125] To prevent excessive power loss in the power-supplying solar street light, the current power level of the receiving solar street light is divided into multiple levels before power is transmitted to the receiving solar street light. Different maximum brightness settings are set for each level. For example, when the power level is above 80%, the solar street light must meet a higher brightness to ensure good lighting effect. In this case, the first target brightness can be set to 90% of the maximum brightness. When the power level is between 50% and 80%, the maximum brightness is set to 70%. When the power level is between 30% and 50%, the maximum brightness is set to 50% to account for power loss during the power supply process and subsequent lighting needs. When the power level is below 30%, the maximum brightness is set to 30% to ensure that the receiving solar street light can maintain basic lighting after receiving power.

[0126] When supplying power to a solar street light, the power supply needs to determine the second target brightness based on its own power and reserved power requirements, and ensure that it can still maintain its basic operation after power supply within the power safety threshold (such as 30%). For example, when the power supply of the solar street light is higher than 70%, the brightness can be appropriately reduced to 60% of the maximum brightness to provide more power to the receiving solar street light. When the power is between 50% and 70%, the brightness is set to 70% of the maximum brightness, which not only ensures a certain level of lighting but also provides a moderate amount of power. When the power is between 30% and 50%, the brightness is maintained at 80% of the maximum brightness, and the power supply is reduced to maintain its own power level. When the power is below 30%, it will no longer supply power to other street lights, and maintain its own maximum brightness to meet the lighting needs of the location.

[0127] Based on the same inventive concept as the above embodiment, this embodiment also provides an intelligent control system, which 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 above-mentioned Bayesian network-based solar street lamp brightness adjustment method.

[0128] like Figure 2 As shown, based on the same inventive concept as the above embodiment, this embodiment also provides a computer-readable storage medium, which stores instructions for the processor to load and execute the above-mentioned solar street light brightness adjustment method based on the Bayesian network.

[0129] In the embodiments of the mobile terminal and computer-readable storage medium provided in this application, all technical features of the above-mentioned control method embodiments are included. The expanded and explained contents of the specification are basically the same as those of the above-mentioned method embodiments and will not be repeated here.

[0130] An embodiment of the present application further provides a computer program product, which includes computer program code. When the computer program code runs on a computer, the computer executes the methods in the various possible implementation modes described above.

[0131] An embodiment of the present application also provides a chip, including a memory and a processor, wherein the memory is used to store computer programs, and the processor is used to call and run the computer programs from the memory, so that a device equipped with the chip executes the methods in the various possible implementation modes as described above.

[0132] 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.

[0133] In this application, the same or similar terminology, technical solutions and / or application scenario descriptions are generally only described in detail 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 this application, for the same or similar terminology, technical solutions and / or application scenario descriptions that are not described in detail later, you can refer to the previous relevant detailed descriptions.

[0134] In this application, the description of each embodiment has its own focus. For parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0135] The various technical features of the technical solution of this application can be combined arbitrarily. 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 this application.

[0136] 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 the 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 enabling 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 using the contents of the description 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.

[0137] 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 and similar parts between the various embodiments can be referenced to each other.

[0138] The foregoing description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be readily conceived by a person skilled in the art within the technical scope disclosed herein are intended to be encompassed within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A solar street light brightness adjustment method based on a Bayesian network, used in an intelligent control system for solar street lights, wherein the intelligent control system comprises a plurality of solar street lights electrically connected in sequence along a road path, characterized in that: The method comprises: Clustering the solar street lamps according to coordinate data, historical light intensity data, and historical power data of each solar street lamp to obtain multiple solar street lamp power supply groups; Using the current ambient light data, current traffic data, and current power data as input, the preset Bayesian network is used to determine the brightness priority of each solar street lamp in the solar street lamp power supply group; Determine the powered solar street lamps and powered solar street lamps of each solar street lamp power supply group based on the light brightness priority; Control the power supply solar street lamps of each solar street lamp power supply group to power the powered solar street lamps, adjust the brightness of the powered solar street lamps to a first target brightness, and adjust the brightness of the power supply solar street lamps to a second target brightness; the first target brightness is greater than the second target brightness.

2. The solar street light brightness adjustment method based on Bayesian network according to claim 1 is characterized in that: Clustering the solar street lamps according to the coordinate data, historical light intensity data, and historical power data of each solar street lamp to obtain multiple solar street lamp power supply groups specifically includes: Randomly assigning an initial cluster label to each of the solar street lamps; the initial cluster label has multiple categories; According to the coordinate data, historical light intensity data and historical power data of each solar street lamp, the category of the initial cluster label is updated based on the Dirichlet discriminant method to obtain a temporary cluster label; The temporary clustering label that meets the preset termination update condition is used as the final clustering label, the final clustering label is used as the category of the clustering label of the corresponding solar street lamp, and the solar street lamps are grouped according to the category of the clustering label to obtain multiple solar street lamp power supply groups.

3. The solar street light brightness adjustment method based on Bayesian network according to claim 2 is characterized in that: The updating of the category of the initial cluster label based on the Dirichlet discriminant method and obtaining the temporary cluster label specifically includes: Determine the probability that each solar street lamp belongs to each category of the initial cluster label according to a preset Dirichlet discriminant function, and use the initial cluster label with the largest probability as a temporary cluster label of the solar street lamp; Based on the temporary clustering label, the parameter variables of the corresponding solar street lamp in the Dirichlet discriminant function are updated and a new temporary clustering label is generated.

4. The solar street light brightness adjustment method based on Bayesian network according to claim 3 is characterized in that: The expression of the Dirichlet discriminant function is: Among them, x i is a vector consisting of the coordinate data, historical light intensity data, and historical power data of the i-th solar street light; -i is the clustering label of all solar street lights except the i-th solar street light; α is the concentration parameter, ranging from 0.1 to 1; n k,-i is the number of solar street lamps in cluster k after excluding the i-th solar street lamp; K is the number of categories of the initial cluster label; n j,-i is the number of solar street lamps in cluster j after excluding the i-th solar street lamp; The mean is μ k , covariance is ∑ k Multivariate Gaussian distribution with mean μ k and covariance∑ k is the parameter variable of the Dirichlet discriminant function to be updated; G0(x i ) is the vector x i The base distribution of .

5. The solar street light brightness adjustment method based on Bayesian network according to claim 1, characterized in that: The method uses the current ambient light data, the current traffic data, and the current power data as inputs and determines the light intensity priority of each solar street lamp in the solar street lamp power supply group based on a preset Bayesian network, specifically including: Constructing a joint tree corresponding to the Bayesian network; the joint tree includes a plurality of connected clique nodes; Determining a potential function of each clique node according to a conditional probability table of a Bayesian network; Distributing the current ambient light data, current traffic data, and current power data as evidence information in the joint tree, and transmitting the evidence information according to the connection relationship between each cluster node; The potential function of each clique node is updated according to the evidence information, and the light brightness priority of each solar street lamp in its corresponding solar street lamp power supply group is determined based on the updated potential function.

6. The solar street light brightness adjustment method based on Bayesian network according to claim 5, characterized in that: The constructing of the joint tree corresponding to the Bayesian network specifically includes: Selecting a node with multiple parent nodes in the Bayesian network as a target node, connecting the parent nodes of the target node in pairs and adding undirected edges; Convert all directed edges of the Bayesian network into undirected edges, and obtain a moral graph consisting of all undirected edges; Selecting a ring in the moral graph whose length is greater than a preset length threshold as a target ring, and adding a chord to the target ring to triangulate the moral graph; the chord is used to connect any two non-adjacent nodes in the target ring with an edge; From the moral graph after triangulation, a clique that cannot add any more nodes and remains fully connected is determined as a maximal clique, and a joint tree is constructed using the maximal clique as a clique node.

7. The solar street light brightness adjustment method based on Bayesian network according to claim 1, characterized in that: Before determining the light intensity priority of each solar street lamp in the solar street lamp power supply group based on the preset Bayesian network, the method further includes: Constructing a basic Bayesian network based on historical ambient light data, historical traffic data, and historical power data; the nodes of the basic Bayesian network include ambient light nodes, traffic nodes, and power nodes; A distance loss probability table is constructed based on the coordinate data, the distance loss probability table is associated with the power node, and the basic Bayesian network structure is updated to obtain a preset Bayesian network structure.

8. The solar street light brightness adjustment method based on Bayesian network according to claim 1, characterized in that: The step of determining the powered solar street lamps and the powered solar street lamps of each solar street lamp power supply group based on the light brightness priority specifically includes: Sorting the solar street lamps of the solar street lamp power supply group according to the priority of light brightness; When the brightness priority ranking of the solar street lamp is within a preset ranking range and the current power of the solar street lamp is lower than a preset power threshold, the solar street lamp is selected as the powered solar street lamp, and other solar street lamps in the solar street lamp power supply group are selected as the power supply solar street lamps.

9. The solar street light brightness adjustment method based on Bayesian network according to any one of claims 1 to 8, characterized in that: Before adjusting the brightness of the powered solar street lamp to the first target brightness and adjusting the brightness of the powered solar street lamp to the second target brightness, the method further includes: Determining the first target brightness according to the current power of the powered solar street lamp; The second target brightness is determined according to the current power of the powered solar street lamp.

10. An intelligent control system, characterized in that: It comprises 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 solar street lamp brightness adjustment method based on the Bayesian network as described in any one of claims 1 to 9.

Images