Indoor and outdoor integrated coverage test and evaluation system for power wireless network

Abstract

The invention relates to the technical field of information technology management, in particular to an indoor and outdoor integrated coverage test and evaluation system for a power wireless network. Comprising a data acquisition module, a positioning and registering module, an environment modeling module, an intelligent evaluation module, an optimization decision module and a feedback updating module. The data acquisition module is used for acquiring multimode wireless signals and environment state information of an electric power communication network; the positioning and registering module is used for realizing high-precision positioning and indoor and outdoor unified space-time calibration of acquired data; the environment modeling module is used for constructing a three-dimensional propagation model including buildings, power equipment and topographic elements. The system supports synchronous acquisition and conjoint analysis of indoor and outdoor data, solves the problems that indoor and outdoor tests are separated and data cannot be aligned in an existing scheme, and realizes global coverage evaluation of a real power communication environment.

Description

Technical Field

[0001] This invention relates to the field of information technology management technology, specifically to an integrated indoor and outdoor coverage testing and evaluation system for power wireless networks. Background Technology

[0002] With the continuous improvement of the digitalization and intelligence level of power systems, business scenarios such as the power Internet of Things, distribution automation, smart terminal data collection, remote monitoring, and status awareness rely heavily on the support of wireless communication networks. To ensure the real-time performance and reliability of power production control data, monitoring data, and user-side information, the coverage quality and communication performance of power wireless networks have become important factors restricting the operation of smart grids.

[0003] Existing power grid wireless network coverage testing and evaluation technologies primarily rely on manual field testing or simple simulation analysis, which generally suffer from the following shortcomings: Traditional coverage testing systems often focus on outdoor scenarios, using single-point sampling or mobile measurement methods to collect signal strength, making it difficult to consider the internal environment of buildings. Due to the complex spaces and severe obstructions in locations such as power control rooms, distribution rooms, and underground utility tunnels, the indoor and outdoor wireless signal propagation models differ significantly, resulting in evaluation results that fail to reflect the true coverage situation. Existing propagation models are mostly based on empirical formulas or two-dimensional simulations, failing to fully consider the three-dimensional influence of buildings, transmission lines, towers, and terrain, and cannot be corrected based on measured data. As the scene structure changes (such as the addition of new buildings and metal reflectors), model errors accumulate, significantly reducing evaluation accuracy. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an integrated indoor and outdoor coverage testing and evaluation system for power wireless networks. This system can generate highly accurate, reliable, context-appropriate, and easy-to-understand technical solutions.

[0005] This solution presents an integrated indoor and outdoor coverage testing and evaluation system for power wireless networks, comprising a data acquisition module, a positioning and registration module, an environmental modeling module, an intelligent evaluation module, an optimization decision-making module, and a feedback update module. The data acquisition module is used to collect multi-mode wireless signals and environmental status information of the power communication network; The positioning and registration module is used to achieve high-precision positioning of the collected data and unified spatiotemporal calibration of indoor and outdoor spaces; The environment modeling module is used to construct a three-dimensional propagation model that includes buildings, power equipment, and terrain features; The intelligent evaluation module is used to integrate collected data and model information, and use artificial intelligence algorithms to generate coverage performance indicators. The optimization decision module is used to automatically generate network parameter optimization strategies based on the intelligent evaluation results; The feedback update module is used to collect test feedback and running data and to update the system model through self-learning.

[0006] Preferably, the data acquisition module includes a multi-mode signal acquisition unit, an interference monitoring unit, and a data synchronization unit: The multi-mode signal acquisition unit is used to simultaneously acquire parameters such as received power, signal-to-noise ratio, latency, packet loss rate, and spectral power density of 5G, LTE, NB-IoT, LoRa, and Wi-Fi signals. The interference monitoring unit is used to perform short-time Fourier transform analysis on the sampled signal, identify the interference energy peak, and record the interference frequency band. The data synchronization unit performs unified timestamp calibration based on a high-precision time reference to ensure that the data time deviation of different sampling devices is ≤1ms.

[0007] Preferably, the positioning and registration module includes a GNSS positioning unit, a UWB positioning unit, an inertial navigation unit, and a data fusion unit. The GNSS positioning unit is used to provide meter-level positioning in outdoor environments; The UWB positioning unit is used to provide centimeter-level positioning in indoor environments; The inertial navigation unit is used to maintain a continuous trajectory in areas where signals are blocked; The data fusion unit dynamically weights and fuses GNSS, UWB, and IMU data based on the extended Kalman filter algorithm, and achieves seamless indoor and outdoor registration with a positioning error of ≤0.2m through time synchronization and trajectory correction.

[0008] Preferably, the environment modeling module includes a 3D modeling unit, a ray propagation calculation unit, and a model calibration unit: The 3D modeling unit uses LiDAR scanning to generate point cloud data of buildings, power towers, transmission lines and terrain, with a resolution of not less than 5cm; The ray propagation calculation unit simulates signal reflection, diffraction, and occlusion based on the ray tracing algorithm to generate a path loss matrix. The model calibration unit corrects the path loss index by the difference between the measured signal strength and the simulation results, with a correction threshold of less than 3dB.

[0009] Preferably, the intelligent evaluation module includes a feature encoding unit, a graph convolutional network unit, a spatiotemporal prediction unit, and an interpretable output unit. The feature encoding unit encodes signal strength, interference, terrain features, and node locations into a unified spatiotemporal feature tensor; The graph convolutional network unit is used to extract spatial topological correlations; The spatiotemporal prediction unit uses a multi-head attention mechanism based on the Transformer structure to capture temporal dynamic features; The interpretable output unit uses the Shapley value method to calculate the feature contribution and generate an index visualization chart to illustrate the impact of various environments or parameters on coverage performance.

[0010] Preferably, the optimization decision module includes a blind spot identification unit, a parameter optimization unit, and a control output unit: The blind spot identification unit automatically detects areas below -90dBm and generates a blind spot mask based on the signal strength distribution output by the intelligent evaluation module. The parameter optimization unit iteratively optimizes the base station transmit power, antenna elevation angle and site location based on the deep reinforcement learning algorithm (DDPG), and its reward function comprehensively considers coverage, interference and power consumption. The control output unit sends the optimized parameters to the corresponding devices through the power communication network management interface to achieve automated closed-loop adjustment.

[0011] Preferably, the system supports multiple types of mobile testing terminals, including handheld terminals, unmanned vehicles, and drones. The drone is equipped with obstacle avoidance radar and path planning module, and adopts a composite path optimization strategy based on AI and genetic algorithm to ensure that the obstacle avoidance distance is ≥2m and the test coverage is ≥95% in complex power line environment.

[0012] Preferably, the feedback update module includes a feedback acquisition unit, a model correction unit, and a knowledge update unit: The feedback collection unit is used to receive user evaluations, test reports, and network operation logs. The model correction unit fine-tunes the neural network parameters of the intelligent evaluation module based on feedback data, and uses lightweight adaptation technology to reduce the amount of parameter updates. The knowledge update unit synchronizes the feedback results to the environment modeling module and the optimization decision module to update the propagation model and optimization strategy.

[0013] Preferably, the system further includes a data fusion and edge computing node, deployed on a portable server in the field, equipped with a GPU acceleration unit, responsible for performing local feature extraction, data compression and temporary evaluation; The edge nodes communicate with the cloud via a 5G private network, with a data transmission latency of less than 50ms.

[0014] Preferably, the system supports a three-dimensional visualization module, which is used to overlay and display coverage heat maps, blind spot distribution maps, and interference source location maps on the GIS platform, so as to realize interactive evaluation and display of the coverage quality of power wireless networks.

[0015] Compared with the prior art, the advantages of this invention are: (1) The environmental modeling module of the present invention combines lidar point cloud scanning and three-dimensional reconstruction technology, which can restore buildings, power towers, transmission lines and terrain elements at a resolution of 5cm and generate a visualized three-dimensional propagation model. The system supports the synchronous acquisition and joint analysis of indoor and outdoor data, which solves the problem of the separation of indoor and outdoor testing and the inability to align data in the existing scheme, and realizes the full-area coverage assessment of the real power communication environment.

[0016] (2) The ray propagation calculation unit in the system simulates the reflection, diffraction, and occlusion effects of signals based on the ray tracing algorithm, and can generate a path loss matrix. At the same time, it performs dynamic correction by utilizing the difference between the measured signal and the simulated value, and controls the correction threshold within 3dB, so that the simulation results are highly consistent with the actual propagation characteristics. This design significantly improves the accuracy of power wireless signal prediction under complex terrain.

[0017] (3) The intelligent evaluation module of the present invention maps signals, interference, terrain and node topology into a unified spatiotemporal feature tensor through the feature encoding unit, extracts spatial topological correlation using graph convolutional network, and further uses a multi-head attention mechanism based on Transformer to capture temporal dynamic features, thereby realizing spatiotemporal joint prediction of signal coverage.

[0018] (4) The system introduces the Shapley value method in the evaluation result output stage to calculate the contribution of input features and generate an index visualization chart and a feature influence ranking table. Through this mechanism, the system not only outputs coverage performance indicators, but also displays the quantitative impact of various environments or parameters on signal coverage performance, providing maintenance personnel with traceable and understandable technical basis. Attached Figure Description

[0019] Figure 1 This is a system architecture diagram of an integrated indoor and outdoor coverage testing and evaluation system for power wireless networks proposed in this invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0021] Reference Figure 1A power wireless network indoor and outdoor integrated coverage testing and evaluation system includes a data acquisition module, a positioning and registration module, an environmental modeling module, an intelligent evaluation module, an optimization decision-making module, and a feedback update module. The data acquisition module is used to collect multi-mode wireless signals and environmental status information from the power communication network; the data acquisition module includes a multi-mode signal acquisition unit, an interference monitoring unit, and a data synchronization unit. The multi-mode signal acquisition unit is used to simultaneously acquire parameters such as received power, signal-to-noise ratio, latency, packet loss rate, and spectral power density of 5G, LTE, NB-IoT, LoRa, and Wi-Fi signals. The interference monitoring unit is used to perform short-time Fourier transform analysis on the sampled signal, identify the interference energy peak, and record the interference frequency band. The data synchronization unit performs unified timestamp calibration based on a high-precision time reference to ensure that the data time deviation of different sampling devices is ≤1ms.

[0022] The positioning and registration module is used to achieve high-precision positioning and unified indoor and outdoor spatiotemporal calibration of the collected data; the positioning and registration module includes a GNSS positioning unit, a UWB positioning unit, an inertial navigation unit, and a data fusion unit. The GNSS positioning unit is used to provide meter-level positioning in outdoor environments; The UWB positioning unit is used to provide centimeter-level positioning in indoor environments; The inertial navigation unit is used to maintain a continuous trajectory in areas where signals are blocked; The data fusion unit dynamically weights and fuses GNSS, UWB, and IMU data based on the extended Kalman filter algorithm, and achieves seamless indoor and outdoor registration with a positioning error of ≤0.2m through time synchronization and trajectory correction.

[0023] The environment modeling module is used to construct a three-dimensional propagation model containing buildings, power equipment, and terrain features; the environment modeling module includes a three-dimensional modeling unit, a ray propagation calculation unit, and a model calibration unit. The 3D modeling unit uses LiDAR scanning to generate point cloud data of buildings, power towers, transmission lines and terrain, with a resolution of not less than 5cm; The ray propagation calculation unit simulates signal reflection, diffraction, and occlusion based on the ray tracing algorithm to generate a path loss matrix. The model calibration unit corrects the path loss index by the difference between the measured signal strength and the simulation results, with a correction threshold of less than 3dB.

[0024] In this invention, the environment modeling module is used to construct a three-dimensional propagation model including buildings, power equipment, and terrain features to support simulation calculations and optimization analysis of indoor and outdoor wireless network signal propagation characteristics. This module includes a three-dimensional modeling unit, a ray propagation calculation unit, and a model calibration unit. These units work collaboratively to achieve the entire process from scene acquisition and signal propagation modeling to simulation model calibration.

[0025] I. Three-dimensional modeling unit The function of the 3D modeling unit is to acquire spatial structure data of a real scene and generate a 3D point cloud model. Its workflow is as follows: Data collection: Use vehicle-mounted or handheld LiDAR scanning equipment (such as Velodyne HDL-32E or Leica RTC360) to perform a 360° panoramic scan of the target area.

[0026] The sampling frequency is set to 10Hz to 20Hz, and the point cloud resolution is preferably no less than 5cm to ensure accurate capture of building shapes, power tower structures, transmission line suspension patterns, and terrain undulations.

[0027] Point cloud preprocessing: Noise points are removed by point cloud filtering algorithms (such as VoxelGrid downsampling and StatisticalOutlierRemoval), and terrain segmentation algorithms (such as RANSAC plane fitting) are used to distinguish between terrain and structures.

[0028] 3D model reconstruction: Perform surface reconstruction (such as the Poisson reconstruction algorithm) on the processed point cloud data to generate a 3D mesh model with normal vectors; At the same time, texture mapping is performed using GIS vector data or UAV imagery data to achieve visual integration of buildings and terrain.

[0029] Object semantic annotation: A semantic segmentation network based on deep learning (such as PointNet++) is used to classify and identify point cloud data, and buildings, power equipment, transmission lines and terrain features in the model are labeled so that different electromagnetic parameters (such as dielectric constant, reflection coefficient, etc.) can be assigned to different media during subsequent propagation calculations.

[0030] II. Ray Propagation Calculation Unit The ray propagation calculation unit is used to simulate the propagation path and energy attenuation process of radio signals in a 3D scene. Its core algorithm is ray tracing. The specific working principle is as follows: Ray emission: Using the location of the base station or transmitting antenna as the ray source, several rays are emitted according to the spherical coordinate system. The number of rays can be set according to the angular resolution, generally with an interval of 1° to 5° per degree of azimuth.

[0031] Propagation Modeling: When a ray encounters different object surfaces during its propagation, the following calculations are performed based on the object's material properties: Reflection: The reflection coefficient is calculated based on the Fresnel reflection model; Diffraction: When rays pass through building edges or steel components of power towers, the diffraction loss is calculated using the Uniform Theory of Diffraction (UTD) model; Penetration and Blockage: When rays pass through walls, windows or trees, the penetration loss is calculated based on the medium thickness and frequency response. If the path is completely blocked, it is discarded.

[0032] Path loss matrix generation: The received power value is obtained by summing the energy of all ray paths arriving at the receiving point. This generates a path loss matrix for the entire region, with a resolution corresponding to the 3D model mesh.

[0033] Output data format: The output includes three-dimensional coordinates (x, y, z), path loss value L, and number of reflections N. ref Diffraction number N dif Data such as path length D are used for subsequent model calibration and signal coverage evaluation.

[0034] III. Model Calibration Unit The model calibration unit is used to correct the error between the simulation model and the measured results, improving the accuracy and generalization ability of the propagation model. Its working principle is as follows: Data collection: Set up several signal sampling points (generally no less than 30) in the test area, and use a spectrum analyzer or network test terminal to record the measured signal strength (RSSI value) and geographical coordinates.

[0035] 2. Error Calculation: Calculate the difference between the measured signal strength and the received power output from the ray tracing simulation. meas sim .

[0036] 3. Path loss index adjustment: If the average error If the value exceeds the set threshold (3 dB), then an adaptive correction algorithm for the path loss exponent n is executed:

[0037] in, Learning rate ( N represents the number of sampling points. The calibration process is iterative until the average error is less than 3dB. 4. Model Validation and Output: The calibrated propagation model is recalculated for coverage and compared with the results of the newly added test points. If the accuracy meets the standard, the final 3D propagation model is output.

[0038] The intelligent evaluation module is used to fuse collected data and model information, and generate coverage performance indicators using artificial intelligence algorithms. The intelligent evaluation module includes a feature encoding unit, a graph convolutional network unit, a spatiotemporal prediction unit, and an interpretable output unit. The feature encoding unit encodes signal strength, interference, terrain features, and node locations into a unified spatiotemporal feature tensor; The graph convolutional network unit is used to extract spatial topological correlations; The spatiotemporal prediction unit uses a multi-head attention mechanism based on the Transformer structure to capture temporal dynamic features; The interpretable output unit uses the Shapley value method to calculate the feature contribution and generate an index visualization chart to illustrate the impact of various environments or parameters on coverage performance.

[0039] The intelligent evaluation module of this invention integrates field-collected data and simulation model information, utilizes artificial intelligence algorithms to intelligently predict and quantitatively evaluate the coverage performance of wireless networks, and generates interpretable index output results. This module includes a feature encoding unit, a graph convolutional network unit, a spatiotemporal prediction unit, and an interpretable output unit. These units work together to form a complete intelligent evaluation link. I. Feature Encoding Unit The feature encoding unit is used to uniformly map multi-source heterogeneous data (including signal measurement data, environmental elements, and model simulation output) into a computable spatiotemporal feature tensor space. 1. Input data includes: signal strength (RSSI, SINR, RSRP, RSRQ, etc.); interference power and neighboring cell signal ratio; terrain and obstacle features (height difference, occlusion rate, land cover type encoding); node location information (latitude and longitude, three-dimensional coordinates, topological connectivity); and the path loss matrix output by the propagation model. 2. Encoding methods: (1) Normalize the signal and interference features to map their value range to 0,1; (2) Use a dimensionality reduction method based on convolutional autoencoder for terrain features to map the original high-dimensional terrain texture into a fixed-length terrain vector. (3) Node topology information is expressed in the adjacency matrix. Indicates matrix elements This indicates that node i and node j are wirelessly reachable; otherwise, it is 0. (4) Finally, all features are concatenated to form a spatiotemporal feature tensor:

[0040] in Indicates the sampling time. Dimensions N represents the number of nodes, and D represents the feature dimension (typically 64–256 dimensions). 3. Output: Generates a unified spatiotemporal feature tensor. , which serves as the input to the graph convolutional network unit.

[0041] II. Graph Convolutional Network Unit This unit is used to model the spatial dependencies between nodes and extract feature interaction information from the topology. 1. Algorithm Structure: A two-layer stacked graph convolutional network structure is adopted. The basic calculation formula is as follows:

[0042] in: ; The adjacency matrix is ​​symmetrically normalized. This is the weight matrix for the first layer; 1. **Activation Function:** ReLU is a non-linear activation function. 2. **Propagation Method:** Graph convolution performs neighborhood aggregation in each iteration, ensuring that each node's features simultaneously reflect a weighted combination of its own features and those of its neighboring nodes. 3. **Output:** The output is a high-dimensional feature representation matrix of the output nodes. This matrix carries information about the spatial topological correlation and propagation environment between nodes. III. Spatiotemporal Prediction Unit This unit is used to capture the temporal dynamic characteristics of signal coverage, enabling modeling and prediction of signal change trends at multiple time points. 1. Input Structure: The input structure is the output of the previous unit... Stacking them along the time dimension forms a sequence tensor:

[0043] Where T is the length of the time window (generally taken as...). 2. Transformer structure components: Encoder part: contains a multi-head attention mechanism to calculate the correlation of features at different time slices.

[0044]

[0045] The decoder section fuses the predicted time-of-flight features with historical features to generate an estimate of signal coverage for future times. The number of multi-head cameras is set to... Head, hidden layer dimension is Dimensions, number of layers Layer. 3. Output metrics: Output the predicted signal coverage strength of each node at future time points. It also calculates performance metrics including Coverage Probability, Mean RSRP, and Interference Index.

[0046] IV. Explainable Output Unit: This unit is used to perform interpretability analysis on the AI ​​model's prediction results, quantify the contribution of each feature to the final prediction result, and generate a visual indicator report. 1. Shapley Value Calculation Method: The Shapley value algorithm based on game theory is used to calculate the marginal contribution of each input feature to the prediction output:

[0047] in For the model output function, 1. Contribution of features. 2. Interpretation process: (1) Randomly sample the model and generate a set of feature perturbations; (2) Calculate the output difference before and after the perturbation through forward inference of the model; (3) Rank the importance of each feature by weighted average of Shapley values. 3. Visual output: Output the calculation results in a graphical way, including: Feature Importance Ranking; Parameter Sensitivity Curve; Coverage Contribution Heatmap, which is used to show the impact of factors such as terrain undulation, node density, and power settings on signal coverage performance. 4. Output results: Generate a comprehensive evaluation report, including coverage uniformity index, interference impact coefficient, prediction error distribution and contribution of Shapley values ​​of each feature, to provide input basis for subsequent network optimization modules.

[0048] The optimization decision module is used to automatically generate network parameter optimization strategies based on intelligent evaluation results; the optimization decision module includes a blind spot identification unit, a parameter optimization unit, and a control output unit. The blind spot identification unit automatically detects areas below -90dBm and generates a blind spot mask based on the signal strength distribution output by the intelligent evaluation module. The parameter optimization unit iteratively optimizes the base station transmit power, antenna elevation angle and site location based on the deep reinforcement learning algorithm (DDPG), and its reward function comprehensively considers coverage, interference and power consumption. The control output unit sends the optimized parameters to the corresponding devices through the power communication network management interface to achieve automated closed-loop adjustment.

[0049] The feedback update module is used to collect test feedback and operational data for system model self-learning updates. The feedback update module includes a feedback acquisition unit, a model correction unit, and a knowledge update unit. The feedback collection unit is used to receive user evaluations, test reports, and network operation logs. The model correction unit fine-tunes the neural network parameters of the intelligent evaluation module based on feedback data, and uses lightweight adaptation technology to reduce the amount of parameter updates. The knowledge update unit synchronizes the feedback results to the environment modeling module and the optimization decision module to update the propagation model and optimization strategy.

[0050] The system supports multiple types of mobile testing terminals, including handheld terminals, unmanned vehicles, and drones. The drone is equipped with obstacle avoidance radar and path planning module, and adopts a composite path optimization strategy based on AI and genetic algorithm to ensure that the obstacle avoidance distance is ≥2m and the test coverage is ≥95% in complex power line environment.

[0051] The system further includes a data fusion and edge computing node, deployed on a portable server in the field, equipped with a GPU acceleration unit, which is responsible for performing local feature extraction, data compression and temporary evaluation; The edge nodes communicate with the cloud via a 5G private network, with a data transmission latency of less than 50ms.

[0052] The system supports a 3D visualization module, which is used to overlay and display coverage heat maps, blind spot distribution maps, and interference source location maps on the GIS platform, enabling interactive evaluation and display of the coverage quality of the power wireless network.

[0053] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0054] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A power wireless network indoor and outdoor integrated coverage testing and evaluation system, characterized in that, It includes a data acquisition module, a positioning and registration module, an environmental modeling module, an intelligent evaluation module, an optimization decision-making module, and a feedback and update module; The data acquisition module is used to collect multi-mode wireless signals and environmental status information of the power communication network; The positioning and registration module is used to achieve high-precision positioning of the collected data and unified spatiotemporal calibration of indoor and outdoor spaces; The environment modeling module is used to construct a three-dimensional propagation model that includes buildings, power equipment, and terrain features; The intelligent evaluation module is used to integrate collected data and model information, and use artificial intelligence algorithms to generate coverage performance indicators. The optimization decision module is used to automatically generate network parameter optimization strategies based on the intelligent evaluation results; The feedback update module is used to collect test feedback and running data and to update the system model through self-learning.

2. The integrated indoor and outdoor coverage testing and evaluation system for power grid wireless networks according to claim 1, characterized in that, The data acquisition module includes a multi-mode signal acquisition unit, an interference monitoring unit, and a data synchronization unit. The multi-mode signal acquisition unit is used to simultaneously acquire parameters such as received power, signal-to-noise ratio, latency, packet loss rate, and spectral power density of 5G, LTE, NB-IoT, LoRa, and Wi-Fi signals. The interference monitoring unit is used to perform short-time Fourier transform analysis on the sampled signal, identify the interference energy peak, and record the interference frequency band. The data synchronization unit performs unified timestamp calibration based on a high-precision time reference to ensure that the data time deviation of different sampling devices is ≤1ms.

3. The integrated indoor and outdoor coverage testing and evaluation system for power wireless networks according to claim 1, characterized in that, The positioning and registration module includes a GNSS positioning unit, a UWB positioning unit, an inertial navigation unit, and a data fusion unit. The GNSS positioning unit is used to provide meter-level positioning in outdoor environments; The UWB positioning unit is used to provide centimeter-level positioning in indoor environments; The inertial navigation unit is used to maintain a continuous trajectory in areas where signals are blocked; The data fusion unit dynamically weights and fuses GNSS, UWB, and IMU data based on the extended Kalman filter algorithm, and achieves seamless indoor and outdoor registration with a positioning error of ≤0.2m through time synchronization and trajectory correction.

4. The integrated indoor and outdoor coverage testing and evaluation system for power wireless networks according to claim 1, characterized in that, The environment modeling module includes a 3D modeling unit, a ray propagation calculation unit, and a model calibration unit. The 3D modeling unit uses LiDAR scanning to generate point cloud data of buildings, power towers, transmission lines and terrain, with a resolution of not less than 5cm; The ray propagation calculation unit simulates signal reflection, diffraction, and occlusion based on the ray tracing algorithm to generate a path loss matrix. The model calibration unit corrects the path loss index by the difference between the measured signal strength and the simulation results, with a correction threshold of less than 3dB.

5. The integrated indoor and outdoor coverage testing and evaluation system for power grid wireless networks according to claim 1, characterized in that, The intelligent evaluation module includes a feature encoding unit, a graph convolutional network unit, a spatiotemporal prediction unit, and an interpretable output unit. The feature encoding unit encodes signal strength, interference, terrain features, and node locations into a unified spatiotemporal feature tensor; The graph convolutional network unit is used to extract spatial topological correlations; The spatiotemporal prediction unit uses a multi-head attention mechanism based on the Transformer structure to capture temporal dynamic features; The interpretable output unit uses the Shapley value method to calculate the feature contribution and generate an index visualization chart to illustrate the impact of various environments or parameters on coverage performance.

6. The integrated indoor and outdoor coverage testing and evaluation system for power wireless networks according to claim 1, characterized in that, The optimization decision module includes a blind spot identification unit, a parameter optimization unit, and a control output unit. The blind spot identification unit automatically detects areas below -90dBm and generates a blind spot mask based on the signal strength distribution output by the intelligent evaluation module. The parameter optimization unit iteratively optimizes the base station transmit power, antenna elevation angle and site location based on the deep reinforcement learning algorithm (DDPG), and its reward function comprehensively considers coverage, interference and power consumption. The control output unit sends the optimized parameters to the corresponding devices through the power communication network management interface to achieve automated closed-loop adjustment.

7. The integrated indoor and outdoor coverage testing and evaluation system for power wireless networks according to claim 1, characterized in that, The system supports multiple types of mobile testing terminals, including handheld terminals, unmanned vehicles, and drones. The drone is equipped with obstacle avoidance radar and path planning module, and adopts a composite path optimization strategy based on AI and genetic algorithm to ensure that the obstacle avoidance distance is ≥2m and the test coverage is ≥95% in complex power line environment.

8. The integrated indoor and outdoor coverage testing and evaluation system for power wireless networks according to claim 1, characterized in that, The feedback update module includes a feedback acquisition unit, a model correction unit, and a knowledge update unit: The feedback collection unit is used to receive user evaluations, test reports, and network operation logs. The model correction unit fine-tunes the neural network parameters of the intelligent evaluation module based on feedback data, and uses lightweight adaptation technology to reduce the amount of parameter updates. The knowledge update unit synchronizes the feedback results to the environment modeling module and the optimization decision module to update the propagation model and optimization strategy.

9. The integrated indoor and outdoor coverage testing and evaluation system for power wireless networks according to claim 1, characterized in that, The system further includes a data fusion and edge computing node, deployed on a portable server in the field, equipped with a GPU acceleration unit, which is responsible for performing local feature extraction, data compression and temporary evaluation; The edge nodes communicate with the cloud via a 5G private network, with a data transmission latency of less than 50ms.

10. The integrated indoor and outdoor coverage testing and evaluation system for power wireless networks according to claim 1, characterized in that, The system supports a 3D visualization module, which is used to overlay and display coverage heat maps, blind spot distribution maps, and interference source location maps on the GIS platform, enabling interactive evaluation and display of the coverage quality of the power wireless network.

Images