Aerial-Assisted Federated Learning UE Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wireless communication networks face challenges in efficiently integrating aerial platforms into beyond 5G networks, particularly in enhancing network performance and increasing the number of participating devices in federated learning processes due to limitations in terrestrial infrastructure and device constraints.
Innovation Solution
The implementation of a system and method that selects user equipment for federated learning based on terrestrial Channel Quality Indicator (CQI) and trajectory information, triggering an aerial link to enable dual communication and expand network coverage and capacity using Low-altitude Aerial Platform (LAP) based aerial cells, thereby enhancing the availability and participation of user equipment in the federated learning process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If aerial platforms are integrated into beyond 5G networks to expand coverage and capacity, then network efficiency and availability are improved, but device complexity and integration difficulty increase
Solution Approach 1:
The patent introduces an aerial platform as an intermediary component between ground-based user equipment and the federated learning server. This aerial platform acts as a mediator that relays communication signals, enabling UEs in coverage areas with poor terrestrial signal quality to participate in federated learning. The aerial platform integrates multiple functions including signal relay, UE selection based on CQI metrics, and coordination with the FL server, thereby improving network availability without requiring complex modifications to existing UE devices.
2Measurement precision
If the number of participating devices in federated learning is increased to improve model training quality, then learning accuracy is improved, but communication overhead and network load increase
Solution Approach 1:
The patent implements selective participation in federated learning based on local channel quality conditions. Specifically, user equipment with poor terrestrial CQI but good aerial link quality are identified and enabled to participate through the aerial platform, while UEs with already good terrestrial connectivity continue to use ground-based communication. This local quality-based selection optimizes the balance between increasing participant diversity for better model training and minimizing unnecessary communication overhead from UEs that would contribute little value.
Solution Approach 2:
The system dynamically changes communication parameters including selecting UEs based on CQI thresholds, adjusting which devices participate in FL rounds, and switching communication paths (terrestrial vs. aerial) based on real-time channel conditions. These parameter changes enable the system to adaptively optimize the trade-off between having more participants for better model accuracy and reducing communication overhead by excluding UEs with poor channel conditions regardless of their data quality.
3Adaptability or versatility
If dual communication with aerial and terrestrial links is enabled to expand network coverage, then adaptability is improved, but device complexity and processing requirements increase
Solution Approach 1:
The patent implements dynamic dual communication capability where user equipment can switch between terrestrial and aerial communication links based on real-time Channel Quality Indicator (CQI) measurements and network conditions. The system dynamically selects the optimal communication path for each UE participating in federated learning, enabling adaptability without requiring permanent dual-link hardware configurations. This dynamic approach allows UEs to have aerial communication capability activated only when needed, reducing overall device complexity while maintaining communication flexibility.
Data Source
AI summary
The present disclosure provides a system and a method for aerial-assisted federated learning at a Federated Learning (FL) server. The method includes receiving a plurality of parameter sets and trajectory information indicating a coverage range by the FL server from a plurality of User Equipment (UEs) and an aerial cell, respectively. Further, the FL server selects at least one UE from the plurality of UEs based on the received plurality of parameter sets and the received trajectory information. Additionally, the FL server triggers an activation of the aerial link between the aerial cell and the selected at least one UE to include the selected at least one UE to a set of federated UEs associated with the FL server.