Real-Time Radio Coverage Map Generation Using GAN and Federated Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for generating radio maps in 5G/6G wireless networks are inefficient due to high computational costs and the need for frequent updates, which drain resources and lead to inaccurate path loss predictions, resulting in poor user experience and system throughput issues.
Innovation Solution
A method using a generative adversarial network (GAN) for real-time radio coverage map generation, aggregating data from multiple sources like satellite images, drones, and geospatial maps, and employing federated learning to train neural networks for efficient and accurate path loss estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If frequent radio map updates are performed using traditional measurement collection methods, then radio map accuracy is improved, but computational cost and resource consumption increase significantly
Solution Approach 1:
The patent creates virtual copies of radio map data through generative adversarial networks. The GAN generates synthetic radio map samples that replicate the statistical properties and spatial characteristics of real measurements, enabling accurate radio environment representation without performing expensive physical measurements or complex ray-tracing simulations for every update scenario.
Solution Approach 2:
The patent performs preliminary training of the GAN model using collected measurement data once, storing the learned radio environment characteristics. Subsequent radio map updates are generated by the trained model without requiring new measurement collections or heavy computations, enabling fast updates while consuming minimal resources.
2Measurement precision
If complex path loss models are used to improve prediction accuracy, then radio map accuracy is improved, but device complexity and computational resources increase
Solution Approach 1:
Instead of implementing complex physical models like ray-tracing or stochastic models in the deployed system, the patent pre-computes these complex scenarios during GAN training and captures their effects in the neural network weights. The trained GAN model then replicates the behavior of these complex models through simple neural network operations, achieving comparable accuracy with much lower runtime complexity.
Solution Approach 2:
The patent replaces traditional mechanical/computational physics-based models (ray-tracing, stochastic simulations) with a data-driven neural network model. The GAN learns the underlying patterns of radio propagation from training data and uses this learned knowledge to generate predictions, substituting heavy computational physics with efficient neural network inference.
3Measurement precision
If traditional measurement-based radio map generation is performed, then radio map accuracy is improved, but update time and productivity decrease
Solution Approach 1:
The patent performs the computationally intensive work of learning radio environment characteristics in advance during the GAN training phase. Once trained, the model can generate updated radio maps instantly by processing current input data through the pre-learned neural network, enabling rapid updates without repeating expensive measurement collections or simulations.
Solution Approach 2:
The GAN generates synthetic radio map samples that copy the essential characteristics of real measurements. This allows the system to produce accurate radio map updates by generating synthetic data rather than collecting and processing real measurement data, dramatically speeding up the update process while maintaining accuracy.
4Area of stationary object
If base stations are installed exponentially to meet user demand, then network coverage is improved, but coverage holes and resource wastage increase
Solution Approach 1:
The patent uses real-time handover information and geospatial data as feedback inputs to the GAN model. The model analyzes current network conditions, user movements, and base station configurations to generate updated radio maps that reflect actual coverage patterns. This feedback loop enables the system to identify coverage holes dynamically and provide information for optimized base station deployment.
Data Source
AI summary
Embodiments herein provide a method for generating a real-time radio coverage map in a wireless network by a network apparatus. The method includes: receiving real-time geospatial information from one or more geographical sources in the wireless network; determining handover information of at least one user equipment (UE) in the wireless network from a plurality of base stations based on the real-time geospatial information; and generating the real-time radio coverage map based on the handover information of at least one UE and the real-time geospatial information.