Adjustable smart home antenna

Abstract

The invention relates to an adjustable smart home antenna, and belongs to the technical field of home antennas. Based on electromagnetic parameters, a three-dimensional radiation pattern of the antenna is imported into a simulation environment, the orientation angle of the antenna is discretized in a digital twinborn body, and light is tracked and simulated; therefore, the full-space coverage quality under all-orientation configuration is comprehensively evaluated, the coverage uniformity and the peak coverage quality serve as multi-target optimization indexes, the optimal orientation and the estimated installation position are calculated through the genetic algorithm, historical movement behavior data information of the user is collected, and the user experience is improved. And the installation position or orientation is dynamically optimized based on the collected historical movement behavior data information of the user. According to the invention, the propagation performance of electromagnetic waves in a specific environment can be simulated with high precision, and complex propagation mechanisms such as wall reflection, diffraction, scattering and the like are accurately considered. Dynamic changes of furniture placement, personnel position and the like in a physical environment are synchronized in real time by the digital twinborn body, and recalculation is automatically triggered when the environment is changed.

Description

Technical Field

[0001] This invention relates to the field of home antenna technology, and more particularly to an adjustable smart home antenna. Background Technology

[0002] As a core radio frequency component in wireless communication systems, enabling electromagnetic signal transmission and reception, the performance of home antennas directly determines the communication distance, data transmission rate, anti-interference capability, and overall user experience of smart home devices. However, unlike traditional communication equipment, smart home devices place higher design demands on antennas. On the one hand, the trend towards miniaturization and integration of devices requires antennas to achieve efficient wireless transmission and reception within limited physical space; on the other hand, with the rapid increase in the number of wireless devices in the home, problems such as concurrent communication of multiple devices, frequency band congestion, and electromagnetic interference are becoming increasingly prominent, placing higher demands on the multi-band compatibility and anti-interference capability of antennas.

[0003] A fundamental challenge in adjusting home antennas is that the adjustment effect is affected by the complex electromagnetic environment of the home space, including the reflection / transmission characteristics of walls, the scattering effect of furniture, and the coherent superposition of multipath propagation. Current technologies rely solely on RSSI changes to guide adjustment, which is a "blind adjustment" method. It cannot obtain the signal distribution characteristics of the entire space, resulting in low adjustment efficiency and difficulty in achieving global optimization. Summary of the Invention

[0004] This invention overcomes the shortcomings of the prior art and provides an adjustable smart home antenna.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: The first aspect of this invention provides an adjustable smart home antenna, comprising: The home environment digital modeling module is used to acquire three-dimensional home models of the home environment, including wall location, wall material information, transmission loss, and furniture size and location, and to construct a digital model of the home environment. An electromagnetic parameter calibration module is used to mark the electromagnetic parameters of each wall in the digital model of the home environment. The ray tracing simulation module is used to import the three-dimensional radiation pattern of the antenna into the simulation environment based on the electromagnetic parameters, discretize the antenna orientation angle in the digital twin, and perform ray tracing and simulation. The optimal orientation and installation location pre-calculation module comprehensively evaluates the full-space coverage quality under all-orientation configurations, using coverage uniformity and peak coverage quality as multi-objective optimization indicators, and calculates the optimal orientation using a genetic algorithm. The installation location determination module is used to collect users' historical movement behavior data and dynamically optimize the installation location or orientation based on the collected historical movement behavior data.

[0006] Furthermore, within the adjustable smart home antenna, a three-dimensional home model is acquired, including information on wall location, wall material, transmission loss, and the dimensions and location of furniture, to construct a digital model of the home environment. This includes: Guide users to use the depth sensor built into the mobile terminal to scan the indoor space, or use the router's multi-antenna array to transmit detection signals and analyze the echoes to analyze the signal's wall-penetrating attenuation characteristics. Based on the wall penetration attenuation characteristics of the signal, the thickness information of the wall is estimated. At the same time, data including the wall thickness, wall location, wall material information, and transmission loss characteristics are generated. Obtain the dimensions and location information of the home furnishings, and construct a three-dimensional home model based on the dimensions and location information of the home furnishings; A home environment model is constructed based on the wall thickness information, wall location, dielectric constant, and transmission loss characteristic data. The home environment model and the three-dimensional home model with the size and location of the home are combined to form a digital home environment model.

[0007] Furthermore, in the adjustable smart home antenna, the electromagnetic parameters of each wall are marked in the digital model of the home environment, specifically as follows: Obtain the dielectric constant and conductivity corresponding to different material types, construct a material database, and input the dielectric constant and conductivity corresponding to different material types into the material database for storage; Obtain the material information of each wall in the digital model of the home environment, and call the data in the material database to obtain the dielectric constant and conductivity corresponding to the material type of each wall in the digital model of the home environment; The electromagnetic parameters of each wall in the digital home environment model are obtained based on the dielectric constant and conductivity corresponding to the material type of each wall. For unknown materials, the electromagnetic parameters are calculated by measuring the actual signal strength at reference points at different locations and using a reverse optimization algorithm. The electromagnetic parameters of each wall are then marked in the digital model of the home environment.

[0008] Furthermore, in the adjustable smart home antenna, the three-dimensional radiation pattern of the antenna is imported into the simulation environment based on the electromagnetic parameters. The antenna orientation angle is discretized in the digital twin, and ray tracing and simulation are performed, specifically as follows: Initialize the orientation of the home antenna, import the three-dimensional radiation pattern of the home antenna into the simulation environment, discretize the antenna orientation angle in the digital model of the home environment, configure each orientation based on the electromagnetic parameters, and emit a large amount of light from the antenna position; Based on the material properties of the wall, the absorbed light data is estimated, and the energy of the transmitted light is calculated according to the absorbed light data and the characteristics of the emitted light. The energy distribution of each light ray after multiple reflections, transmissions and diffractions is tracked to reach different areas of the room. The full-space path gain, received signal strength, and signal-to-interference-plus-noise ratio distribution are calculated based on the energy distribution of each region.

[0009] Furthermore, in adjustable smart home antennas, the overall spatial coverage quality under omnidirectional configuration is comprehensively evaluated. Using coverage uniformity and peak coverage quality as multi-objective optimization indicators, a genetic algorithm is used to calculate the optimal orientation and estimated installation location. Specifically: Based on the energy distribution of each region, the full-space path gain, received signal strength and signal-to-interference-plus-noise ratio distribution are calculated, and the full-space coverage quality under the all-directional configuration is comprehensively evaluated to obtain coverage uniformity and peak coverage quality. A genetic algorithm is introduced, and the number of generations is set based on the genetic algorithm. Coverage uniformity and peak coverage quality evaluation indicators are set to determine whether the coverage uniformity and peak coverage quality meet the coverage uniformity and peak coverage quality evaluation indicators. When the coverage uniformity and peak coverage quality meet the evaluation indicators for coverage uniformity and peak coverage quality, the antenna shall be configured according to its current installation location and orientation. When the coverage uniformity and peak coverage quality do not meet the evaluation criteria for coverage uniformity and peak coverage quality, the current orientation is reset until the coverage uniformity and peak coverage quality meet the evaluation criteria for coverage uniformity and peak coverage quality.

[0010] Furthermore, the adjustable smart home antenna collects historical user movement behavior data, and dynamically optimizes the installation location or orientation based on this data. Specifically, this includes: With user authorization, the system continuously collects the user's three-dimensional location coordinates in the home to form time-series trajectory data. The collection frequency is no less than a preset frequency threshold. The raw trajectory data is then subjected to Kalman filtering for noise reduction, stop point detection, and trajectory segmentation. Among them, speeds below the threshold and durations exceeding the threshold are judged as dwellings. Statistical features are extracted from the segmented trajectory fragments, including the frequency of visits to each room and the distribution of dwell time. The spatial geometry of preferred areas and commonly used paths in different time periods is also analyzed, and the trajectories are clustered using a hierarchical clustering algorithm to identify several typical movement patterns. The LSTM prediction model is trained based on historical trajectory data. It takes the user location sequence of the past N time steps as input, outputs the location prediction of the next K time steps, and dynamically optimizes the orientation of home antennas based on the location of the next K time steps.

[0011] Furthermore, in adjustable smart home antennas, the orientation of the antenna is dynamically optimized based on its position over the next K time steps, specifically as follows: After the prediction model outputs the future location sequence, the deviation angle between the current orientation of the home antenna and the direction of the predicted location is calculated. If the deviation angle exceeds the preset threshold, and the distance between the predicted location and the antenna location does not exceed the effective coverage range of the antenna; Trigger pre-alignment adjustment. The timing of the pre-alignment adjustment is determined in advance based on the predicted arrival time of the location. The adjustment is initiated a preset number of seconds before the user is expected to arrive, ensuring that the antenna completes its turn exactly when the user arrives at the target location. Meanwhile, the prediction model outputs the position prediction variance for each future time step as a confidence index. When the prediction confidence is lower than the preset value, it automatically downgrades to reactive tracking mode and performs weighted fusion of predictive pre-alignment and reactive tracking.

[0012] Furthermore, adjustable smart home antennas also include: The frequency of visits and duration of stay in each room are statistically analyzed, and the frequency characteristics of each location node at a predetermined time are statistically analyzed based on the frequency of visits and duration of stay in each room. The location nodes with frequency characteristics greater than a preset frequency characteristic are identified, and it is determined whether all the location nodes with frequency characteristics greater than the preset frequency characteristic are within the effective coverage range of the antenna. When all location nodes with frequency characteristics greater than the preset frequency characteristics are within the effective coverage range of the antenna, the antenna is configured according to its current installation location and displayed in a preset manner. When the location nodes whose frequency characteristics are greater than the preset frequency characteristics are not all within the effective coverage range of the antenna, the installation position of the current home antenna is reconfigured until all location nodes whose frequency characteristics are greater than the preset frequency characteristics are within the effective coverage range of the antenna.

[0013] This invention addresses the shortcomings of the prior art and has the following beneficial effects: This invention constructs a digital model of the home environment by acquiring a 3D home model containing information on wall locations, wall materials, transmission loss, and furniture dimensions and locations. The electromagnetic parameters of each wall are then annotated in this digital model. Based on these electromagnetic parameters, the 3D radiation pattern of an antenna is imported into a simulation environment. The antenna orientation angle is discretized in the digital twin, and ray tracing and simulation are performed. This allows for a comprehensive evaluation of the overall coverage quality under all-orientation configurations. Coverage uniformity and peak coverage quality are used as multi-objective optimization indicators. A genetic algorithm is used to calculate the optimal orientation and estimated installation location. Furthermore, historical user movement data is collected, and the installation location or orientation is dynamically optimized based on this data. This invention can simulate the propagation performance of electromagnetic waves in a specific environment with high precision, accurately considering complex propagation mechanisms such as wall reflection, diffraction, and scattering. The digital twin synchronizes with real-time changes in the physical environment, such as furniture placement and personnel positions, and automatically triggers recalculation when the environment changes. By applying digital twin and ray tracing technologies to the field of home antenna direction adjustment, this solution differs from the "trial and error" adjustment of existing technologies. Through physically precise electromagnetic simulation, the optimal direction can be predicted "before adjustment," fundamentally solving the problems of blind and inefficient direction adjustment. Combined with machine learning to accelerate the prediction model, millisecond-level response to environmental changes is achieved. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained from these drawings without creative effort.

[0015] Figure 1 A schematic diagram of the overall module of the adjustable smart home antenna is shown. Detailed Implementation

[0016] 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.

[0017] 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.

[0018] like Figure 1As shown, the first aspect of the present invention provides an adjustable smart home antenna, comprising: The home environment digital modeling module is used to acquire three-dimensional home models of the home environment, including wall location, wall material information, transmission loss, and furniture size and location, and to construct a digital model of the home environment. Electromagnetic parameter calibration module, used to annotate the electromagnetic parameters of each wall in the digital model of the home environment; The ray tracing simulation module is used to import the three-dimensional radiation pattern of the antenna into the simulation environment based on electromagnetic parameters, discretize the antenna orientation angle in the digital twin, and perform ray tracing and simulation. The optimal orientation and installation location pre-calculation module comprehensively evaluates the full-space coverage quality under all-orientation configurations, using coverage uniformity and peak coverage quality as multi-objective optimization indicators, and calculates the optimal orientation using a genetic algorithm. The installation location determination module is used to collect users' historical movement behavior data and dynamically optimize the installation location or orientation based on the collected historical movement behavior data.

[0019] It should be noted that this invention can simulate the propagation performance of electromagnetic waves in specific environments with high precision, accurately considering complex propagation mechanisms such as wall reflection, diffraction, and scattering. The digital twin synchronizes in real time with dynamic changes in the physical environment, such as furniture placement and personnel positions, and automatically triggers recalculation when the environment changes. Applying digital twin and ray tracing technology to the field of home antenna direction adjustment, unlike the "trial and error" adjustment of existing technologies, this solution can predict the optimal direction "before adjustment" through physically accurate electromagnetic simulation, fundamentally solving the problem of blind and inefficient direction adjustment. Combined with machine learning to accelerate the prediction model, millisecond-level response to environmental changes is achieved.

[0020] Furthermore, within the adjustable smart home antenna, a three-dimensional home model is acquired, including information on wall location, wall material, transmission loss, and the dimensions and location of furniture, to construct a digital model of the home environment. This includes: Guide users to use the depth sensor built into the mobile terminal to scan the indoor space, or use the router's multi-antenna array to transmit detection signals and analyze the echoes to analyze the signal's wall-penetrating attenuation characteristics. Based on the signal's wall penetration attenuation characteristics, the wall thickness information is estimated. At the same time, data including wall thickness, wall location, wall material information, and transmission loss characteristics are generated. Obtain the dimensions and location information of the home furnishings, and construct a 3D home model based on the dimensions and location information of the home furnishings; A home environment model is constructed based on the wall thickness, wall location, dielectric constant, and transmission loss characteristics. This home environment model, along with a three-dimensional home model showing the dimensions and location of the home, forms a digital home environment model.

[0021] It should be noted that depth sensors include LiDAR in iPhones or ToF sensors in Android devices. Different material types result in different dielectric constants and transmission losses when light penetrates objects. Furthermore, walls of different thicknesses also bring different wall penetration attenuation characteristics. Therefore, this method can form a digital model of the home environment to dynamically simulate the actual light penetration.

[0022] Furthermore, in the adjustable smart home antenna, the electromagnetic parameters of each wall are marked in the digital model of the home environment, specifically: Obtain the dielectric constant and conductivity corresponding to different material types, construct a material database, and input the dielectric constant and conductivity corresponding to different material types into the material database for storage; Obtain material information for each wall in the digital model of the home environment, and call data from the material database to obtain the dielectric constant and conductivity corresponding to the material type of each wall in the digital model of the home environment; The electromagnetic parameters of each wall in the digital model of the home environment are obtained based on the dielectric constant and conductivity corresponding to the material type of each wall. For unknown materials, the electromagnetic parameters are calculated by measuring the actual signal strength at reference points at different locations, using a reverse optimization algorithm, and then marking the electromagnetic parameters of each wall in the digital model of the home environment.

[0023] It should be noted that the electromagnetic parameters of each wall are labeled in the 3D model. For known wall materials, such as gypsum board, brick wall, and concrete, the corresponding dielectric constant and conductivity are retrieved from the material database; for unknown materials, the electromagnetic parameters are deduced by measuring the actual signal strength at reference points at different locations and then using a reverse optimization algorithm.

[0024] Furthermore, in adjustable smart home antennas, the three-dimensional radiation pattern of the antenna is imported into the simulation environment based on electromagnetic parameters. The antenna orientation angle is discretized in the digital twin, and ray tracing and simulation are performed, specifically as follows: Initialize the orientation of the home antenna, import the three-dimensional radiation pattern of the home antenna into the simulation environment, discretize the antenna orientation angle in the digital model of the home environment, configure each orientation based on electromagnetic parameters, and emit a large amount of light from the antenna position; Based on the material properties of the wall, the absorbed light data is estimated, and the energy of the transmitted light is calculated based on the absorbed light data and the characteristics of the emitted light. The energy distribution of each ray of light after multiple reflections, transmissions and diffractions is tracked to reach different areas of the room. The full-space path gain, received signal strength, and signal-to-interference-plus-noise ratio distribution are calculated based on the energy distribution in each region.

[0025] It should be noted that the absorption of light data will vary depending on the type and thickness of the wall material. By calculating the energy of the transmitted light based on the absorbed light data and the characteristics of the emitted light, the path gain, received signal strength, and signal-to-interference-plus-noise ratio distribution in the entire space can be more accurately predicted, thereby improving the rationality of antenna adjustment.

[0026] Furthermore, in adjustable smart home antennas, the overall spatial coverage quality under omnidirectional configuration is comprehensively evaluated. Using coverage uniformity and peak coverage quality as multi-objective optimization indicators, a genetic algorithm is used to calculate the optimal orientation and estimated installation location. Specifically: Based on the energy distribution of each region, the full-space path gain, received signal strength and signal-to-interference-plus-noise ratio distribution are calculated, and the full-space coverage quality under the all-directional configuration is comprehensively evaluated to obtain coverage uniformity and peak coverage quality. A genetic algorithm is introduced, and the number of generations is set based on the genetic algorithm. Evaluation indicators for coverage uniformity and peak coverage quality are set to determine whether coverage uniformity and peak coverage quality meet the evaluation indicators. Among them, coverage uniformity is the minimization of the RSS variance of each region, and peak coverage quality is the maximization of the RSS of key regions.

[0027] When the coverage uniformity and peak coverage quality meet the evaluation indicators for coverage uniformity and peak coverage quality, the antennas should be configured according to their current orientation. When the coverage uniformity and peak coverage quality do not meet the evaluation criteria for coverage uniformity and peak coverage quality, the orientation of the current home antenna is reset, and the process is iterated and inherited until the coverage uniformity and peak coverage quality meet the evaluation criteria for coverage uniformity and peak coverage quality.

[0028] Furthermore, the adjustable smart home antenna collects historical user movement data and dynamically optimizes its installation location or orientation based on this data. Specifically, this includes: With user authorization, the system continuously collects the user's three-dimensional location coordinates in the home to form time-series trajectory data. The collection frequency is no less than a preset frequency threshold. The raw trajectory data is then subjected to Kalman filtering for noise reduction, stop point detection, and trajectory segmentation. The system continuously collects the user's three-dimensional location coordinates within the home using methods such as multi-antenna CSI positioning, Wi-Fi RTT ranging, UWB positioning, or multi-sensor fusion to generate time-series trajectory data. The acquisition frequency is no less than 10Hz to ensure the temporal resolution of the trajectory. The raw trajectory data undergoes Kalman filtering for noise reduction, stop point detection, and trajectory segmentation.

[0029] Among them, speeds below the threshold and durations exceeding the threshold are judged as dwellings. Statistical features are extracted from the segmented trajectory fragments, including the frequency of visits to each room and the distribution of dwell time. The spatial geometry of preferred areas and commonly used paths in different time periods is also analyzed, and the trajectories are clustered using a hierarchical clustering algorithm to identify several typical movement patterns. By clustering the trajectories using a hierarchical clustering algorithm, several typical movement patterns were identified, such as movement from the living room to the bedroom at night and movement from the bedroom to the study on weekday mornings.

[0030] The LSTM prediction model is trained based on historical trajectory data. It takes the user location sequence of the past N time steps as input, outputs the location prediction of the next K time steps, and dynamically optimizes the orientation of home antennas based on the location of the next K time steps.

[0031] It's important to note that the LSTM prediction model is trained based on historical trajectory data. It takes the user's location sequence over the past N time steps as input, where N is typically between 5 and 20, corresponding to a historical window of 0.5 to 2 seconds. The output is the location prediction for the next K time steps, where K is typically between 5 and 50, corresponding to a prediction window of 0.5 to 5 seconds. The LSTM model is specifically designed for time series prediction problems and can effectively capture the long-term and short-term dependencies in the trajectory. The model can be continuously updated online. Whenever new trajectory data arrives, the model is incrementally trained in mini-batch mode, allowing the predictive ability to adapt to changes in lifestyle, such as user relocation or adjustments to sleep schedules.

[0032] It's important to note that user movement within the home is not random but exhibits significant regularity and patterns: fixed daily routines (e.g., bedroom → bathroom → kitchen → living room), frequently visited areas during specific time periods (study during work hours, living room in the evening), and consistent route preferences. This solution leverages these regularities, using machine learning algorithms to extract movement patterns from historical trajectories. It predicts the user's location several seconds to tens of seconds in advance, even before they begin moving, and adjusts the antenna orientation accordingly.

[0033] Furthermore, in adjustable smart home antennas, the orientation of the antenna is dynamically optimized based on its position over the next K time steps, specifically as follows: After the prediction model outputs the future location sequence, the deviation angle between the current orientation of the home antenna and the direction of the predicted location is calculated. If the deviation angle exceeds the preset threshold, and the distance between the predicted location and the antenna location does not exceed the effective coverage range of the antenna; After the prediction model outputs the future location sequence, the system calculates the deviation angle between the antenna's current orientation and the predicted location. If the deviation angle exceeds a preset threshold, such as 10° in the horizontal direction or 5° in the pitch direction, and the distance between the predicted location and the antenna location meets certain conditions and does not exceed the antenna's effective coverage range, then pre-alignment adjustment is triggered. The timing of the pre-alignment adjustment is determined in advance based on the predicted location's arrival time—the system starts adjustment Δt seconds before the user's expected arrival, where Δt = mechanical rotation adjustment time + safety margin, ensuring that the antenna completes its turning precisely when the user reaches the target location.

[0034] Trigger pre-alignment adjustment. The timing of the pre-alignment adjustment is determined in advance based on the predicted arrival time of the location. The adjustment is initiated a preset number of seconds before the user is expected to arrive, ensuring that the antenna completes its turn exactly when the user arrives at the target location. Meanwhile, the prediction model outputs the position prediction variance for each future time step as a confidence index. When the prediction confidence is lower than the preset value, it automatically downgrades to reactive tracking mode and performs weighted fusion of predictive pre-alignment and reactive tracking.

[0035] It should be noted that this method introduces user movement behavior learning and prediction into the field of antenna orientation adjustment, achieving a paradigm shift from "reactive tracking" to "predictive pre-alignment." Unlike existing solutions that can only respond after the user's position changes, this solution uses deep learning of the user's movement patterns to complete antenna alignment before the user reaches the target location, fundamentally eliminating signal degradation caused by adjustment delay. The fusion decision-making mechanism of prediction confidence and reactive tracking ensures the system's robustness.

[0036] Furthermore, adjustable smart home antennas also include: The frequency of visits and length of stay for each room are statistically analyzed, and the frequency characteristics of each location node at the predetermined time are statistically analyzed based on the frequency of visits and length of stay for each room. The location nodes with frequency characteristics greater than the preset frequency characteristics are identified, and it is determined whether all location nodes with frequency characteristics greater than the preset frequency characteristics are within the effective coverage range of the antenna. When all location nodes with frequency characteristics greater than the preset frequency characteristics are within the effective coverage range of the antenna, the configuration is performed according to the current installation location of the home antenna, and the display is performed according to the preset method. When the location nodes with frequency characteristics greater than the preset frequency characteristics are not all within the effective coverage range of the antenna, the installation position of the current home antenna is reconfigured until all location nodes with frequency characteristics greater than the preset frequency characteristics are within the effective coverage range of the antenna.

[0037] It should be noted that by combining the user's dwell time within a preset period, the installation location of the home antenna can be optimized, ensuring that all nodes with frequency characteristics greater than the preset frequency characteristics are within the effective coverage range of the antenna, thus further improving the rationality of home antenna adjustment.

[0038] In addition, this method also includes: Reference points were divided into a 0.5m × 0.5m grid within the home space. At each reference point, a data acquisition device equipped with a multi-antenna array was used to receive Wi-Fi signals from the antennas, and the CSI amplitude and phase information of each subcarrier was extracted to form a multi-dimensional feature vector. A convolutional neural network combined with a spatial attention mechanism is used to extract the degree of the multidimensional feature vector. The spatial attention layer dynamically focuses on the signal region with the highest information value, and the feature vectors of all reference points and their corresponding coordinate labels are stored in the fingerprint database. Deep reinforcement learning is embedded, and home environment simulators or actual operating data are used as training environments. The PPO algorithm is used for policy gradient updates. During the training process, the agent explores the optimal adjustment strategy under different user movement trajectories. The data in the fingerprint database is input into the CNN model to obtain the user's location estimate. The current location and historical trajectory are input into the PPO policy network, which outputs the optimal adjustment direction and angle. The antenna is then rotated to the target orientation by a motor.

[0039] It should be noted that CSI fingerprint positioning and deep reinforcement learning are integrated into home antenna adjustment to achieve predictive antenna tracking and adjustment. The introduction of deep reinforcement learning enables the antenna to autonomously learn the optimal adjustment strategy without labeled data, and the computational efficiency of the PPO algorithm is far superior to traditional optimization methods, possessing millisecond-level real-time response capabilities.

[0040] In addition, this method also includes: Using the home gateway as a central aggregation server, all devices in the home with antenna adjustment capabilities are registered as clients. The global optimization goal is to maximize the average signal quality of all terminals in the house while minimizing co-channel interference between devices. Each device deploys a local deep Q network, taking the local antenna orientation, the RSSI of terminals within the local coverage area, and the interference intensity received from neighboring devices as inputs, and outputs local antenna orientation adjustment actions. Each device periodically uploads its local model's gradient update parameters to the gateway after encrypting them using homomorphic encryption. The gateway then uses a federated averaging algorithm to aggregate the model parameter updates from each client. The weighted average coefficient is proportional to the amount of data and the number of training rounds for each device. The aggregated global model is then distributed to each device to replace the local model. The observation status of each device includes RSSI of signals received from neighboring devices. The model is trained to actively reduce the radiated energy pointing towards neighboring devices when adjusting its own antenna orientation, thereby forming a distributed cooperative interference suppression throughout the house. Furthermore, through polarization diversity-maximum ratio combining technology, the antenna polarization is adaptively adjusted in both the vertical and horizontal axes.

[0041] It's important to note that modern homes typically deploy multiple devices with antenna adjustment capabilities, including main routers, Mesh sub-nodes, smart gateways, and various IoT base stations. These devices adjust their antenna directions independently, lacking coordination. By integrating a federated learning model, the global optimization objective is to maximize the average signal quality of all devices throughout the house while minimizing co-channel interference between devices. The model is trained to proactively reduce radiated energy pointing towards neighboring devices when adjusting its own antenna orientation, thus forming a distributed cooperative interference suppression mechanism throughout the house. This achieves coordinated optimization of antenna orientation across the entire house, further reducing interference between devices.

[0042] Furthermore, an LSTM prediction model is trained based on historical trajectory data. Taking the user's location sequence over the past N time steps as input, the model outputs location predictions for the next K time steps, specifically including: The system continuously collects the user's three-dimensional location coordinates through multi-antenna CSI positioning, Wi-Fi RTT ranging, or UWB fusion positioning, and constructs a continuous trajectory map based on the user's three-dimensional location coordinates, and discretizes the home environment into a graph. In the diagram, the node set includes the geometric center of each room, each doorway / corridor turning point, and the key coordinates of each densely furnished area (such as the location with the most visits). During the construction of edges, if a straight path between two nodes does not pass through any walls, then there is an undirected edge. Each edge is assigned a weight, and each node corresponds to a learnable embedding vector, which is mapped to a static embedding through a small fully connected network. This static embedding is then concatenated with the learnable embedding and used as the input to the graph convolution. For each time step, the user's location is mapped to a probability distribution on the graph, the distance between the user's location and all nodes is calculated, and a soft attribution vector is generated using radial basis functions. For each time step, the user's location is represented as a node activation vector on the graph. The node activation vector on the graph is input into the temporal attention layer, and the model outputs the user's location sequence for the next K time steps.

[0043] It should be noted that by representing the user's location as the node activation vector on the graph, the graph structure is not a simple Euclidean distance graph, but is strictly constructed based on physical connectivity (walls cannot be crossed). This allows the model to inherently understand the reachability relationships between rooms, avoid predicting physically impossible trajectories through walls, and improve the accuracy of predictions.

[0044] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0045] 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. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0046] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0047] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0048] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium 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 methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0049] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An adjustable smart home antenna, characterized in that, include: The home environment digital modeling module is used to acquire three-dimensional home models of the home environment, including wall location, wall material information, transmission loss, and furniture size and location, and to construct a digital model of the home environment. An electromagnetic parameter calibration module is used to mark the electromagnetic parameters of each wall in the digital model of the home environment. The ray tracing simulation module is used to import the three-dimensional radiation pattern of the antenna into the simulation environment based on the electromagnetic parameters, discretize the antenna orientation angle in the digital twin, and perform ray tracing and simulation. The optimal orientation pre-calculation module comprehensively evaluates the full-space coverage quality under all orientation configurations, using coverage uniformity and peak coverage quality as multi-objective optimization indicators, and calculates the optimal orientation using a genetic algorithm. The installation location determination module is used to collect users' historical movement behavior data and dynamically optimize the installation location or orientation based on the collected historical movement behavior data.

2. The adjustable smart home antenna according to claim 1, characterized in that, Obtain a 3D home model containing information on wall locations, wall materials, light transmission loss, and furniture dimensions and locations to construct a digital model of the home environment. This includes: Guide users to use the depth sensor built into the mobile terminal to scan the indoor space, or use the router's multi-antenna array to transmit detection signals and analyze the echoes to analyze the signal's wall-penetrating attenuation characteristics. Based on the wall penetration attenuation characteristics of the signal, the thickness information of the wall is estimated. At the same time, data including the wall thickness, wall location, wall material information, and transmission loss characteristics are generated. Obtain the dimensions and location information of the home furnishings, and construct a three-dimensional home model based on the dimensions and location information of the home furnishings; A home environment model is constructed based on the wall thickness information, wall location, dielectric constant, and transmission loss characteristic data. The home environment model and the three-dimensional home model with the size and location of the home are combined to form a digital home environment model.

3. An adjustable smart home antenna according to claim 1, characterized in that, The electromagnetic parameters of each wall are marked in the digital model of the home environment, specifically: Obtain the dielectric constant and conductivity corresponding to different material types, construct a material database, and input the dielectric constant and conductivity corresponding to different material types into the material database for storage; Obtain the material information of each wall in the digital model of the home environment, and call the data in the material database to obtain the dielectric constant and conductivity corresponding to the material type of each wall in the digital model of the home environment; The electromagnetic parameters of each wall in the digital home environment model are obtained based on the dielectric constant and conductivity corresponding to the material type of each wall. For unknown materials, the electromagnetic parameters are calculated by measuring the actual signal strength at reference points at different locations and using a reverse optimization algorithm. The electromagnetic parameters of each wall are then marked in the digital model of the home environment.

4. An adjustable smart home antenna according to claim 1, characterized in that, Based on the aforementioned electromagnetic parameters, the three-dimensional radiation pattern of the antenna is imported into the simulation environment. The antenna orientation angle is discretized in the digital twin, and ray tracing and simulation are performed, specifically as follows: Initialize the orientation of the home antenna, import the three-dimensional radiation pattern of the home antenna into the simulation environment, discretize the antenna orientation angle in the digital model of the home environment, configure each orientation based on the electromagnetic parameters, and emit a large amount of light from the antenna position; Based on the material properties of the wall, the absorbed light data is estimated, and the energy of the transmitted light is calculated according to the absorbed light data and the characteristics of the emitted light. The energy distribution of each light ray after multiple reflections, transmissions and diffractions is tracked to reach different areas of the room. The full-space path gain, received signal strength, and signal-to-interference-plus-noise ratio distribution are calculated based on the energy distribution of each region.

5. An adjustable smart home antenna according to claim 4, characterized in that, The overall spatial coverage quality under all-orientation configuration is comprehensively evaluated. Using coverage uniformity and peak coverage quality as multi-objective optimization indicators, a genetic algorithm is used to calculate the optimal orientation. Specifically: Based on the energy distribution of each region, the full-space path gain, received signal strength and signal-to-interference-plus-noise ratio distribution are calculated, and the full-space coverage quality under the all-directional configuration is comprehensively evaluated to obtain coverage uniformity and peak coverage quality. A genetic algorithm is introduced, and the number of generations is set based on the genetic algorithm. Coverage uniformity and peak coverage quality evaluation indicators are set to determine whether the coverage uniformity and peak coverage quality meet the coverage uniformity and peak coverage quality evaluation indicators. When the coverage uniformity and peak coverage quality meet the evaluation indicators for coverage uniformity and peak coverage quality, the configuration shall be made according to the current orientation; When the coverage uniformity and peak coverage quality do not meet the evaluation criteria for coverage uniformity and peak coverage quality, the current orientation is reset until the coverage uniformity and peak coverage quality meet the evaluation criteria for coverage uniformity and peak coverage quality.

6. An adjustable smart home antenna according to claim 4, characterized in that, Collecting users' historical movement behavior data and dynamically optimizing the installation location or orientation based on this data, specifically including: With user authorization, the system continuously collects the user's three-dimensional location coordinates in the home to form time-series trajectory data. The collection frequency is no less than a preset frequency threshold. The raw trajectory data is then subjected to Kalman filtering for noise reduction, stop point detection, and trajectory segmentation. Among them, speeds below the threshold and durations exceeding the threshold are judged as dwellings. Statistical features are extracted from the segmented trajectory fragments, including the frequency of visits to each room and the distribution of dwell time. The spatial geometry of preferred areas and commonly used paths in different time periods is also analyzed, and the trajectories are clustered using a hierarchical clustering algorithm to identify several typical movement patterns. The LSTM prediction model is trained based on historical trajectory data. It takes the user location sequence of the past N time steps as input, outputs the location prediction of the next K time steps, and dynamically optimizes the orientation of home antennas based on the location of the next K time steps.

7. An adjustable smart home antenna according to claim 6, characterized in that, The orientation of the home antenna is dynamically optimized based on its position over the next K time steps, specifically as follows: After the prediction model outputs the future location sequence, the deviation angle between the current orientation of the home antenna and the direction of the predicted location is calculated. If the deviation angle exceeds the preset threshold, and the distance between the predicted location and the antenna location does not exceed the effective coverage range of the antenna; Trigger pre-alignment adjustment. The timing of the pre-alignment adjustment is determined in advance based on the predicted arrival time of the location. The adjustment is initiated a preset number of seconds before the user is expected to arrive, ensuring that the antenna completes its turn exactly when the user arrives at the target location. Meanwhile, the prediction model outputs the position prediction variance for each future time step as a confidence index. When the prediction confidence is lower than the preset value, it automatically downgrades to reactive tracking mode and performs weighted fusion of predictive pre-alignment and reactive tracking.

8. An adjustable smart home antenna according to claim 6, characterized in that, Also includes: The frequency of visits and duration of stay in each room are statistically analyzed, and the frequency characteristics of each location node at a predetermined time are statistically analyzed based on the frequency of visits and duration of stay in each room. The location nodes with frequency characteristics greater than a preset frequency characteristic are identified, and it is determined whether all the location nodes with frequency characteristics greater than the preset frequency characteristic are within the effective coverage range of the antenna. When all location nodes with frequency characteristics greater than the preset frequency characteristics are within the effective coverage range of the antenna, the antenna is configured according to its current installation location and displayed in a preset manner. When the location nodes whose frequency characteristics are greater than the preset frequency characteristics are not all within the effective coverage range of the antenna, the installation position of the current home antenna is reconfigured until all location nodes whose frequency characteristics are greater than the preset frequency characteristics are within the effective coverage range of the antenna.

Images