Intelligent lighting calculation method and system based on tunnel environment parameters

Abstract

The invention provides an intelligent lighting calculation method and system based on tunnel environment parameters, and relates to the lighting control technology, and the method comprises the steps: constructing a heterogeneous hypergraph model according to historical vehicle information, historical region information and historical environment parameters; inputting the actually measured vehicle information, the regional information and the environmental parameters into the heterogeneous hypergraph model to obtain updated node features and hyperedge weights; on the basis of the updated node features and hyperedge weights, fuzzy rules are generated through fuzzy logic, and weighted fusion is carried out on the fuzzy rules to obtain a control strategy; and obtaining an illumination demand of each region of the tunnel based on the control strategy, and obtaining an optimal illumination parameter of each lamp through a genetic algorithm based on the illumination demand. According to the method, the output of the hypergraph model is combined with the fuzzy logic to generate the control rule, the hybrid method not only can reserve the modeling capability of the hypergraph for the complex relationship, but also can utilize the advantage of the fuzzy logic to process uncertainty, the generated rule can adapt to environmental changes, and the intelligence of a control system is improved.

Description

Technical Field

[0001] The present invention relates to the field of lighting control technology, and in particular to an intelligent lighting calculation method and system based on tunnel environment parameters. Background Art

[0002] Tunnel intelligent lighting systems are an important component of modern transportation infrastructure. Through automatic adjustment of illumination, tunnel intelligent lighting control systems can achieve instant and precise lighting required for various weather and traffic conditions, thereby eliminating energy waste and insufficient lighting.

[0003] Traditional lighting methods typically use segmented dimming based on traffic flow, which lacks flexibility. In recent years, dynamic control based on environmental sensing has emerged, such as dynamic control based on traffic flow or light intensity, or lighting control using neural networks based on historical data. However, these existing methods often rely on single parameters or static rules, lacking the ability to model and dynamically decide on multi-source heterogeneous data, and struggle to balance safety, comfort, and energy efficiency. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent lighting calculation method and system based on tunnel environmental parameters to improve the above-mentioned problems. To achieve the above-mentioned purpose, the technical solutions adopted by the present invention are as follows:

[0005] In a first aspect, the present application provides an intelligent lighting calculation method based on tunnel environment parameters, comprising:

[0006] Construct a heterogeneous hypergraph model based on historical vehicle information, historical area information, and historical environmental parameters;

[0007] Input the measured vehicle information, regional information, and environmental parameters into the heterogeneous hypergraph model to obtain updated node features and hyperedge weights;

[0008] Based on the updated node features and hyperedge weights, fuzzy rules are generated through fuzzy logic, and the fuzzy rules are weightedly fused to obtain the control strategy;

[0009] The lighting requirements of each area of ​​the tunnel are obtained based on the control strategy, and based on the lighting requirements, the optimal lighting parameters of each lamp are obtained through a genetic algorithm.

[0010] In a second aspect, the present application further provides an intelligent lighting calculation system based on tunnel environment parameters, comprising:

[0011] The first module is used to build a heterogeneous hypergraph model based on historical vehicle information, historical area information and historical environmental parameters;

[0012] The second module is used to input measured vehicle information, regional information, and environmental parameters into the heterogeneous hypergraph model to obtain updated node features and hyperedge weights;

[0013] The third module is used to generate fuzzy rules through fuzzy logic based on the updated node features and hyperedge weights, and perform weighted fusion on the fuzzy rules to obtain the control strategy;

[0014] The fourth module is used to obtain the lighting requirements of each area of ​​the tunnel based on the control strategy, and obtain the optimal lighting parameters of each lamp through a genetic algorithm based on the lighting requirements.

[0015] The beneficial effects of the present invention are:

[0016] This invention integrates multimodal data, including vehicle, regional, and environmental data, through heterogeneous hypergraph modeling. This captures the dynamic interactions of various factors in the tunnel, improving perception accuracy. It converts complex hypergraph outputs into fuzzy rules, simulating human thinking, improving logical interpretability, and facilitating engineering debugging and verification. A genetic algorithm is used to optimize lighting parameters in each area, achieving an adaptive balance between safety, energy efficiency, and comfort.

[0017] Other features and advantages of the present invention will be set forth in the following description, and in part will become apparent from the description, or may be learned by practicing embodiments of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] 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.

[0019] Figure 1 This is a flow chart of the intelligent lighting calculation method based on tunnel environment parameters according to an embodiment of the present application;

[0020] Figure 2 This is a structural diagram of an intelligent lighting computing device based on tunnel environment parameters according to an embodiment of the present application.

[0021] Markings in the figure: 100-first module; 110-first construction unit; 120-second construction unit; 130-third construction unit; 140-fourth construction unit; 150-training unit; 200-second module; 300-third module; 310-decoupling unit; 320-mapping unit; 330-extraction unit; 340-fusion unit; 400-fourth module; 410-encoding unit; 420-fifth construction unit; 430-sixth construction unit; 440-seventh construction unit; 450-optimization unit. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0023] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0024] Example 1

[0025] See also Figure 1 , this embodiment provides an intelligent lighting calculation method based on tunnel environment parameters, including steps S100, S200, S300 and S400;

[0026] S100: Constructing a heterogeneous hypergraph model based on historical vehicle information, historical area information, and historical environmental parameters; specifically, including:

[0027] S110, clustering vehicles based on their locations, speeds, and vehicle types, and constructing multiple vehicle cluster nodes;

[0028] Since there may be a large number of vehicles, treating each vehicle as a node will significantly increase the number of nodes. To improve computational efficiency, this application will use the DBSCAN clustering algorithm to cluster vehicles and form vehicle group nodes, as follows:

[0029] S111. Acquire vehicle images, extract the position features, speed features, and vehicle type features of each vehicle through image recognition, and construct a feature vector for each vehicle;

[0030] S112. Clustering the vehicles using a clustering algorithm based on the feature vector of each vehicle to obtain multiple vehicle clusters;

[0031] This step uses the DBSCAN clustering algorithm for clustering, where the parameters of DBSCAN are adaptively adjusted. For example, the cluster neighborhood radius eps is adjusted according to the vehicle density. When the vehicle distribution density is high (such as in congestion), a larger eps is selected, and when the vehicle density is low, a smaller eps can be selected.

[0032] If there are outliers during clustering, the outliers can be classified into the nearest cluster and the cluster center can be readjusted;

[0033] S113, generating an embedded feature of each vehicle cluster based on the vehicle information in each cluster;

[0034] This step calculates the spatial distribution center of all vehicles in the cluster as the location of the cluster. The average speed of all vehicles is used as the speed of the cluster. The distribution of vehicle types within the cluster is also calculated, and the proportion of each type within the cluster is calculated. This generates embedded features based on the location, speed, and type information of the cluster.

[0035] S114: construct a vehicle cluster node based on the embedded features of the vehicle cluster.

[0036] S120, dividing the tunnel into regions according to a preset distance, and constructing multiple regional nodes based on the spatial characteristics of each region;

[0037] In this step, the tunnel can be divided into equal intervals of 30-50 meters. Alternatively, the functions and structures of different sections of the tunnel can be considered when dividing the areas. For example, if the 60-meter area at the tunnel entrance (or exit) is divided into one area, and there is a 112-meter curve in the middle of the tunnel, the 112-meter curve can be divided into one area.

[0038] After the areas are divided according to the characteristics of the tunnel structure, the areas are numbered in the order of the areas, and the corresponding information such as the length, slope, curve and height of these areas are obtained;

[0039] Each region is regarded as a region node, and the above number and information are used as the embedding features of the region node.

[0040] S130, constructing multiple environmental nodes based on information from sensors inside and outside the tunnel;

[0041] Sensors arranged inside and outside the tunnel are used to measure data such as light intensity, temperature and humidity, and visibility. Sensors are generally arranged at equal intervals. Certain special or important areas can be adjusted according to the situation. For example, light intensity sensors at tunnel entrances and exits need to be arranged more densely to accurately monitor external light changes.

[0042] S140, constructing a hyperedge by connecting the relevant nodes;

[0043] Connect the environment nodes in the same area to each other, because the environment parameters will affect each other and have interaction. In addition, hyperedges can be used to connect the vehicle cluster nodes with nearby regional nodes and corresponding environment nodes.

[0044] The features of the nodes connected by the hyperedge are spliced ​​to obtain the initial hyperedge features; the weight of the hyperedge is used to represent the strength of the association between nodes. In this step, the features of the hyperedge are input into a pre-built multi-layer perceptron (MLP) to obtain the hyperedge weight, and the hyperedge weight can be dynamically updated with the convolution of the heterogeneous hypergraph.

[0045] S150. Perform convolution training on the heterogeneous hypergraph to obtain a heterogeneous hypergraph model.

[0046] Convolution training includes vertex convolution and hyperedge convolution, which are used to aggregate features between vertices and hyperedges respectively, thereby effectively transmitting and integrating the connectivity and correlation between nodes. Through heterogeneous hypergraph convolution, the model can effectively capture and represent complex high-order relationships.

[0047] In addition, since the environment and driving conditions in the tunnel are constantly changing, in order to adapt to changes in data, this application adopts a dynamic heterogeneous hypergraph construction method. Dynamic heterogeneous hypergraph construction uses the KNN (nearest neighbor) or KMeans (K-means) algorithm to construct a dynamic heterogeneous hypergraph. The KNN method is used to generate basic hyperedges. In the KNN method, for each node, its distance to other nodes is calculated, and the K vertices closest to it are selected. These K vertices are connected to the current vertex to form a hyperedge. KMeans clustering is used to expand the set of adjacent hyperedges by dividing the vertices into different clusters, each cluster constituting a hyperedge; the KNN method and KMeans can be used in combination to extract local and global relationships.

[0048] S200: Inputting measured vehicle information, regional information, and environmental parameters into a heterogeneous hypergraph model to obtain updated node features and hyperedge weights;

[0049] S300, based on the updated node features and hyperedge weights, generating fuzzy rules through fuzzy logic, and performing weighted fusion on the fuzzy rules to obtain a control strategy;

[0050] Fuzzy logic is a mathematical method for dealing with uncertainty and ambiguity. It extends traditional binary logic (Boolean logic) by allowing logical values ​​to range between completely true (1) and completely false (0). The core idea of ​​fuzzy logic is to simulate human thinking through fuzzy sets and fuzzy rules, thereby better handling ambiguity and uncertainty in the real world.

[0051] S310, performing semantic decoupling on the updated node features to obtain fuzzy variables;

[0052] The output of the heterogeneous hypergraph includes node features for different nodes. Each node feature can express a corresponding semantic meaning. For example, vehicle node features express the degree of traffic congestion, environment node features express field of view visibility, and regional features express risk rate. This step maps the node features into fuzzy variables based on their semantics to generate fuzzy rules.

[0053] S311, generating a predefined semantic dimension for each node according to the node type;

[0054] Based on the type of each node, query the pre-defined semantic dimension table, such as:

[0055] Vehicle nodes [congestion level, speed anomaly, vehicle type distribution];

[0056] Regional nodes [congestion level, risk factor, and visual clarity];

[0057] Environmental nodes [risk factor, humidity, visual clarity];

[0058] S312: Based on the semantic dimension, all node features are processed by a structured projection matrix to obtain reduced-dimensional semantic features;

[0059] First, count the relevant nodes of each semantic dimension. For example, the risk coefficient is related to the regional node and the environmental node. The historical accident rate of the regional node and the characteristic information such as curves and slopes are obtained. At the same time, the characteristics of all environmental nodes in the area (humidity, visibility, etc.) are obtained. The characteristics of the regional node and the characteristics of the environmental node both express risk information. Therefore, the characteristics of the risk coefficient can be represented by a combination of the characteristics of the two nodes. However, since the combined feature dimension is high, it is not convenient for subsequent calculations. Therefore, dimensionality reduction processing is required in this step.

[0060] Specifically, data dimensionality reduction can be performed through Linear Discriminant Analysis (LDA) or Kernel PCA to obtain the dimensionality reduction semantic features of the risk coefficient.

[0061] Based on the same processing idea, the dimensionality reduction semantic features of the congestion degree can be calculated according to the vehicle density, speed and the area in which it is located. The dimensionality reduction semantic features of the field of view clarity can be calculated according to the visibility data measured in a certain area and the curve data in the area.

[0062] S313: Obtain fuzzy variables according to the dimension-reduced semantic features.

[0063] This step uses a pre-trained mapping model to map the reduced-dimensionality semantic features into specific numerical values. For example, based on the reduced-dimensionality semantic features of the risk coefficient after the above step, the current risk coefficient value for a certain area is 0.6, which is used as a fuzzy variable. Based on the same idea, fuzzy variables such as congestion level and field of view clarity can also be derived.

[0064] S320, using a preset membership function to perform fuzzy mapping on each fuzzy variable to obtain the membership degree of each fuzzy variable under each language value;

[0065] The membership function can be triangular function, trapezoidal function and Sigmoid function. The triangular function is suitable for most continuous variables, the trapezoidal function is suitable for variables with clear safety thresholds, and the Sigmoid function is used in scenarios with rapid response to extreme risks.

[0066] First, we need to classify the linguistic values ​​(such as low risk, medium risk, and high risk) based on the scope and domain of the fuzzy variable and determine the range of values ​​for each linguistic value. For example, if the fuzzy variable corresponding to the risk factor is 0-1, 0-0.3 is low risk, 0.3-0.7 is medium risk, and 0.6-1 is high risk (the ranges can overlap and are ambiguous).

[0067] Assuming a triangular membership function, the formula is as follows:

[0068]

[0069] μ(x) is the degree of membership, x is the fuzzy variable, a is the left boundary, b is the vertex, and c is the right boundary; vertex b is usually the middle value of the two boundary points.

[0070] From the above formula, we can see that the membership degree of the risk coefficient fuzzy variable (0.6) is 0 under the low-risk language value;

[0071] Calculate the membership of the risk coefficient fuzzy variable (0.6) under the medium risk language value (the vertex is 0.5): μ(x) = (0.7-0.6) / (0.7-0.5) = 0.5;

[0072] Calculate the membership of the risk coefficient fuzzy variable (0.6) under the high-risk language value (the peak is 0.8):

[0073] μ(x)=(0.6-0.6) / (1-0.8)=0;

[0074] Therefore, the membership of the risk coefficient fuzzy variable (0.6) under the three linguistic values ​​is 0, 0.5, and 0;

[0075] S330, extracting fuzzy rules from a rule base according to the membership degree, and determining the weight of each fuzzy rule according to the updated hyperedge weight;

[0076] This step is to match the fuzzified input variables with predefined rules. For example, according to the risk coefficient fuzzy variable membership of 0, 0.5, 0, the brightness of the area is controlled at 70% of the maximum value in the rule base.

[0077] According to the fuzzy variable membership corresponding to the field of view clarity, the brightness control parameters and color temperature control parameters of the area can be extracted from the rule base;

[0078] In emergency situations, such as an accident in a tunnel, full lighting is mandatory.

[0079] The weight of each fuzzy rule is determined as follows:

[0080] Obtain the nodes involved in generating the fuzzy rule. For example, the fuzzy rule for risk coefficient extraction is generated by the participation of the environment node and the regional node. That is, obtain the weight of the hyperedge connecting the environment node and the regional node, and use the hyperedge weight as the weight of the fuzzy rule.

[0081] S340. Screen and weightedly fuse the fuzzy rules according to their weights to obtain a control strategy.

[0082] When the extracted fuzzy rules are repeated and contradictory, for example, the brightness of the area is controlled at 70% based on the risk coefficient, and the brightness of the area is controlled at 80% based on the field of view clarity, that is, the two parameters are weightedly fused according to the weight of the fuzzy rules to obtain the comprehensive brightness control parameter.

[0083] When some of the extracted rules have decision-making power (such as forcing full lighting in an emergency), all other contradictory rules are eliminated to ensure that safety has the highest priority.

[0084] All control parameters of a certain area are combined to obtain the final control strategy.

[0085] S400 : Obtaining lighting requirements for each area of ​​the tunnel based on a control strategy, and obtaining optimal lighting parameters for each lamp through a genetic algorithm based on the lighting requirements.

[0086] Each divided area has multiple lamps, and the types of lamps may be different. In order to optimize lighting energy consumption and visual comfort, this step uses a genetic algorithm to determine the lighting parameters of each lamp. The details are as follows:

[0087] Based on the control strategy, the control parameters are expanded into range values. The point values ​​can be reduced or increased by a certain percentage to determine the lower and upper limits of the parameter range. The lighting requirements of each area are obtained, and the lighting parameters of each lamp in the area are encoded into gene fragments to generate the initial chromosome. The lighting parameters of the lamps include brightness and color temperature.

[0088] According to the optional range of the parameters, parameters are randomly selected in the range for encoding to generate the initial population. Subsequent mutation operations also need to meet the range constraints.

[0089] When coding, the lighting characteristics of the area also need to be considered. For example, the lighting in the tunnel entrance area should gradually transition from the outside to the inside, that is, the parameter changes of the lamps are regular. Therefore, for such special areas, additional coding constraints need to be added.

[0090] Based on the lighting parameters of the lamps, the lamp gradient difference, color temperature deviation index and transition parameters are calculated to construct the comfort item.

[0091] The lamp gradient difference is to calculate the brightness and color temperature difference between adjacent lamps. The smaller the difference, the more comfortable it is. The color temperature deviation index is to calculate the degree to which the color temperature of the lamp deviates from the preset human eye comfortable color temperature value. The smaller the deviation, the more comfortable it is. The transition parameter is to calculate the parameter difference of the lamps at the junction of different areas to ensure a smooth transition between different areas.

[0092] The lamp gradient difference, color temperature deviation index and transition parameter are weighted and summed according to the preset weights to obtain the comfort item.

[0093] Energy consumption items are constructed based on the lighting parameters of lamps;

[0094] Construct a fitness function based on comfort terms and energy consumption terms;

[0095] Based on the fitness function, the chromosomes are selected, crossed and mutated to obtain the optimal lighting parameters of the lamp.

[0096] Example 2

[0097] See also Figure 2 This embodiment provides an intelligent lighting calculation system based on tunnel environment parameters, including:

[0098] The first module 100 is used to construct a heterogeneous hypergraph model based on historical vehicle information, historical area information and historical environmental parameters;

[0099] The second module 200 is used to input the measured vehicle information, regional information and environmental parameters into the heterogeneous hypergraph model to obtain updated node features and hyperedge weights;

[0100] The third module 300 is used to generate fuzzy rules through fuzzy logic based on the updated node features and hyperedge weights, and perform weighted fusion on the fuzzy rules to obtain a control strategy;

[0101] The fourth module 400 is used to obtain the lighting requirements of each area of ​​the tunnel based on the control strategy, and obtain the optimal lighting parameters of each lamp through a genetic algorithm based on the lighting requirements.

[0102] As an optional implementation, the first module 100 includes:

[0103] A first constructing unit 110 is configured to cluster vehicles based on their locations, speeds, and vehicle types, and construct a plurality of vehicle cluster nodes;

[0104] The second construction unit 120 is configured to divide the tunnel into regions according to a preset distance and construct a plurality of region nodes based on the spatial characteristics of each region;

[0105] The third construction unit 130 is used to construct multiple environmental nodes based on information from sensors inside and outside the tunnel;

[0106] A fourth construction unit 140 is configured to construct a hyperedge by connecting related nodes;

[0107] The training unit 150 is used to perform convolution training on the heterogeneous hypergraph to obtain a heterogeneous hypergraph model.

[0108] As an optional implementation, the third module 300 includes:

[0109] a decoupling unit 310 for performing semantic decoupling on the updated node features to obtain fuzzy variables;

[0110] A mapping unit 320 is configured to perform fuzzy mapping on each fuzzy variable using a preset membership function to obtain the membership degree of each fuzzy variable under each linguistic value;

[0111] an extraction unit 330 for extracting fuzzy rules from a rule base according to the membership degree, and determining a weight of each fuzzy rule according to the updated hyperedge weight;

[0112] The fusion unit 340 is used to screen and weighted fuse the fuzzy rules according to their weights to obtain a control strategy.

[0113] As an optional implementation, the control strategy includes lighting requirements for each area of ​​the tunnel; the fourth module 400 includes:

[0114] The encoding unit 410 is used to encode the lighting parameters of each lamp in the area into gene fragments based on the lighting requirements of each area to generate an initial chromosome; the lighting parameters of the lamps include brightness and color temperature;

[0115] A fifth constructing unit 420 is configured to calculate the lamp gradient difference, the color temperature deviation index, and the transition parameter based on the lighting parameters of the lamp to construct a comfort item;

[0116] A sixth constructing unit 430 is configured to construct an energy consumption item based on the lighting parameters of the lamp;

[0117] A seventh constructing unit 440 is configured to construct a fitness function based on the comfort item and the energy consumption item;

[0118] The optimization unit 450 is used to perform selection, crossover and mutation operations on the chromosome based on the fitness function to obtain the optimal lighting parameters of the lamp.

[0119] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0120] The above 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 easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. An intelligent lighting calculation method based on tunnel environment parameters, characterized in that: include: Construct a heterogeneous hypergraph model based on historical vehicle information, historical area information, and historical environmental parameters; Input the measured vehicle information, regional information, and environmental parameters into the heterogeneous hypergraph model to obtain updated node features and hyperedge weights; Based on the updated node features and hyperedge weights, fuzzy rules are generated through fuzzy logic, and the fuzzy rules are weightedly fused to obtain the control strategy; The lighting requirements of each area of ​​the tunnel are obtained based on the control strategy, and based on the lighting requirements, the optimal lighting parameters of each lamp are obtained through a genetic algorithm.

2. The intelligent lighting calculation method based on tunnel environment parameters according to claim 1 is characterized in that: A heterogeneous hypergraph model is constructed based on historical vehicle information, regional information, and environmental parameters, including: Cluster vehicles based on their location, speed, and vehicle type information to build multiple vehicle cluster nodes; Divide the tunnel into regions according to preset distances, and construct multiple regional nodes based on the spatial characteristics of each region; Construct multiple environmental nodes based on the information from sensors inside and outside the tunnel; By connecting the relevant nodes, a hyperedge is constructed; Convolutional training is performed on the heterogeneous hypergraph to obtain a heterogeneous hypergraph model.

3. The intelligent lighting calculation method based on tunnel environment parameters according to claim 1 is characterized in that: Based on the updated node features and hyperedge weights, fuzzy rules are generated through fuzzy logic, and the fuzzy rules are weightedly fused to obtain the control strategy, including: Perform semantic decoupling on the updated node features to obtain fuzzy variables; Use the preset membership function to fuzzify each fuzzy variable and obtain the membership degree of each fuzzy variable under each language value; extracting fuzzy rules from a rule base according to the membership degree, and determining the weight of each fuzzy rule according to the updated hyperedge weight; The fuzzy rules are screened and weightedly fused according to their weights to obtain the control strategy.

4. The intelligent lighting calculation method based on tunnel environment parameters according to claim 2, characterized in that: The control strategy includes the lighting requirements for each zone of the tunnel; The genetic algorithm is used to optimize the lighting parameters of each tunnel section, including: Based on the lighting requirements of each area, the lighting parameters of each lamp in the area are encoded into gene fragments to generate initial chromosomes; the lighting parameters of the lamps include brightness and color temperature; Based on the lighting parameters of the lamps, the lamp gradient difference, color temperature deviation index and transition parameters are calculated to construct the comfort item. Energy consumption items are constructed based on the lighting parameters of lamps; Construct a fitness function based on comfort terms and energy consumption terms; Based on the fitness function, the chromosomes are selected, crossed and mutated to obtain the optimal lighting parameters of the lamp.

5. The intelligent lighting calculation method based on tunnel environment parameters according to claim 2 is characterized in that: Cluster vehicles based on their location, speed, and vehicle type information, and build multiple vehicle cluster nodes, including: Obtain vehicle images, extract the location features, speed features, and vehicle type features of each vehicle through image recognition, and construct a feature vector for each vehicle; Based on the feature vector of each vehicle, the vehicles are clustered using a clustering algorithm to obtain multiple vehicle clusters; Generate embedding features for each vehicle cluster based on the vehicle information in each cluster; The vehicle cluster nodes are constructed based on the embedded features of the vehicle cluster.

6. The intelligent lighting calculation method based on tunnel environment parameters according to claim 2, characterized in that: The updated node features are semantically decoupled to obtain fuzzy variables, including: Generate predefined semantic dimensions for each node based on the node type; Based on the semantic dimension, all node features are processed by a structured projection matrix to obtain reduced-dimensional semantic features; A fuzzy variable is obtained according to the dimension-reduced semantic features.

7. An intelligent lighting calculation system based on tunnel environment parameters, characterized in that: include: The first module is used to build a heterogeneous hypergraph model based on historical vehicle information, historical area information and historical environmental parameters; The second module is used to input measured vehicle information, regional information, and environmental parameters into the heterogeneous hypergraph model to obtain updated node features and hyperedge weights; The third module is used to generate fuzzy rules through fuzzy logic based on the updated node features and hyperedge weights, and perform weighted fusion on the fuzzy rules to obtain the control strategy; The fourth module is used to obtain the lighting requirements of each area of ​​the tunnel based on the control strategy, and obtain the optimal lighting parameters of each lamp through a genetic algorithm based on the lighting requirements.

8. The intelligent lighting calculation system based on tunnel environment parameters according to claim 1, characterized in that: The first module includes: A first construction unit is used to cluster vehicles based on their locations, speeds, and vehicle types, and to construct a plurality of vehicle cluster nodes; The second construction unit is used to divide the tunnel into regions according to a preset distance and construct a plurality of regional nodes based on the spatial characteristics of each region; A third construction unit is used to construct multiple environmental nodes based on information from sensors inside and outside the tunnel; A fourth construction unit is configured to construct a hyperedge by connecting related nodes; The training unit is used to perform convolution training on the heterogeneous hypergraph to obtain a heterogeneous hypergraph model.

9. The intelligent lighting calculation method based on tunnel environment parameters according to claim 1, characterized in that: The third module includes: The decoupling unit is used to semantically decouple the updated node features to obtain fuzzy variables; A mapping unit, configured to perform fuzzy mapping on each fuzzy variable using a preset membership function to obtain the membership degree of each fuzzy variable under each language value; an extraction unit, configured to extract fuzzy rules from a rule base according to the membership degree, and determine a weight of each fuzzy rule according to the updated hyperedge weight; The fusion unit is used to screen and weightedly fuse the fuzzy rules according to their weights to obtain the control strategy.

10. The intelligent lighting calculation method based on tunnel environment parameters according to claim 7, characterized in that: The control strategy includes the lighting requirements of each area of ​​the tunnel; the fourth module includes: An encoding unit, configured to encode the lighting parameters of each lamp in each area into gene fragments based on the lighting requirements of each area, thereby generating an initial chromosome; the lighting parameters of the lamps include brightness and color temperature; A fifth construction unit is configured to calculate the lamp gradient difference, the color temperature deviation index, and the transition parameter based on the lighting parameters of the lamp to construct a comfort item; A sixth construction unit, configured to construct an energy consumption item based on the lighting parameters of the lamp; A seventh construction unit is used to construct a fitness function based on the comfort term and the energy consumption term; The optimization unit is used to perform selection, crossover and mutation operations on chromosomes based on the fitness function to obtain the optimal lighting parameters of the lamp.

Images