A flying robot for intelligent lighting system automated configuration

Abstract

The application provides a flying robot for automatic configuration of an intelligent lighting system, and a configuration method, which comprises the following steps: constructing a digital twin model; based on a function area label and a preset lighting standard knowledge graph, a rule reasoning engine combines a natural light dynamic value to generate differentiated illumination, color temperature and uniformity requirements; the three-dimensional position, angle and power of a lamp are calculated, and a layout scheme containing space coordinates and parameters is output; the robot reaches a predetermined position, injects an identifier, topology and function area label into the lamp through a wireless protocol, so that the lamp self-assembles a regional perception mesh network; before installation, light simulation is carried out in the digital twin body, if the simulation and requirements deviate by more than a threshold value, the layout is iterated; after reaching the standard, a configuration instruction is issued; in a running cycle, the robot regularly collects actual illumination data, compares the target value and feeds back the knowledge graph to iteratively update the rules. The application reduces the cost, improves the adaptability, enables the configuration system to have a continuous optimization capability, and improves the lighting quality and energy efficiency level.

Description

Technical Field

[0001] This invention relates to the field of robotics, and more specifically to a flying robot for the automated configuration of intelligent lighting systems. Background Technology

[0002] In existing technologies, when intelligent lighting systems are deployed in actual physical spaces, the corresponding locations of lighting devices and physical spaces need to be manually searched and checked one by one. However, this method is acceptable when there are few devices installed (or large household appliances placed on the ground), but when there are many devices or the devices are installed at high positions (such as on the roof), it takes too long to find the devices by eye and cannot meet the adaptability of lighting configuration to the actual needs of different functional areas. Summary of the Invention

[0003] Based on the above-mentioned problems, this invention proposes a flying robot for the automated configuration of intelligent lighting systems. Through this invention, the configuration time cost and manpower input are reduced, the adaptability of lighting configuration to the actual needs of different functional areas is improved, the configuration system has the ability to continuously optimize, and the lighting quality and energy efficiency level of long-term operation are improved.

[0004] In view of this, one aspect of the present invention proposes a flying robot for automated configuration of an intelligent lighting system, comprising: The flying robot performs autonomous flight traversal within the target space, simultaneously collecting three-dimensional point cloud data, light intensity distribution data, and infrared thermal imaging data. Based on multimodal sensor fusion technology, it constructs a digital twin model that includes spatial geometric features, functional area semantic labels, and real-time light environment features. The functional area semantic labels are automatically labeled as at least one of office area, meeting area, rest area, or passage area through image recognition of object features within the space. Based on the semantic tags of functional areas in the digital twin model, combined with the preset lighting standard knowledge graph, the target illuminance value, color temperature range and lighting uniformity requirements of each functional area are automatically derived using a rule reasoning engine. The target illuminance value is dynamically adjusted according to the contribution of natural light in the real-time light environment characteristics to generate a differentiated lighting requirement parameter set. Based on the differentiated lighting demand parameter set and the spatial geometric features of the digital twin model, the flying robot uses a multi-objective optimization fitness function to calculate the three-dimensional installation position, installation angle and power configuration scheme of the lamps. The multi-objective optimization fitness function simultaneously optimizes three objective dimensions: lighting uniformity, energy consumption minimization and installation convenience, and outputs a layout scheme containing the spatial coordinates and configuration parameters of each lamp. The flying robot carries an Internet of Things (IoT) communication module and flies sequentially to the predetermined installation positions of each lamp according to the layout scheme. It establishes a temporary connection with the lamp to be installed through a near-field wireless communication protocol and injects configuration information including lamp identifiers, network topology relationships, and semantic tags of the functional areas, so that each lamp can automatically form a mesh IoT topology with regional awareness capabilities. Before the physical lighting fixtures are installed, the layout scheme is simulated in the digital twin model to calculate the deviation between the simulated lighting effect and the set of differentiated lighting requirements parameters. When the deviation exceeds a preset threshold, the iterative optimization of the optimal layout planning step of the lighting fixtures is automatically triggered until the deviation meets the requirements and the final configuration command is output to each lighting fixture. After configuration, the flying robot periodically flies to collect actual lighting effect data within a preset period. It compares and analyzes the actual lighting data with the target illuminance value, extracts configuration deviation features, and feeds them back to the lighting standard knowledge graph for incremental learning, thereby updating the reasoning rules of the rule reasoning engine.

[0005] Optionally, the flying robot performs autonomous flight traversal within the target space, simultaneously collecting 3D point cloud data, illumination intensity distribution data, and infrared thermal imaging data. The steps involved in constructing a digital twin model based on multimodal sensor fusion technology, incorporating spatial geometric features, functional area semantic labels, and real-time lighting environment features, include: The flying robot generates a full-coverage flight path based on pre-acquired target space boundary information using a hybrid trajectory mode that combines three-dimensional spiral ascent with horizontal grid scanning. During flight, it dynamically adjusts the trajectory through real-time obstacle detection to ensure that the sensor field of view covers the space without omission. At each sampling point along the full-coverage flight path, the flying robot synchronously activates a lidar, a illuminometer, and an infrared thermal imager through a hardware triggering mechanism to collect three-dimensional point cloud data, light intensity distribution data, and infrared thermal imaging data, respectively. It also uniformly labels each set of collected data with a timestamp and six-degree-of-freedom pose information to establish a spatiotemporal correspondence between multimodal data. After registering the three-dimensional point cloud data with infrared thermal imaging data, the spatial geometric features and thermal radiation features of objects in the space are extracted. The objects such as tables, chairs, conference tables, sofas, and aisle markers are identified through a pre-trained deep learning model. Based on the identified object category combinations and their spatial distribution relationships, the area where the objects are located is automatically labeled as at least one of office area, meeting area, rest area, or aisle area to generate functional area semantic labels. Spatially register the light intensity distribution data with the three-dimensional point cloud data, map real-time illuminance values ​​onto the three-dimensional grid nodes of the digital twin model, and simultaneously calculate and store light environment characteristic parameters such as average illuminance, illuminance uniformity, and natural light window position for each functional area. The spatial geometric features, the functional area semantic labels, and the light environment feature parameters are fused according to a unified coordinate system and data structure to construct a three-layer digital twin model containing a geometric layer, a semantic layer, and a light environment layer. An association index between the data in each layer is established to support subsequent lighting requirement reasoning and layout planning.

[0006] Optionally, the step of automatically deriving the target illuminance value, color temperature range, and lighting uniformity requirements of each functional area based on the functional area semantic tags in the digital twin model, combined with a preset lighting standard knowledge graph, using a rule reasoning engine, and dynamically adjusting the target illuminance value according to the contribution of natural light in the real-time light environment features to generate a differentiated lighting requirement parameter set includes: The rule reasoning engine is used to read the semantic tags of each functional area in the digital twin model. The lighting standard node corresponding to each functional area semantic tag is retrieved in the lighting standard knowledge graph. The benchmark target illuminance value, recommended color temperature range and minimum uniformity coefficient stored in the lighting standard node are extracted, and the special lighting rules associated with the functional area are identified. The rule reasoning engine determines the floor height, area, and adjacency relationship of each functional area based on the spatial geometric features in the digital twin model, and performs contextual correction on the benchmark target illuminance value in combination with the special lighting rules. Specifically, for areas with floor heights exceeding the conventional value, the target illuminance value is increased to compensate for light loss, and transition lighting requirements are set for adjacent functional areas to avoid visual abrupt changes. The location, area, and current illuminance of the natural light window in each functional area are extracted from the real-time light environment characteristics. The natural light contribution of each functional area is calculated. The natural light contribution represents the degree to which natural light meets the lighting needs of the area. Areas significantly affected by natural light and areas that rely on artificial lighting year-round are also identified. For functional areas where the contribution of natural light is higher than a preset threshold, the target illuminance value after context correction is reduced based on the quantified value of the contribution of natural light to achieve natural light compensation, thereby generating a dynamic target illuminance value for the functional area. The dynamic target illuminance value adapts to changes in natural light. For areas with low contribution of natural light, the target illuminance value after context correction remains unchanged. The target illuminance value, recommended color temperature range, minimum uniformity coefficient, and natural light compensation strategy identifier of each functional area after dynamic compensation adjustment are integrated into the lighting requirement parameters of that functional area. The lighting requirement parameters of all functional areas are summarized to form a differentiated lighting requirement parameter set, and a mapping relationship between functional area identifiers and corresponding lighting requirement parameters is established in the differentiated lighting requirement parameter set.

[0007] Optionally, the flying robot, based on the differentiated lighting demand parameter set and the spatial geometric features of the digital twin model, uses a multi-objective optimization fitness function to calculate the three-dimensional installation position, installation angle, and power configuration scheme of the luminaires. This multi-objective optimization fitness function simultaneously optimizes three objective dimensions: lighting uniformity, energy consumption minimization, and installation convenience. The step of outputting a layout scheme containing the spatial coordinates and configuration parameters of each luminaire includes: Based on the spatial geometric features of the digital twin model, the flying robot identifies installable structures such as ceiling planes, beams, and ceiling joists in the target space. On these installable structures, it generates a set of three-dimensional candidate installation points according to preset safety distance rules, and labels each candidate installation point with its structural type, load-bearing capacity, and construction convenience level. At the same time, it excludes infeasible installation points located in ventilation openings, fire-fighting facilities, or structural weak points. Based on the target illuminance value and color temperature range of each functional area in the set of differentiated lighting demand parameters, candidate lamp models that meet the luminous efficacy requirements are selected from the preset lamp library. The effective lighting radius of each lamp model at different installation heights is calculated. The initial number of lamps required for each functional area is estimated by combining the area of ​​each functional area, forming a configuration candidate set that includes lamp models, initial quantities and applicable functional areas. The multi-objective optimization fitness function takes the candidate installation point set and the configuration candidate set as input to construct an optimization model that includes an objective function of lighting uniformity, an objective function of energy consumption minimization, and an objective function of installation convenience. The installation location selection, installation angle, and power configuration of each lamp are adjusted through an iterative search mechanism. During the iteration process, the illuminance distribution of each functional area, the total energy consumption of the system, and the construction complexity score are calculated under each configuration combination. The Pareto front screening method is used to retain multiple sets of non-dominated solutions. Constraint verification is performed on each candidate layout scheme in the multiple sets of non-dominated solutions to check whether it meets the minimum uniformity coefficient, the transition lighting requirements of adjacent functional areas, and the maximum load limit of a single structural point specified in the set of differentiated lighting requirements parameters. Schemes that do not meet the constraints are locally adjusted and corrected, including fine-tuning the position of the lamps, replacing the low-power lamp model, or adding auxiliary lighting points. The optimal layout scheme is selected from the verified and corrected non-dominated solutions according to the preset comprehensive evaluation criteria. The comprehensive evaluation criteria take into account the weighted scores of three objective dimensions and the robustness of the scheme. The three-dimensional spatial coordinates, installation angle, power level, lamp model and functional area identifier of each lamp in the optimal layout scheme are extracted to form a layout scheme containing complete configuration parameters and output to the subsequent IoT topology self-organization step.

[0008] Optionally, the flying robot carries an IoT communication module and flies sequentially to the predetermined installation positions of each lamp according to the layout scheme. It establishes a temporary connection with the lamps to be installed via a near-field wireless communication protocol and injects configuration information including lamp identifiers, network topology relationships, and semantic labels of the functional areas, enabling each lamp to automatically form a mesh IoT topology with area awareness capabilities. This step includes: The flying robot reads the three-dimensional spatial coordinates and functional area identifiers of all lamps in the layout scheme, generates a configuration access sequence according to the principle of functional area priority and shortest distance within the area, prioritizes the configuration of lamps in the same functional area to quickly establish a communication backbone network within the area, and assigns a globally unique lamp identifier to each lamp in the access sequence. The flying robot flies sequentially to the predetermined installation position of each lamp according to the configured access sequence. When it approaches the target position, it hovers and locates itself by using the spatial geometric features of the digital twin model. It then activates the Internet of Things communication module to send near-field wireless communication signals, establishes a point-to-point temporary connection with the target lamp in the configuration state, and verifies the lamp's identity and communication reliability through a handshake protocol. The flying robot generates a configuration information package for the currently connected lighting fixture. The configuration information package includes the globally unique identifier of the lighting fixture, the semantic label of its functional area, the spatial coordinates and power configuration parameters of the lighting fixture in the layout scheme, as well as a list of neighboring lighting fixtures and communication priorities calculated based on the area-aware routing strategy. The list of neighboring lighting fixtures includes the identifiers of several lighting fixtures that are closest to each other in the same area and the identifiers of gateway lighting fixtures across areas. The configuration information package is injected into the storage unit of the target lighting fixture through the temporary connection. The lamps with injected configuration information actively search and establish communication links with neighboring lamps according to the list of neighboring lamps. They prioritize establishing high-priority links with neighboring lamps in the same functional area to form a subnet within the area, and then establish low-priority links with cross-regional gateway lamps to achieve inter-regional interconnection. As the flying robot completes the configuration of more lamps, the communication links between each lamp are gradually connected to form a mesh topology covering all functional areas. After the flying robot completes the injection of configuration information for all lamps, it broadcasts a topology construction completion command to the network. Upon receiving the command, each lamp activates its region awareness function based on the semantic tags of the functional areas, enabling lamps within the same functional area to identify each other's region affiliation and form a collaborative control group. Lamps in different functional areas can execute differentiated lighting strategies according to the semantic tags of the functional areas, thereby realizing the self-organization of an intelligent lighting network with region awareness capabilities.

[0009] Optionally, before the physical lighting fixtures are installed, a lighting simulation of the layout scheme is performed in the digital twin model to calculate the deviation between the simulated lighting effect and the set of differentiated lighting requirement parameters. When the deviation exceeds a preset threshold, the iterative optimization of the optimal layout planning step for the lighting fixtures is automatically triggered until the deviation meets the requirements, and then the final configuration command is output to each lighting fixture. This includes the following steps: In the digital twin model, a corresponding virtual lamp instance is created according to the configuration parameters of each lamp in the layout scheme. The photometric characteristic data of the real physical lamp is loaded for each virtual lamp instance, including the light intensity distribution curve, beam angle, color temperature spectrum and power-luminous efficacy relationship. The virtual lamp instance is placed in the corresponding position of the digital twin model according to the three-dimensional spatial coordinates and installation angle specified in the layout scheme. Based on the spatial geometric features and material reflection properties in the digital twin model, path tracing simulation is performed on the light emitted by each virtual lamp instance to calculate the propagation, reflection, refraction and attenuation process of light in space. The light contribution of all virtual lamps is accumulated at the key evaluation points of each functional area to generate a simulated lighting effect dataset containing the illuminance value, color temperature value and illuminance uniformity of each evaluation point. The simulated lighting effect dataset is compared and analyzed with the differentiated lighting requirement parameter set by functional area. The degree of deviation of each functional area in three dimensions, namely target illuminance value, color temperature range and lighting uniformity requirement, is calculated. The location and amount of the maximum deviation point are extracted for each functional area. The deviation information of all functional areas is summarized to form a comprehensive deviation index. The key influencing factors that cause deviation are identified, including improper lamp position, insufficient power configuration or light blockage. When the comprehensive deviation index exceeds the preset threshold, the system automatically extracts the key influencing factors and the maximum deviation point information as optimization constraints and feeds them back to the optimal layout planning step of the lighting fixtures. Under the premise of keeping the configuration of the functional areas that meet the requirements unchanged, the system re-executes the multi-objective optimization fitness function for the functional areas with excessive deviation, adjusts the installation position, installation angle or power configuration of the relevant lighting fixtures, generates an improved layout scheme, and returns to this virtual-real linkage verification step to re-perform the illumination simulation, forming an iterative closed loop of verification-optimization-re-verification. When the comprehensive deviation index meets the preset threshold requirement, the layout scheme is confirmed as the final optimized scheme. The complete configuration parameters of each lamp in the final optimized scheme are extracted, including spatial coordinates, installation angle, power level, dimming curve and start-up sequence. A set of configuration instructions that can be directly executed is generated. The configuration instruction set is distributed to each lamp through the Internet of Things topology, so that each lamp can perform actual lighting output according to the parameters that have been finally verified.

[0010] Optionally, in the operation of constructing a digital twin model containing spatial geometric features, functional region semantic labels, and real-time light environment features based on multimodal sensor fusion technology, the digital twin model is constructed using an adaptive weighted fusion algorithm, and the confidence evaluation formula for the adaptive weighted fusion is:

[0011] in, This represents the overall confidence score of the fused digital twin model; M represents the total number of sensor modes involved in the fusion. The basic weighting coefficients for the m-th mode are preset based on the inherent accuracy of the sensor. This represents the dynamic quality factor of the m-th mode under the current environment, calculated in real time using the signal-to-noise ratio; This represents the completeness index of the m-th modality data, reflecting the data coverage rate; Indicates the spatial density influence coefficient; The spatial density of the sampling points is represented by the number of point clouds per unit volume.

[0012] Optionally, the dynamic adjustment of the target illuminance value based on the contribution of natural light in the real-time light environment characteristics generates a set of differentiated lighting demand parameters. The dynamic adjustment of the target illuminance value is calculated using the following natural light compensation model:

[0013] in, This represents the target illuminance value of artificial lighting required for this functional area at time t; This represents the reference illuminance value for this functional area, derived from the lighting standards knowledge graph. The natural light utilization efficiency coefficient depends on the light transmission performance and location of the window. This represents the real-time illuminance value of natural light measured at time t; It indicates the angle of incidence of natural light and reflects the effectiveness of the light; This represents the circadian rhythm regulation function, which outputs correction coefficients at different times according to the needs of the biological clock. The circadian rhythm regulation function Defined as:

[0014] in, Indicates the amplitude parameter of rhythm regulation; Indicates the number of hours in a given day; This indicates the time corresponding to the peak lighting demand; It indicates the rhythm cycle, usually 24 hours.

[0015] Optionally, the multi-objective optimization fitness function is specifically:

[0016] in, This represents the overall adaptability score of the layout scheme; the higher the score, the better the scheme. , , Represents the normalized weight coefficients of the three optimization objectives; Indicators representing the uniformity of lighting; N represents the total power consumption of the system; N represents the total number of lamps. The function representing the installation complexity of the nth lamp; The uniformity of illumination index The calculation formula is:

[0017] Where K represents the total number of illuminance assessment points within the functional area; This represents the calculated illuminance value at the k-th evaluation point; This represents the target illuminance value determined for the k-th evaluation point based on its functional area. Indicates the distance attenuation coefficient; This represents the normalized distance from the k-th evaluation point to the nearest luminaire.

[0018] Optionally, the network topology is determined using the following area-aware routing cost function:

[0019] in, This represents the routing cost between lighting node i and node j; , , Weight coefficients representing physical distance, cross-regional penalty, and semantic difference; This represents the normalized physical distance between node i and node j; This represents a cross-functional area indicator function, which takes the value of 1 when nodes i and j belong to different functional areas, and 0 otherwise. This represents the semantic similarity between node i and node j; The semantic similarity Defined as:

[0020] Where P represents the total number of semantic attribute dimensions; This represents the importance weight of the p-th semantic attribute; This represents the feature value of node i on the p-th semantic attribute; This represents the feature value of node j on the p-th semantic attribute.

[0021] The technical solution of this invention provides a flying robot for automated configuration of intelligent lighting systems. By combining the flying robot's autonomous spatial traversal capabilities with digital twin technology, it achieves a transformation from manual measurement to intelligent automation in lighting system configuration, significantly reducing configuration time costs and manpower input. Through semantic perception-based differential lighting demand reasoning, it improves the adaptability of lighting configuration to the actual needs of different functional areas. Through a virtual-physical linkage pre-verification mechanism, it effectively avoids rework issues after physical installation. Through an adaptive learning feedback mechanism, it enables the configuration system to have continuous optimization capabilities, improving the lighting quality and energy efficiency level during long-term operation. Attached Figure Description

[0022] Figure 1 This is a flowchart of an automated configuration method for an intelligent lighting system performed by a flying robot, according to an embodiment of the present invention.

[0023] Figure 2 This is a schematic diagram of a possible form of a flying robot for automated configuration of an intelligent lighting system, provided by an embodiment of the present invention.

[0024] Figure 3 This is a structural block diagram of a flying robot for automated configuration of an intelligent lighting system, provided in one embodiment of the present invention. Detailed Implementation

[0025] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0026] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0027] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0028] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0029] The following reference Figures 1 to 3 This invention describes a flying robot for the automated configuration of an intelligent lighting system, provided by some embodiments of the invention.

[0030] like Figure 1 As shown, one embodiment of the present invention provides a flying robot for automated configuration of an intelligent lighting system, the configuration method including: The flying robot performs autonomous flight traversal within the target space, simultaneously collecting three-dimensional point cloud data, light intensity distribution data, and infrared thermal imaging data. Based on multimodal sensor fusion technology, it constructs a digital twin model containing spatial geometric features, functional area semantic labels, and real-time light environment features. The functional area semantic labels are automatically labeled as at least one of the building areas such as office area, meeting area, rest area, or passage area through image recognition of object features in the space. It is understood that the semantic labels of the functional areas may also include building areas such as roof areas and exterior wall areas.

[0031] Based on the semantic tags of functional areas in the digital twin model, combined with the preset lighting standard knowledge graph, the target illuminance value, color temperature range and lighting uniformity requirements of each functional area are automatically derived using a rule reasoning engine. The target illuminance value is dynamically adjusted according to the contribution of natural light in the real-time light environment characteristics to generate a differentiated lighting requirement parameter set. Based on the differentiated lighting demand parameter set and the spatial geometric features of the digital twin model, the flying robot uses a multi-objective optimization fitness function to calculate the three-dimensional installation position, installation angle and power configuration scheme of the lamps. The multi-objective optimization fitness function simultaneously optimizes three objective dimensions: lighting uniformity, energy consumption minimization and installation convenience, and outputs a layout scheme containing the spatial coordinates and configuration parameters of each lamp. The flying robot carries an Internet of Things (IoT) communication module and flies sequentially to the predetermined installation positions of each lamp according to the layout scheme. It establishes a temporary connection with the lamp to be installed through a near-field wireless communication protocol and injects configuration information including lamp identifiers, network topology relationships, and semantic tags of the functional areas, so that each lamp can automatically form a mesh IoT topology with regional awareness capabilities. Before the physical lighting fixtures are installed, the layout scheme is simulated in the digital twin model to calculate the deviation between the simulated lighting effect and the set of differentiated lighting requirements parameters. When the deviation exceeds a preset threshold, the iterative optimization of the optimal layout planning step of the lighting fixtures is automatically triggered until the deviation meets the requirements and the final configuration command is output to each lighting fixture. After configuration, the flying robot periodically flies to collect actual lighting effect data within a preset period. It compares and analyzes the actual lighting data with the target illuminance value, extracts configuration deviation features, and feeds them back to the lighting standard knowledge graph for incremental learning. This updates the reasoning rules of the rule reasoning engine to improve the accuracy of subsequent configuration tasks.

[0032] The technical solution adopted in this embodiment combines the spatial autonomous traversal capability of a flying robot with digital twin technology to realize the transformation of lighting system configuration from manual measurement to intelligent automation, significantly reducing configuration time costs and manpower input; through semantic perception-based differential lighting demand reasoning, the adaptability of lighting configuration to the actual needs of different functional areas is improved; through the virtual-physical linkage pre-verification mechanism, rework problems after physical installation are effectively avoided; and through the adaptive learning feedback mechanism, the configuration system has the ability to continuously optimize, improving the lighting quality and energy efficiency level of long-term operation.

[0033] In some possible embodiments of the present invention, the flying robot performs autonomous flight traversal within the target space, simultaneously collecting 3D point cloud data, illumination intensity distribution data, and infrared thermal imaging data. The steps of constructing a digital twin model containing spatial geometric features, functional region semantic labels, and real-time lighting environment features based on multimodal sensor fusion technology include: The flying robot generates a full-coverage flight path based on pre-acquired target space boundary information using a hybrid trajectory mode that combines three-dimensional spiral ascent with horizontal grid scanning. During flight, it dynamically adjusts the trajectory through real-time obstacle detection to ensure that the sensor field of view covers the space without omission. At each sampling point along the full-coverage flight path, the flying robot synchronously activates a lidar, a illuminometer, and an infrared thermal imager through a hardware triggering mechanism to collect three-dimensional point cloud data, light intensity distribution data, and infrared thermal imaging data, respectively. It also uniformly labels each set of collected data with a timestamp and six-degree-of-freedom pose information to establish a spatiotemporal correspondence between multimodal data. After registering the three-dimensional point cloud data with infrared thermal imaging data, the spatial geometric features and thermal radiation features of objects in the space are extracted. The objects such as tables, chairs, conference tables, sofas, and aisle markers are identified through a pre-trained deep learning model. Based on the identified object category combinations and their spatial distribution relationships, the area where the objects are located is automatically labeled as at least one of office area, meeting area, rest area, or aisle area to generate functional area semantic labels. In this step, the 3D point cloud data and infrared thermal imaging data are aligned in spatial coordinate system based on the unified timestamp and six-degree-of-freedom pose information. For each 3D point in the point cloud, its corresponding pixel position in the thermal imaging is located, and the temperature value and thermal radiation intensity of that pixel are assigned to the 3D point, generating fused point cloud data containing spatial coordinates, geometric shape information, and thermal radiation attributes. Simultaneously, data points with registration errors exceeding a threshold are discarded to ensure fusion quality. The fused point cloud data is then input into a pre-trained 3D object detection deep learning model. This model performs multimodal feature extraction based on the geometric shape and thermal radiation features of the point cloud, and identifies and segments independent object instances within the space using a convolutional neural network. For each object instance, its category label is predicted, including office desk, office chair, conference table, conference chair, sofa, etc. The system identifies objects such as coffee tables, passageway signs, green plants, or others, and outputs a list of object recognition results including object category labels, 3D bounding boxes, and confidence scores. Based on the category labels and spatial positions of the 3D bounding boxes of each object instance in the object recognition result list, the system calculates the spatial proximity and category association between objects. Spatially adjacent objects with functional associations are clustered into candidate functional regions. Semantic inference is performed on each candidate functional region according to preset functional region determination rules. For example, a cluster containing multiple desks and chairs is labeled as an office area, a cluster containing a conference table and multiple conference chairs is labeled as a conference area, a cluster containing a sofa and coffee table is labeled as a rest area, and a cluster containing passageway signs or a long, narrow area without main furniture is labeled as a passageway area. Semantic labels for functional regions bound to spatial positions are generated. Multimodal registration and fusion improve the accuracy and robustness of object recognition, especially when there is insufficient lighting or objects have similar appearances, thermal radiation features provide an effective supplement; multimodal feature extraction by deep learning models can make comprehensive judgments by using both geometric and thermal features; spatial clustering and functional inference are based on the combination relationship of objects rather than a single object, making the division of functional areas more in line with the actual spatial usage logic and human cognitive habits.

[0034] Spatially register the light intensity distribution data with the three-dimensional point cloud data, map real-time illuminance values ​​onto the three-dimensional grid nodes of the digital twin model, and simultaneously calculate and store light environment characteristic parameters such as average illuminance, illuminance uniformity, and natural light window position for each functional area. The spatial geometric features, the functional area semantic labels, and the light environment feature parameters are fused according to a unified coordinate system and data structure to construct a three-layer digital twin model containing a geometric layer, a semantic layer, and a light environment layer. An association index between the data in each layer is established to support subsequent lighting requirement reasoning and layout planning.

[0035] In this embodiment, a hybrid trajectory mode ensures complete data acquisition in complex spaces, avoiding blind spots caused by occlusion; hardware synchronous triggering and unified spatiotemporal annotation ensure accurate alignment of multimodal data, improving the accuracy of the fusion model; semantic inference based on object combinations is more consistent with the actual spatial usage logic than simply identifying individual objects, improving the rationality of functional area division; the accurate mapping of light environment features and spatial geometry provides a reliable data foundation for subsequent natural light compensation calculations; the three-layer architecture of the digital twin model realizes the decoupling and association of information in different dimensions, facilitating the extraction of data on demand by different configured algorithms, improving the scalability and computational efficiency of the system.

[0036] In some possible embodiments of the present invention, the steps of automatically deriving the target illuminance value, color temperature range, and lighting uniformity requirements of each functional area based on the functional area semantic tags in the digital twin model, combined with a preset lighting standard knowledge graph, using a rule reasoning engine, and dynamically adjusting the target illuminance value according to the contribution of natural light in the real-time light environment characteristics to generate a differentiated lighting requirement parameter set include: The rule reasoning engine is used to read the semantic tags of each functional area in the digital twin model. The lighting standard node corresponding to each functional area semantic tag is retrieved in the lighting standard knowledge graph. The benchmark target illuminance value, recommended color temperature range and minimum uniformity coefficient stored in the lighting standard node are extracted, and the special lighting rules associated with the functional area are identified. In this step, the rule-based reasoning engine traverses the semantic layer data structure of the digital twin model, extracts the semantic tags, spatial range identifiers, and associated attribute information of each functional area, standardizes the semantic tags to match the node naming conventions in the knowledge graph, identifies composite functional areas and decomposes them into multiple basic functional types, and extracts additional attributes of the functional area, including whether it is near a window, whether it is an open layout, or whether it has special purpose annotations. The rule-based reasoning engine uses the standardized functional area semantic tags as query keys to perform node retrieval in the lighting standard knowledge graph, locates the corresponding lighting standard node, and reads the benchmark target illuminance value and recommended color temperature range upper and lower limits stored in that node. The system uses a minimum uniformity coefficient. If a node contains multiple illuminance level options, the corresponding level parameter is selected based on the spatial range identifier of the functional area. The system also extracts the reference standard source information associated with the node for subsequent traceability. Based on the additional attribute information of the functional area, applicable special lighting rules are queried from the association rule base of the lighting standard node. These rules include glare control rules for window areas, illuminance transition rules for open layouts, light loss compensation rules for high-rise and high-space areas, or specific rules required by industry standards. The identified special lighting rules are then bound to the functional area, and the triggering conditions, adjustment parameters, and priority identifiers defined in the rules are extracted to form a complete lighting standard information package for the functional area. Semantic parsing operations enable semantic integration between the digital twin model and the knowledge graph, ensuring query accuracy. Node retrieval and parameter extraction directly obtain lighting benchmark values ​​that conform to industry standards, ensuring the standardization of the configuration. Special rule recognition enables the system to handle the personalized needs of complex scenarios, improving the refinement and practicality of the configuration.

[0037] The rule reasoning engine determines the floor height, area, and adjacency relationship of each functional area based on the spatial geometric features in the digital twin model, and performs contextual correction on the benchmark target illuminance value in combination with the special lighting rules. Specifically, for areas with floor heights exceeding the conventional value, the target illuminance value is increased to compensate for light loss, and transition lighting requirements are set for adjacent functional areas to avoid visual abrupt changes. In this step, the operation of contextual correction for the benchmark target illuminance value specifically includes: extracting the floor height, horizontal area, length-to-width ratio, and boundary position relationship of each functional area from the geometric layer data of the digital twin model; comparing the extracted geometric parameters with the triggering conditions defined in the special lighting rules item by item; identifying the set of rules that meet the triggering conditions, including light loss compensation rules triggered when the floor height exceeds the standard value, longitudinal illuminance attenuation correction rules triggered in narrow spaces, or transition lighting rules triggered by adjacent different functional areas; for each triggered special lighting rule, calculating the illuminance correction amount according to the calculation logic defined by the rule and the spatial geometric parameters, wherein the light loss compensation rule is based on the difference between the actual floor height and the standard floor height. The process involves determining the percentage increase in illuminance, using a longitudinal attenuation correction rule to determine differentiated illuminance compensation coefficients for different locations based on the aspect ratio of the space, and using a transition lighting rule to determine the width of the transition zone and the illuminance gradient based on the target illuminance difference between adjacent areas. All corrections generated by triggering rules are summarized, and the superposition or mutual exclusion relationships between rules are addressed. Using the baseline target illuminance value as the base value, corrections calculated by each special lighting rule are applied sequentially according to rule priority. For superimposed corrections, an accumulation method is used to update the illuminance value; for mutually exclusive corrections, a higher-priority rule overrides a lower-priority rule. This generates a corrected target illuminance value that integrates all contextual factors, and the applied rule list and correction basis are labeled for this corrected target illuminance value to support subsequent traceability analysis. This step achieves automatic contextual perception through geometric parameter extraction and rule trigger determination, avoiding the subjectivity of manual judgment; rule parameter calculation ensures that the correction amount is based on objective spatial characteristics, improving the scientific nature of the configuration; and the correction value fusion mechanism handles the complex situation of multiple rule superpositions, ensuring that the final illuminance value both meets standards and adapts to actual spatial conditions.

[0038] The location, area, and current illuminance of the natural light window in each functional area are extracted from the real-time light environment characteristics. The natural light contribution of each functional area is calculated. The natural light contribution represents the degree to which natural light meets the lighting needs of the area. Areas significantly affected by natural light and areas that rely on artificial lighting year-round are also identified. In this step, for each functional area, the spatial average of the illuminance measurements from all natural light sampling points within that area is calculated as the representative value of the area's natural light illuminance. This representative value is then compared to the baseline target illuminance value for that functional area to obtain the natural light satisfaction ratio. When this ratio approaches or exceeds a set sufficient threshold, the natural light is considered to be highly satisfied with the area's lighting needs; when the ratio is below a set insufficient threshold, the satisfaction is considered low; and when it falls between the two thresholds, it is considered partially satisfied. The satisfaction level is further adjusted using window area ratio and window orientation information to reflect the stability of the natural light. Using historical illuminance distribution data collected by the flying robot at different times, the natural light illuminance variation curve for each functional area within a complete solar cycle is extracted. The standard deviation, peak-to-valley difference, and effective lighting duration of this curve are then calculated. Regions with large standard deviations and peak-to-valley differences indicate significant fluctuations in natural light and a substantial impact from it. Regions with long effective lighting durations indicate sufficient availability of natural light. The temporal fluctuation characteristics and the natural light satisfaction ratio are comprehensively evaluated to form a comprehensive index of natural light contribution. Based on this comprehensive index, functional regions are classified and labeled. Regions with high natural light satisfaction ratios and moderate temporal fluctuations are labeled as areas significantly affected by natural light and marked with a dynamic natural light compensation strategy. Regions with consistently low natural light satisfaction ratios and small window area proportions are labeled as areas that rely on artificial lighting year-round and marked with a fixed illuminance output strategy. Regions in between are labeled as mixed lighting areas and marked with a time-segmented adjustment strategy. The strategy labeling information and the functional region semantic labels are stored together in the digital twin model for subsequent dynamic compensation and adjustment. In this scheme, the calculation of natural light satisfaction quantifies the actual contribution of natural light to lighting demand through a ratio method, providing a numerical basis for energy-saving optimization; the analysis of temporal fluctuation characteristics reveals the dynamic change law of natural light, enabling the system to predict lighting demand at different times; regional classification and strategy labeling realize the automatic matching of differentiated lighting control strategies, improving energy efficiency and comfort.

[0039] For functional areas where the contribution of natural light is higher than a preset threshold, the target illuminance value after context correction is reduced based on the quantified value of the contribution of natural light to achieve natural light compensation, thereby generating a dynamic target illuminance value for the functional area. The dynamic target illuminance value adapts to changes in natural light. For areas with low contribution of natural light, the target illuminance value after context correction remains unchanged. The target illuminance value, recommended color temperature range, minimum uniformity coefficient, and natural light compensation strategy identifier of each functional area after dynamic compensation adjustment are integrated into the lighting requirement parameters of that functional area. The lighting requirement parameters of all functional areas are summarized to form a differentiated lighting requirement parameter set, and a mapping relationship between functional area identifiers and corresponding lighting requirement parameters is established in the differentiated lighting requirement parameter set.

[0040] This embodiment ensures that lighting requirements meet industry standards and specifications through knowledge graph queries, avoiding the arbitrariness of subjective settings; context-aware rule reasoning enables lighting parameters to adapt to the specificities of the actual space, improving the targeting of the configuration; precise quantification of the contribution of natural light provides a scientific basis for energy-saving optimization; the dynamic compensation and adjustment mechanism realizes the synergy between artificial lighting and natural light, reducing energy consumption while ensuring lighting quality; the generation of differentiated parameter sets provides clear optimization targets for subsequent lighting layout planning, forming a closed loop from demand analysis to scheme design in the entire configuration process, improving the intelligence level and user comfort of the lighting system.

[0041] It is understood that, in some possible embodiments of the present invention, the method for constructing a rule-based reasoning engine includes: collecting national standard documents, industry specification manuals, and expert experience knowledge in the field of lighting engineering; extracting the conditional and conclusion parts of lighting design rules from the standard text using natural language processing technology; converting the extracted rules into structured rule expressions in the form of "IF-THEN-ELSE," each rule expression containing a set of preconditions, reasoning actions, and output conclusions; and labeling each rule with its applicable scenario scope, confidence weight, and rule source identifier, forming a multi-level rule base covering basic lighting standards, scenario-correction rules, and special scenario rules; and constructing a reasoning engine architecture based on forward chaining reasoning, which includes a fact base for storing known facts in the current reasoning process, a rule matcher for identifying triggerable rules, and a conflict resolver for handling multiple rules... The priority selection mechanism for determining when conditions are met implements a reasoning loop mechanism, which includes reading input facts from the fact base, matching candidate rules that meet the preconditions in the rule base, selecting the highest priority rule through a conflict resolver to execute the reasoning action, adding the reasoning conclusion to the fact base as a new fact, and repeating the reasoning loop until the target conclusion is reached or no rules are available. Logical consistency verification is performed on all rules in the rule base to detect contradictory rule pairs, loop reasoning paths, or redundant rules. A logical reasoning verification tool automatically marks conflicting rules and prompts for manual review. A rule version management mechanism records the timestamp and changes for each rule modification. An incremental update interface for rules allows receiving new rules or rule correction suggestions from the adaptive learning feedback step. Newly added or modified rules must pass consistency verification and regression testing before being formally added to the rule base to ensure that they do not disrupt existing reasoning logic. In this embodiment, the formal extraction of domain knowledge transforms scattered standard documents into machine-executable rules, improving the reusability of knowledge and the degree of automation of reasoning; the forward chain reasoning mechanism simulates the thought process of human lighting designers, making the reasoning results interpretable and traceable; rule consistency verification avoids erroneous reasoning caused by logical conflicts; and the dynamic update mechanism enables the reasoning engine to continuously absorb new knowledge and adapt to constantly changing lighting standards and application scenarios, ensuring the long-term effectiveness and intelligence level of the system.

[0042] In some possible embodiments of the present invention, the flying robot, based on the differentiated lighting demand parameter set and the spatial geometric features of the digital twin model, uses a multi-objective optimization fitness function to calculate the three-dimensional installation position, installation angle, and power configuration scheme of the luminaires. The multi-objective optimization fitness function simultaneously optimizes three objective dimensions: lighting uniformity, energy consumption minimization, and installation convenience. The step of outputting a layout scheme containing the spatial coordinates and configuration parameters of each luminaire includes: Based on the spatial geometric features of the digital twin model, the flying robot identifies installable structures such as ceiling planes, beams, and ceiling joists in the target space. On these installable structures, it generates a set of three-dimensional candidate installation points according to preset safety distance rules, and labels each candidate installation point with its structural type, load-bearing capacity, and construction convenience level. At the same time, it excludes infeasible installation points located in ventilation openings, fire-fighting facilities, or structural weak points. Based on the target illuminance value and color temperature range of each functional area in the set of differentiated lighting demand parameters, candidate lamp models that meet the luminous efficacy requirements are selected from the preset lamp library. The effective lighting radius of each lamp model at different installation heights is calculated. The initial number of lamps required for each functional area is estimated by combining the area of ​​each functional area, forming a configuration candidate set that includes lamp models, initial quantities and applicable functional areas. The multi-objective optimization fitness function takes the candidate installation point set and the configuration candidate set as input to construct an optimization model that includes an objective function of lighting uniformity, an objective function of energy consumption minimization, and an objective function of installation convenience. The installation location selection, installation angle, and power configuration of each lamp are adjusted through an iterative search mechanism. During the iteration process, the illuminance distribution of each functional area, the total energy consumption of the system, and the construction complexity score are calculated under each configuration combination. The Pareto front screening method is used to retain multiple sets of non-dominated solutions. Constraint verification is performed on each candidate layout scheme in the multiple sets of non-dominated solutions to check whether it meets the minimum uniformity coefficient, the transition lighting requirements of adjacent functional areas, and the maximum load limit of a single structural point specified in the set of differentiated lighting requirements parameters. Schemes that do not meet the constraints are locally adjusted and corrected, including fine-tuning the position of the lamps, replacing the low-power lamp model, or adding auxiliary lighting points. The optimal layout scheme is selected from the verified and corrected non-dominated solutions according to the preset comprehensive evaluation criteria. The comprehensive evaluation criteria take into account the weighted scores of three objective dimensions (i.e., lighting uniformity, energy consumption minimization, and installation convenience) and the robustness of the scheme. The three-dimensional spatial coordinates, installation angle, power level, lamp model and functional area identifier of each lamp in the optimal layout scheme are extracted to form a layout scheme containing complete configuration parameters and output to the subsequent IoT topology self-organization step.

[0043] This embodiment avoids blind searching through structured generation of candidate installation points, improving optimization efficiency and ensuring the physical feasibility of installation. Intelligent initialization of luminaire type and quantity provides a reasonable starting point for the optimization algorithm, accelerating convergence. The iterative solution mechanism of multi-objective optimization achieves a synergistic balance between lighting performance, energy consumption, and construction costs, avoiding performance degradation in other dimensions caused by single-objective optimization. Constraint verification ensures the practicality and safety of the solution. Pareto front screening and comprehensive evaluation provide diverse options for decision-making, enabling the final solution to meet technical indicators while also considering the convenience of engineering implementation, thus improving the overall quality and efficiency of lighting system deployment.

[0044] In some possible embodiments of the present invention, the flying robot carries an Internet of Things (IoT) communication module and flies sequentially to the predetermined installation positions of each lamp according to the layout scheme. It establishes a temporary connection with the lamps to be installed via a near-field wireless communication protocol and injects configuration information including lamp identifiers, network topology relationships, and semantic tags of the functional areas, enabling each lamp to automatically form a mesh IoT topology with area awareness capabilities. The steps include: The flying robot reads the three-dimensional spatial coordinates and functional area identifiers of all lamps in the layout scheme, generates a configuration access sequence according to the principle of functional area priority and shortest distance within the area, prioritizes the configuration of lamps in the same functional area to quickly establish a communication backbone network within the area, and assigns a globally unique lamp identifier to each lamp in the access sequence. The flying robot flies sequentially to the predetermined installation position of each lamp according to the configured access sequence. When it approaches the target position, it hovers and locates itself by using the spatial geometric features of the digital twin model. It then activates the Internet of Things communication module to send near-field wireless communication signals, establishes a point-to-point temporary connection with the target lamp in the configuration state, and verifies the lamp's identity and communication reliability through a handshake protocol. The flying robot generates a configuration information package for the currently connected lighting fixture. The configuration information package includes the globally unique identifier of the lighting fixture, the semantic label of its functional area, the spatial coordinates and power configuration parameters of the lighting fixture in the layout scheme, as well as a list of neighboring lighting fixtures and communication priorities calculated based on the area-aware routing strategy. The list of neighboring lighting fixtures includes the identifiers of several lighting fixtures that are closest to each other in the same area and the identifiers of gateway lighting fixtures across areas. The configuration information package is injected into the storage unit of the target lighting fixture through the temporary connection. In this step, the flying robot extracts the spatial coordinates and functional area identifier of the currently connected light fixture from the layout scheme, calculates the three-dimensional Euclidean distance between the light fixture and all other light fixtures, selects a preset number of light fixtures within the same functional area as neighbor nodes within the area, identifies other functional areas adjacent to the current functional area, and selects the light fixture closest to the boundary in each adjacent area as a cross-area gateway node. The light fixture identifiers of the neighbor nodes and gateway nodes within the area are aggregated to form a neighbor light fixture list. Based on the area-aware routing strategy, high communication priority is assigned to neighbors within the area, medium communication priority to gateway nodes, and low communication priority to other reachable light fixtures. A globally unique identifier is generated for the currently connected light fixture, using a hierarchical encoding structure including device type code, functional area code, and device serial number to ensure the uniqueness and readability of the identifier. The system extracts configuration parameters such as the 3D spatial coordinates, installation angle, power rating, and dimming range of the luminaire. The luminaire identifier, its functional area semantic label, spatial coordinates, power configuration parameters, a list of neighboring luminaires, and communication priority are encapsulated into a configuration information packet according to a predefined data format. This data format uses a lightweight binary protocol to adapt to the bandwidth limitations of IoT transmission. A configuration information packet write command is sent to the target luminaire via the temporary connection. An encrypted transmission protocol protects the security of the configuration information during wireless transmission. After receiving the configuration information packet, the target luminaire verifies the data integrity check code, parses the configuration information packet, and writes it to the luminaire's non-volatile storage unit to ensure that the configuration is not lost after power failure. After completing the persistent storage of the configuration information, the luminaire returns a configuration success confirmation message to the flying robot. The flying robot records the configuration status of the luminaire and disconnects the temporary connection to prepare for configuring the next luminaire. In this step, intelligent neighbor luminaire selection is based on spatial proximity and functional area correlation, which optimizes the network topology and reduces unnecessary communication overhead; hierarchical coded identifiers and structured encapsulation improve the manageability of configuration information; encrypted transmission and persistent storage ensure the security and reliability of the configuration process, ensuring that luminaires can accurately obtain and save configuration parameters for a long time.

[0045] The lamps with injected configuration information actively search and establish communication links with neighboring lamps according to the list of neighboring lamps. They prioritize establishing high-priority links with neighboring lamps in the same functional area to form a subnet within the area, and then establish low-priority links with cross-regional gateway lamps to achieve inter-regional interconnection. As the flying robot completes the configuration of more lamps, the communication links between each lamp are gradually connected to form a mesh topology covering all functional areas. After the flying robot completes the injection of configuration information for all lamps, it broadcasts a topology construction completion command to the network. Upon receiving the command, each lamp activates its region awareness function based on the semantic tags of the functional areas, enabling lamps within the same functional area to identify each other's region affiliation and form a collaborative control group. Lamps in different functional areas can execute differentiated lighting strategies according to the semantic tags of the functional areas, thereby realizing the self-organization of an intelligent lighting network with region awareness capabilities.

[0046] This embodiment accelerates the establishment of the communication backbone network within the same area through functional area-first configuration sequence planning, improving topology construction efficiency; precise positioning and near-field connectivity ensure the reliability of the configuration process, avoiding misconfiguration and communication interference; the neighbor luminaire list and communication priority preset in the configuration information packet enable luminaires to intelligently select communication paths, reducing the trial and error time of topology self-organization; the dynamic construction mechanism realizes the gradual formation of the topology structure as the configuration progresses, and network construction can begin without waiting for all devices to be ready; the activation of area awareness enables the lighting network to have functional area-level collaborative control capabilities, laying the foundation for subsequent scene-based lighting adjustment and energy efficiency optimization, and significantly improving the flexibility and response speed of the intelligent lighting system.

[0047] In some possible embodiments of the present invention, before the physical lighting fixtures are installed, a lighting simulation of the layout scheme is performed in the digital twin model to calculate the deviation between the simulated lighting effect and the set of differentiated lighting requirement parameters. When the deviation exceeds a preset threshold, the iterative optimization of the optimal layout planning step of the lighting fixtures is automatically triggered until the deviation meets the requirements, and then the final configuration command is output to each lighting fixture. This includes the following steps: In the digital twin model, a corresponding virtual lamp instance is created according to the configuration parameters of each lamp in the layout scheme. The photometric characteristic data of the real physical lamp is loaded for each virtual lamp instance, including the light intensity distribution curve, beam angle, color temperature spectrum and power-luminous efficacy relationship. The virtual lamp instance is placed in the corresponding position of the digital twin model according to the three-dimensional spatial coordinates and installation angle specified in the layout scheme. Based on the spatial geometric features and material reflection properties in the digital twin model, path tracing simulation is performed on the light emitted by each virtual lamp instance to calculate the propagation, reflection, refraction and attenuation process of light in space. The light contribution of all virtual lamps is accumulated at the key evaluation points of each functional area to generate a simulated lighting effect dataset containing the illuminance value, color temperature value and illuminance uniformity of each evaluation point. The simulated lighting effect dataset is compared and analyzed with the differentiated lighting requirement parameter set by functional area. The degree of deviation of each functional area in three dimensions, namely target illuminance value, color temperature range and lighting uniformity requirements, is calculated. The location and amount of the maximum deviation point are extracted for each functional area. The deviation information of all functional areas is summarized to form a comprehensive deviation index. The key influencing factors that cause deviation are identified, including improper lamp position (the deviation value from the preset standard position exceeds the preset deviation threshold), insufficient power configuration (power is lower than the preset power value) or light blockage. When the comprehensive deviation index exceeds the preset threshold, the system automatically extracts the key influencing factors and the maximum deviation point information as optimization constraints and feeds them back to the optimal layout planning step of the lighting fixtures. Under the premise of keeping the configuration of the functional areas that meet the requirements unchanged, the system re-executes the multi-objective optimization fitness function for the functional areas with excessive deviation, adjusts the installation position, installation angle or power configuration of the relevant lighting fixtures, generates an improved layout scheme, and returns to this virtual-real linkage verification step to re-perform the illumination simulation, forming an iterative closed loop of verification-optimization-re-verification. When the comprehensive deviation index meets the preset threshold requirement, the layout scheme is confirmed as the final optimized scheme. The complete configuration parameters of each lamp in the final optimized scheme are extracted, including spatial coordinates, installation angle, power level, dimming curve and start-up sequence. A set of configuration instructions that can be directly executed is generated. The configuration instruction set is distributed to each lamp through the Internet of Things topology, so that each lamp can perform actual lighting output according to the parameters that have been finally verified.

[0048] This embodiment achieves precise mapping of physical lighting fixtures in digital space through virtual lighting fixture instantiation, laying the foundation for high-fidelity simulation. Ray tracing simulation calculations consider the optical physics processes in the real environment, enabling simulation results to accurately predict actual lighting effects and avoiding errors from empirical estimations. Multidimensional deviation quantification provides a comprehensive evaluation from multiple lighting quality dimensions, ensuring the integrity of the verification. An adaptive iterative optimization mechanism identifies and resolves configuration defects before physical installation, avoiding the time and cost waste caused by rework after installation. The verification-optimization closed loop ensures the reliability of the final solution, enabling the output configuration commands to achieve the expected lighting effect on the first attempt, significantly improving the success rate of lighting system deployment and user satisfaction.

[0049] In some possible embodiments of the present invention, the flying robot periodically flies to collect actual lighting effect data within a preset period after configuration, compares and analyzes the actual lighting data with the target illuminance value, extracts configuration deviation features, and feeds them back to the lighting standard knowledge graph for incremental learning, updating the reasoning rules of the rule reasoning engine to improve the accuracy of subsequent configuration tasks. The steps include: After configuration, the flying robot automatically starts according to the preset inspection cycle. It uses the full-coverage flight path generated in the spatial semantic modeling step to conduct regular inspection flights of the target space. It collects illuminance, color temperature and lighting uniformity data under actual operating conditions at key evaluation points in each functional area. At the same time, it records the natural light conditions, lamp working status and space usage at the time of collection, forming actual lighting effect data containing timestamps and environmental context information. The actual lighting effect data is compared item by item with the target illuminance value, color temperature range and lighting uniformity requirements of the corresponding functional areas in the set of differentiated lighting demand parameters. The functional areas and time periods with systematic deviations are identified. The causes of the deviations are analyzed, including insufficient illuminance due to lamp light decay, estimation deviation of natural light compensation strategy, and discrepancies between the actual usage mode of the functional area and the preset or changes in spatial layout. The configuration deviation feature vector containing deviation direction, deviation amplitude, deviation frequency and related environmental factors is extracted. The configuration deviation feature vectors are classified and archived according to the semantic tags of functional areas. For recurring systematic deviations, the relevant lighting standard nodes are located in the lighting standard knowledge graph. The benchmark target illuminance value, recommended color temperature range or natural light compensation coefficient stored in the node are adjusted according to the deviation characteristics. For newly discovered special lighting demand patterns, new rule nodes are created in the knowledge graph and an association relationship with the semantic tags of functional areas is established to achieve continuous enrichment and refinement of the knowledge graph. Based on the incremental updates of the knowledge graph, the rule reasoning engine automatically extracts the updated lighting standard parameters and new rule nodes, reconstructs or corrects the existing context-aware reasoning rules, including adjusting the floor height correction coefficient, optimizing the transition lighting judgment conditions of adjacent areas, or adding special reasoning rules for specific usage patterns, and performs internal consistency verification on the updated reasoning rules to avoid rule conflicts. Before the next configuration task is executed, the updated lighting standard knowledge graph and rule reasoning engine are used to perform regression verification using historical configuration cases to evaluate whether the learning update has improved the accuracy of lighting requirement reasoning and the adaptability of the configuration scheme. For the verified update content, it is formally applied to subsequent configuration tasks, and for the verified update content, it is rolled back and marked as pending further observation to ensure the stability and reliability of the learning feedback mechanism.

[0050] This embodiment establishes a long-term tracking mechanism for configuration effectiveness through regular inspection data collection, enabling the system to promptly identify problems in actual operation. Multi-dimensional deviation feature extraction not only identifies deviation phenomena but also deeply analyzes the root causes of deviations, providing a basis for precise optimization. Incremental updates of the knowledge graph enable adaptive evolution of lighting standards, allowing the system to continuously learn and improve from practice. Optimized updates of inference rules enable the rule inference engine to cope with increasingly complex and diverse lighting scenarios, improving the intelligence level of configuration. The learning effect verification mechanism ensures that each update is a beneficial improvement rather than introducing new problems, guaranteeing continuous improvement in system performance and giving the entire lighting configuration system self-evolution capabilities, becoming increasingly accurate and efficient in long-term use.

[0051] In some possible embodiments of the present invention, in the operation of constructing a digital twin model containing spatial geometric features, functional region semantic labels, and real-time light environment features based on multimodal sensor fusion technology, the digital twin model is constructed through an adaptive weighted fusion algorithm, and the confidence evaluation formula for the adaptive weighted fusion is as follows:

[0052] in, This represents the overall confidence score of the fused digital twin model; M represents the total number of sensor modes involved in the fusion. The basic weighting coefficients for the m-th mode are preset based on the inherent accuracy of the sensor. This represents the dynamic quality factor of the m-th mode under the current environment, calculated in real time using the signal-to-noise ratio; This represents the completeness index of the m-th modality data, reflecting the data coverage rate; Indicates the spatial density influence coefficient; The spatial density of the sampling points is represented by the number of point clouds per unit volume.

[0053] In this embodiment, the confidence evaluation formula adjusts the modal weights in real time through dynamic quality factors, automatically increasing the contribution of infrared thermal imaging in cases of insufficient lighting or occlusion, and increasing the weight of point cloud data in open areas, thus ensuring the robustness of the digital twin model in complex environments. The spatial density correction term avoids over-reliance on sparse sampling areas, improving the overall reliability of the model.

[0054] In some possible embodiments of the present invention, the dynamic adjustment of the target illuminance value based on the contribution of natural light in the real-time light environment characteristics to generate a set of differentiated lighting demand parameters is calculated using the following natural light compensation model:

[0055] in, This represents the target illuminance value of artificial lighting required for this functional area at time t; This represents the reference illuminance value for this functional area, derived from the lighting standards knowledge graph. The natural light utilization efficiency coefficient depends on the light transmission performance and location of the window. This represents the real-time illuminance value of natural light measured at time t; It indicates the angle of incidence of natural light and reflects the effectiveness of the light; This represents the circadian rhythm regulation function, which outputs correction coefficients at different times according to the needs of the biological clock. The circadian rhythm regulation function Defined as:

[0056] in, Indicates the amplitude parameter of rhythm regulation; Indicates the number of hours in a given day; This indicates the time corresponding to the peak lighting demand; It indicates the rhythm cycle, usually 24 hours.

[0057] The natural light compensation model in this embodiment avoids over-compensation for low-angle weak light by correcting the incident angle cosine, thus improving energy efficiency; the circadian rhythm adjustment function makes the lighting configuration conform to the needs of the human body's biological clock, automatically increasing the illuminance in the early morning and evening to combat fatigue, thereby improving the comfort and work efficiency of space users.

[0058] In some possible embodiments of the present invention, the multi-objective optimization fitness function is specifically:

[0059] in, This represents the overall adaptability score of the layout scheme; the higher the score, the better the scheme. , , Represents the normalized weight coefficients of the three optimization objectives; Indicators representing the uniformity of lighting; N represents the total power consumption of the system; N represents the total number of lamps. The function representing the installation complexity of the nth lamp; The uniformity of illumination index The calculation formula is:

[0060] Where K represents the total number of illuminance assessment points within the functional area; This represents the calculated illuminance value at the k-th evaluation point; This represents the target illuminance value determined for the k-th evaluation point based on its functional area. Indicates the distance attenuation coefficient; This represents the normalized distance from the k-th evaluation point to the nearest luminaire.

[0061] In this embodiment, the multi-objective fitness function constrains energy consumption and installation costs while pursuing lighting uniformity, avoiding over-configuration. The distance attenuation term in the lighting uniformity index gives higher tolerance for deviations in areas far from the luminaires, which is consistent with actual visual perception characteristics. This reduces the redundancy in the number of luminaires caused by pursuing extreme uniformity, and achieves a balance between performance and cost.

[0062] In some possible embodiments of the present invention, the network topology is determined by the following area-aware routing cost function:

[0063] in, This represents the routing cost between lighting node i and node j; , , Weight coefficients representing physical distance, cross-regional penalty, and semantic difference; This represents the normalized physical distance between node i and node j; This represents a cross-functional area indicator function, which takes the value of 1 when nodes i and j belong to different functional areas, and 0 otherwise. This represents the semantic similarity between node i and node j; The semantic similarity Defined as:

[0064] Where P represents the total number of semantic attribute dimensions; This represents the importance weight of the p-th semantic attribute; This represents the feature value of node i on the p-th semantic attribute; This represents the feature value of node j on the p-th semantic attribute.

[0065] In this embodiment, the region-aware routing cost function prioritizes the establishment of communication links within the same functional area, reducing the latency and interference of cross-regional communication; semantic similarity calculation prioritizes networking of lamps with the same lighting mode, facilitating the realization of regional-level linkage control; and the cross-regional penalty term ensures the consistency of the topology and spatial functional layout, improving the response speed and collaborative efficiency of lighting scene switching.

[0066] In some possible embodiments of the present invention, the extraction of the configuration deviation features is calculated by the following weighted deviation vector:

[0067] in, R represents the global configuration deviation vector, used for knowledge graph updates; R represents the total number of functional regions. This represents the area proportion weight of the r-th functional area; This represents the measured lighting parameter vector for the r-th functional area, which includes components such as illuminance, color temperature, and uniformity. This represents the target lighting parameter vector for the r-th functional region; The Euclidean norm of the target parameter vector is used for normalization. This represents the time decay function, reflecting the reliability of data at different usage durations; The time decay function Defined as:

[0068] in, This represents the cumulative actual usage time of the r-th functional area; The half-life parameter is used to control the rate at which the weight of short-term data increases. The upper limit for indicating data validity; It represents the aging and decay index, which controls the rate at which long-term data becomes obsolete.

[0069] In this embodiment, the weighted bias vector highlights the bias impact of large areas through area proportion weights, avoiding interference from small area anomalies with global learning; the time decay function rapidly improves data credibility in the early stage of use to accelerate learning convergence, and gradually reduces the weight of outdated data in the later stage of use to track environmental changes, achieving dynamic balance in the learning process; the vectorized bias expression preserves the correlation between multi-dimensional lighting parameters, making knowledge graph updates more accurate and comprehensive.

[0070] Please refer to Figure 2 and Figure 3 For example, the following describes one possible implementation of the present invention. It should be noted that this implementation is for illustrative purposes only and does not constitute a limitation on the present invention.

[0071] The hardware modules of the flying robot include, but are not limited to: image sensors, video processing chips, LBN communication modules, directional antennas, power management modules, ultrasonic sensors, infrared distance sensors, etc. Among them, the image sensors communicate bidirectionally with the video processing chips via a MIPI interface to transmit the acquired image data in real time; the video processing chips connect to the flight controller via a UART interface to output information such as device position coordinates and recognition results, providing a basis for flight control; the LBN communication module interconnects with the flight controller via a UART interface, receiving power adjustment and networking commands, and providing feedback on the device's communication status; the directional antenna connects to the LBN communication module via an SMA interface, with the antenna pointing vertically upwards; the power management module supplies power to components such as the flight controller (5V), video processing unit (12V), and communication module (3.3V) through a DC-DC conversion circuit; the ultrasonic sensors and infrared distance sensors are directly connected to the ADC interface of the flight controller to provide environmental obstacle and distance data.

[0072] The software modules possessed by flying robots include, but are not limited to: APP interaction module: runs on Android / iOS / Harmony systems, supports room information configuration (room number, number of devices), flight robot control (start / pause / return), real-time display of network configuration progress, and device status query; Flight control module: integrates path planning algorithm (based on AI algorithm, combined with room floor plan to generate optimal flight path), obstacle avoidance algorithm (real-time analysis of ultrasonic data, dynamic adjustment of flight trajectory), and fixed-point hovering algorithm (integrating GPS and inertial navigation data to ensure positioning accuracy within 10cm). AI video recognition module: Based on the YOLOv8 lightweight model, the training set includes the shape features of common smart lighting devices (round / square lampshade, mounting bracket, brand logo), and supports device recognition in indoor low light environments (recognition accuracy ≥98%). LBN communication adapter module: compatible with RW-LBN protocol, supports network command encapsulation (device ID binding, room ID allocation), and dynamic adjustment of transmission power (based on infrared distance sensor data, the power is adjusted to -10dBm when the distance to the device is 10cm to avoid interference). Status feedback module: Collects data such as the flying robot's battery level, device identification results, and network success rate in real time, and feeds it back to the mobile APP via the LBN protocol.

[0073] The core algorithms used by flying robots include, but are not limited to: Equipment recognition algorithm: After the image sensor acquires images of the roof area, the video processing chip performs grayscale and noise reduction on the images, extracts equipment features through the YOLOv8 model, matches them with the built-in equipment feature library, and determines the equipment location coordinates (establishing a two-dimensional coordinate system with the upper left corner of the room as the origin). Fixed-point flight algorithm: The flight controller plans a straight path from the current position to 10cm directly below the device based on the device coordinates output by the AI ​​recognition module and combined with ultrasonic obstacle avoidance data. The hovering accuracy is ±10cm by adjusting the speed of the brushless motor through PID. Network communication algorithm: After approaching the device, the LBN communication module adjusts the directional antenna to align with the device, sends a network start command, and after the device responds, writes the room ID and device number preset by the mobile APP into the device storage unit, and sends a confirmation command after the binding is completed.

[0074] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0075] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0076] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0077] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0078] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0079] If the integrated units described above are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0080] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0081] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

[0082] While the present invention has been disclosed above, it is not limited thereto. Any person skilled in the art can easily conceive of variations or substitutions without departing from the spirit and scope of the present invention, and various modifications and alterations can be made, including combinations of the different functions and implementation steps described above, as well as software and hardware implementation methods, all of which are within the protection scope of the present invention.

Claims

1. A flying robot for automated configuration of an intelligent lighting system, characterized in that, Configuration methods include: The flying robot performs autonomous flight traversal within the target space, simultaneously collecting three-dimensional point cloud data, light intensity distribution data, and infrared thermal imaging data. Based on multimodal sensor fusion technology, it constructs a digital twin model that includes spatial geometric features, functional area semantic labels, and real-time light environment features. The functional area semantic labels are automatically labeled as at least one of office area, meeting area, rest area, or passage area through image recognition of object features within the space. Based on the semantic tags of functional areas in the digital twin model, combined with the preset lighting standard knowledge graph, the target illuminance value, color temperature range and lighting uniformity requirements of each functional area are automatically derived using a rule reasoning engine. The target illuminance value is dynamically adjusted according to the contribution of natural light in the real-time light environment characteristics to generate a differentiated lighting requirement parameter set. Based on the differentiated lighting demand parameter set and the spatial geometric features of the digital twin model, the flying robot uses a multi-objective optimization fitness function to calculate the three-dimensional installation position, installation angle and power configuration scheme of the lamps. The multi-objective optimization fitness function simultaneously optimizes three objective dimensions: lighting uniformity, energy consumption minimization and installation convenience, and outputs a layout scheme containing the spatial coordinates and configuration parameters of each lamp. The flying robot carries an Internet of Things (IoT) communication module and flies sequentially to the predetermined installation positions of each lamp according to the layout scheme. It establishes a temporary connection with the lamp to be installed through a near-field wireless communication protocol and injects configuration information including lamp identifiers, network topology relationships, and semantic tags of the functional areas, so that each lamp can automatically form a mesh IoT topology with regional awareness capabilities. Before the physical lighting fixtures are installed, the layout scheme is simulated in the digital twin model to calculate the deviation between the simulated lighting effect and the set of differentiated lighting requirements parameters. When the deviation exceeds a preset threshold, the iterative optimization of the optimal layout planning step of the lighting fixtures is automatically triggered until the deviation meets the requirements and the final configuration command is output to each lighting fixture. After configuration, the flying robot periodically flies to collect actual lighting effect data within a preset period. It compares and analyzes the actual lighting data with the target illuminance value, extracts configuration deviation features, and feeds them back to the lighting standard knowledge graph for incremental learning, thereby updating the reasoning rules of the rule reasoning engine.

2. The flying robot for automated configuration of an intelligent lighting system according to claim 1, characterized in that, The steps involved in constructing a digital twin model of a flying robot that autonomously traverses the target space, simultaneously collecting 3D point cloud data, illumination intensity distribution data, and infrared thermal imaging data, and based on multimodal sensor fusion technology, include: The flying robot generates a full-coverage flight path based on pre-acquired target space boundary information using a hybrid trajectory mode that combines three-dimensional spiral ascent with horizontal grid scanning. During flight, it dynamically adjusts the trajectory through real-time obstacle detection to ensure that the sensor field of view covers the space without omission. At each sampling point along the full-coverage flight path, the flying robot synchronously activates a lidar, a illuminometer, and an infrared thermal imager through a hardware triggering mechanism to collect three-dimensional point cloud data, light intensity distribution data, and infrared thermal imaging data, respectively. It also uniformly labels each set of collected data with a timestamp and six-degree-of-freedom pose information to establish a spatiotemporal correspondence between multimodal data. After registering the three-dimensional point cloud data with infrared thermal imaging data, the spatial geometric features and thermal radiation features of objects in the space are extracted. The objects such as tables, chairs, conference tables, sofas, and aisle markers are identified through a pre-trained deep learning model. Based on the identified object category combinations and their spatial distribution relationships, the area where the objects are located is automatically labeled as at least one of office area, meeting area, rest area, or aisle area to generate functional area semantic labels. Spatially register the light intensity distribution data with the three-dimensional point cloud data, map real-time illuminance values ​​onto the three-dimensional grid nodes of the digital twin model, and simultaneously calculate and store light environment characteristic parameters such as average illuminance, illuminance uniformity, and natural light window position for each functional area. The spatial geometric features, the functional area semantic labels, and the light environment feature parameters are fused according to a unified coordinate system and data structure to construct a three-layer digital twin model containing a geometric layer, a semantic layer, and a light environment layer. An association index between the data in each layer is established to support subsequent lighting requirement reasoning and layout planning.

3. The flying robot for automated configuration of an intelligent lighting system according to claim 2, characterized in that, Based on the functional area semantic tags in the digital twin model, combined with a preset lighting standard knowledge graph, the steps of automatically deriving the target illuminance value, color temperature range, and lighting uniformity requirements for each functional area using a rule-based reasoning engine, and dynamically adjusting the target illuminance value according to the contribution of natural light in the real-time lighting environment features to generate a differentiated lighting requirement parameter set include: The rule reasoning engine is used to read the semantic tags of each functional area in the digital twin model. The lighting standard node corresponding to each functional area semantic tag is retrieved in the lighting standard knowledge graph. The benchmark target illuminance value, recommended color temperature range and minimum uniformity coefficient stored in the lighting standard node are extracted, and the special lighting rules associated with the functional area are identified. The rule reasoning engine determines the floor height, area, and adjacency relationship of each functional area based on the spatial geometric features in the digital twin model, and performs contextual correction on the benchmark target illuminance value in combination with the special lighting rules. Specifically, for areas with floor heights exceeding the conventional value, the target illuminance value is increased to compensate for light loss, and transition lighting requirements are set for adjacent functional areas to avoid visual abrupt changes. The location, area, and current illuminance of the natural light window in each functional area are extracted from the real-time light environment characteristics. The natural light contribution of each functional area is calculated. The natural light contribution represents the degree to which natural light meets the lighting needs of the area. Areas significantly affected by natural light and areas that rely on artificial lighting year-round are also identified. For functional areas where the contribution of natural light is higher than a preset threshold, the target illuminance value after context correction is reduced based on the quantified value of the contribution of natural light to achieve natural light compensation, thereby generating a dynamic target illuminance value for the functional area. The dynamic target illuminance value adapts to changes in natural light. For areas with low contribution of natural light, the target illuminance value after context correction remains unchanged. The target illuminance value, recommended color temperature range, minimum uniformity coefficient, and natural light compensation strategy identifier of each functional area after dynamic compensation adjustment are integrated into the lighting requirement parameters of that functional area. The lighting requirement parameters of all functional areas are summarized to form a differentiated lighting requirement parameter set, and a mapping relationship between functional area identifiers and corresponding lighting requirement parameters is established in the differentiated lighting requirement parameter set.

4. The flying robot for automated configuration of an intelligent lighting system according to claim 3, characterized in that, The flying robot, based on the differentiated lighting demand parameter set and the spatial geometric features of the digital twin model, uses a multi-objective optimization fitness function to calculate the three-dimensional installation position, installation angle, and power configuration scheme of the luminaires. This multi-objective optimization fitness function simultaneously optimizes three objective dimensions: lighting uniformity, energy consumption minimization, and installation convenience. The steps to output a layout scheme containing the spatial coordinates and configuration parameters of each luminaire include: Based on the spatial geometric features of the digital twin model, the flying robot identifies installable structures such as ceiling planes, beams, and ceiling joists in the target space. On these installable structures, it generates a set of three-dimensional candidate installation points according to preset safety distance rules, and labels each candidate installation point with its structural type, load-bearing capacity, and construction convenience level. At the same time, it excludes infeasible installation points located in ventilation openings, fire-fighting facilities, or structural weak points. Based on the target illuminance value and color temperature range of each functional area in the set of differentiated lighting demand parameters, candidate lamp models that meet the luminous efficacy requirements are selected from the preset lamp library. The effective lighting radius of each lamp model at different installation heights is calculated. The initial number of lamps required for each functional area is estimated by combining the area of ​​each functional area, forming a configuration candidate set that includes lamp models, initial quantities and applicable functional areas. The multi-objective optimization fitness function takes the candidate installation point set and the configuration candidate set as input to construct an optimization model that includes an objective function of lighting uniformity, an objective function of energy consumption minimization, and an objective function of installation convenience. The installation location selection, installation angle, and power configuration of each lamp are adjusted through an iterative search mechanism. During the iteration process, the illuminance distribution of each functional area, the total energy consumption of the system, and the construction complexity score are calculated under each configuration combination. The Pareto front screening method is used to retain multiple sets of non-dominated solutions. Constraint verification is performed on each candidate layout scheme in the multiple sets of non-dominated solutions to check whether it meets the minimum uniformity coefficient, the transition lighting requirements of adjacent functional areas, and the maximum load limit of a single structural point specified in the set of differentiated lighting requirements parameters. Schemes that do not meet the constraints are locally adjusted and corrected, including fine-tuning the position of the lamps, replacing the low-power lamp model, or adding auxiliary lighting points. The optimal layout scheme is selected from the verified and corrected non-dominated solutions according to the preset comprehensive evaluation criteria. The comprehensive evaluation criteria take into account the weighted scores of three objective dimensions and the robustness of the scheme. The three-dimensional spatial coordinates, installation angle, power level, lamp model and functional area identifier of each lamp in the optimal layout scheme are extracted to form a layout scheme containing complete configuration parameters and output to the subsequent IoT topology self-organization step.

5. The flying robot for automated configuration of an intelligent lighting system according to claim 4, characterized in that, The flying robot, carrying an IoT communication module, flies sequentially to the predetermined installation positions of each lamp according to the layout scheme. It establishes temporary connections with the lamps to be installed via a near-field wireless communication protocol and injects configuration information including lamp identifiers, network topology relationships, and semantic tags for the functional areas. This enables the lamps to automatically form a mesh IoT topology with area awareness capabilities. The steps include: The flying robot reads the three-dimensional spatial coordinates and functional area identifiers of all lamps in the layout scheme, generates a configuration access sequence according to the principle of functional area priority and shortest distance within the area, prioritizes the configuration of lamps in the same functional area to quickly establish a communication backbone network within the area, and assigns a globally unique lamp identifier to each lamp in the access sequence. The flying robot flies sequentially to the predetermined installation position of each lamp according to the configured access sequence. When it approaches the target position, it hovers and locates itself by using the spatial geometric features of the digital twin model. It then activates the Internet of Things communication module to send near-field wireless communication signals, establishes a point-to-point temporary connection with the target lamp in the configuration state, and verifies the lamp's identity and communication reliability through a handshake protocol. The flying robot generates a configuration information package for the currently connected lighting fixture. The configuration information package includes the globally unique identifier of the lighting fixture, the semantic label of its functional area, the spatial coordinates and power configuration parameters of the lighting fixture in the layout scheme, as well as a list of neighboring lighting fixtures and communication priorities calculated based on the area-aware routing strategy. The list of neighboring lighting fixtures includes the identifiers of several lighting fixtures that are closest to each other in the same area and the identifiers of gateway lighting fixtures across areas. The configuration information package is injected into the storage unit of the target lighting fixture through the temporary connection. The lamps with injected configuration information actively search and establish communication links with neighboring lamps according to the list of neighboring lamps. They prioritize establishing high-priority links with neighboring lamps in the same functional area to form a subnet within the area, and then establish low-priority links with cross-regional gateway lamps to achieve inter-regional interconnection. As the flying robot completes the configuration of more lamps, the communication links between each lamp are gradually connected to form a mesh topology covering all functional areas. After the flying robot completes the injection of configuration information for all lamps, it broadcasts a topology construction completion command to the network. Upon receiving the command, each lamp activates its region awareness function based on the semantic tags of the functional areas, enabling lamps within the same functional area to identify each other's region affiliation and form a collaborative control group. Lamps in different functional areas can execute differentiated lighting strategies according to the semantic tags of the functional areas, thereby realizing the self-organization of an intelligent lighting network with region awareness capabilities.

6. The flying robot for automated configuration of an intelligent lighting system according to claim 5, characterized in that, Before the physical lighting fixtures are installed, a lighting simulation of the layout scheme is performed in the digital twin model to calculate the deviation between the simulated lighting effect and the set of differentiated lighting requirement parameters. When the deviation exceeds a preset threshold, the iterative optimization of the optimal lighting fixture layout planning step is automatically triggered until the deviation meets the requirements, at which point the final configuration command is output to each lighting fixture. This includes the following steps: In the digital twin model, a corresponding virtual lamp instance is created according to the configuration parameters of each lamp in the layout scheme. The photometric characteristic data of the real physical lamp is loaded for each virtual lamp instance, including the light intensity distribution curve, beam angle, color temperature spectrum and power-luminous efficacy relationship. The virtual lamp instance is placed in the corresponding position of the digital twin model according to the three-dimensional spatial coordinates and installation angle specified in the layout scheme. Based on the spatial geometric features and material reflection properties in the digital twin model, path tracing simulation is performed on the light emitted by each virtual lamp instance to calculate the propagation, reflection, refraction and attenuation process of light in space. The light contribution of all virtual lamps is accumulated at the key evaluation points of each functional area to generate a simulated lighting effect dataset containing the illuminance value, color temperature value and illuminance uniformity of each evaluation point. The simulated lighting effect dataset is compared and analyzed with the differentiated lighting requirement parameter set by functional area. The degree of deviation of each functional area in three dimensions, namely target illuminance value, color temperature range and lighting uniformity requirement, is calculated. The location and amount of the maximum deviation point are extracted for each functional area. The deviation information of all functional areas is summarized to form a comprehensive deviation index. The key influencing factors that cause deviation are identified, including improper lamp position, insufficient power configuration or light blockage. When the comprehensive deviation index exceeds the preset threshold, the system automatically extracts the key influencing factors and the maximum deviation point information as optimization constraints and feeds them back to the optimal layout planning step of the lighting fixtures. Under the premise of keeping the configuration of the functional areas that meet the requirements unchanged, the system re-executes the multi-objective optimization fitness function for the functional areas with excessive deviation, adjusts the installation position, installation angle or power configuration of the relevant lighting fixtures, generates an improved layout scheme, and returns to this virtual-real linkage verification step to re-perform the illumination simulation, forming an iterative closed loop of verification-optimization-re-verification. When the comprehensive deviation index meets the preset threshold requirement, the layout scheme is confirmed as the final optimized scheme. The complete configuration parameters of each lamp in the final optimized scheme are extracted, including spatial coordinates, installation angle, power level, dimming curve and start-up sequence. A set of configuration instructions that can be directly executed is generated. The configuration instruction set is distributed to each lamp through the Internet of Things topology, so that each lamp can perform actual lighting output according to the parameters that have been finally verified.

7. The flying robot for automated configuration of an intelligent lighting system according to claim 6, characterized in that, In the process of constructing a digital twin model based on multimodal sensor fusion technology, which includes spatial geometric features, functional region semantic labels, and real-time light environment features, the digital twin model is constructed through an adaptive weighted fusion algorithm. The confidence evaluation formula for the adaptive weighted fusion is as follows: in, This represents the overall confidence score of the fused digital twin model; M represents the total number of sensor modes involved in the fusion. The basic weighting coefficients for the m-th mode are preset based on the inherent accuracy of the sensor. This represents the dynamic quality factor of the m-th mode under the current environment, calculated in real time using the signal-to-noise ratio; This represents the completeness index of the m-th modality data, reflecting the data coverage rate; Indicates the spatial density influence coefficient; The spatial density of the sampling points is represented by the number of point clouds per unit volume.

8. The flying robot for automated configuration of an intelligent lighting system according to claim 7, characterized in that, The target illuminance value is dynamically adjusted based on the contribution of natural light in the real-time light environment characteristics to generate a set of differentiated lighting demand parameters. The dynamic adjustment of the target illuminance value is calculated using the following natural light compensation model: in, This represents the target illuminance value of artificial lighting required for this functional area at time t; This represents the reference illuminance value for this functional area, derived from the lighting standards knowledge graph. The natural light utilization efficiency coefficient depends on the light transmission performance and location of the window. This represents the real-time illuminance value of natural light measured at time t; It indicates the angle of incidence of natural light and reflects the effectiveness of the light; This represents the circadian rhythm regulation function, which outputs correction coefficients at different times according to the needs of the biological clock. The circadian rhythm regulation function Defined as: in, Indicates the amplitude parameter of rhythm regulation; Indicates the number of hours in a given day; This indicates the time corresponding to the peak lighting demand; It indicates the rhythm cycle, usually 24 hours.

9. The flying robot for automated configuration of an intelligent lighting system according to claim 8, characterized in that, The multi-objective optimization fitness function is specifically: in, This represents the overall adaptability score of the layout scheme; the higher the score, the better the scheme. , , Represents the normalized weight coefficients of the three optimization objectives; Indicators representing the uniformity of lighting; N represents the total power consumption of the system; N represents the total number of lamps. The function representing the installation complexity of the nth lamp; The uniformity of illumination index The calculation formula is: Where K represents the total number of illuminance assessment points within the functional area; This represents the calculated illuminance value at the k-th evaluation point; This represents the target illuminance value determined for the k-th evaluation point based on its functional area. Indicates the distance attenuation coefficient; This represents the normalized distance from the k-th evaluation point to the nearest luminaire.

10. The flying robot for automated configuration of an intelligent lighting system according to claim 9, characterized in that, The network topology is determined using the following area-aware routing cost function: in, This represents the routing cost between lighting node i and node j; , , Weight coefficients representing physical distance, cross-regional penalty, and semantic difference; This represents the normalized physical distance between node i and node j; This represents a cross-functional area indicator function, which takes the value of 1 when nodes i and j belong to different functional areas, and 0 otherwise. This represents the semantic similarity between node i and node j; The semantic similarity Defined as: Where P represents the total number of semantic attribute dimensions; This represents the importance weight of the p-th semantic attribute; This represents the feature value of node i on the p-th semantic attribute; This represents the feature value of node j on the p-th semantic attribute.

Images