CSI-Free AirComp Voting for Federated Edge Learning

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing federated edge learning (FEEL) systems face communication bottlenecks due to the need for channel state information (CSI) in over-the-air computation (AirComp) schemes, which are not trivial to develop and require substantial overhead, especially in multipath channels.

Innovation Solution

An AirComp scheme using frequency-shift keying (FSK) over orthogonal frequency division multiplexing (OFDM) subcarriers for voting, eliminating the need for CSI by separating votes on orthogonal resources and employing non-coherent detection, while incorporating randomization symbols to reduce peak-to-mean envelope power ratio (PMEPR).

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If over-the-air computation (AirComp) schemes are used for federated edge learning, then communication efficiency is improved, but channel state information (CSI) overhead and system complexity increase

Engineering Contradiction:

Improvecommunication efficiencyVSAvoidCSI overhead

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent extracts and removes the CSI requirement from the AirComp system by using non-coherent detection. Instead of requiring CSI for channel compensation, the system separates votes on orthogonal resources and detects them without coherent processing, effectively taking out the CSI overhead component while maintaining communication efficiency

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent introduces orthogonal resources (orthogonal subcarriers or time slots) as an intermediary mechanism. This intermediary allows votes to be separated and transmitted without requiring CSI, as the orthogonality provides inherent separation that eliminates the need for channel state knowledge at the receiver

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If AirComp schemes with channel inversion are used, then aggregation accuracy is improved, but device complexity and computational overhead increase

Engineering Contradiction:

Improveaggregation accuracyVSAvoidcomputational overhead

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent removes the channel inversion operation from the system by adopting non-coherent detection. Instead of computing inverse channel coefficients at each device, the system uses orthogonal separation and energy-based detection, extracting the essential voting function while eliminating the computationally intensive channel inversion step

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent replaces expensive channel inversion computations with simple, low-cost operations. Each device performs only local voting and transmits on orthogonal resources, using inexpensive energy detection at the receiver instead of complex coherent detection, effectively using cheap short-living computational objects

Inventive Principle:

Principle #27Cheap short-living objects (Disposable)

3Productivity

If votes are transmitted on the same resources, then bandwidth efficiency is improved, but interference and detection difficulty increase

Engineering Contradiction:

Improvebandwidth efficiencyVSAvoiddetection difficulty

Core Design Contradiction:

ProductivityVSDifficulty of detecting and measuring

Solution Approach 1:

The patent segments the voting transmissions by assigning different orthogonal resources (subcarriers or time slots) to different votes. This segmentation separates the votes in the frequency or time domain, making them easily detectable through simple energy measurement on each orthogonal resource without mutual interference

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS12572799B2Methods for reliable over-the-air computation and federated edge learning

Publication Date: 2026.03.10 UNIVERSITY OF SOUTH CAROLINA

AI summary

System and method for an over-the-air computation (AirComp) scheme for federated edge learning (FEEL) doesn't require channel state information (CSI) at the edge devices (EDs) or edge server(ES). The disclosure adopts the majority vote (MV) principle and defines multiple subcarriers and orthogonal frequency division multiplexing (OFDM) symbols for voting options, which reduces to frequency-shift keying (FSK) over OFDM subcarriers as a special case. Thus, FSK-based over-the-air computation is provided for federated edge learning without channel state information. Since the votes from EDs are separated on orthogonal resources, the scheme eliminates the need for truncated-channel inversion (TCI) at the EDs and allows the ES to detect MV with a non-coherent detector.