Quantization Impact on the Accuracy and Communication Efficiency Trade-off in Federated Learning for Aerospace Predictive Maintenance

TL;DR

This study evaluates how different levels of quantization affect the accuracy and communication efficiency of federated learning models in aerospace predictive maintenance, demonstrating INT4 as an optimal balance.

Contribution

It provides a comprehensive analysis of quantization effects on FL accuracy and communication, highlighting INT4's effectiveness and FPGA implementation feasibility.

Findings

INT4 quantization maintains accuracy similar to FP32 with 8x reduced communication.

Naive IID partitioning underestimates variance; true Non-IID evaluation shows quantization instability.

INT2 quantization causes instability and non-reproducibility under heterogeneity.

Abstract

Federated learning (FL) enables privacy-preserving predictive maintenance across distributed aerospace fleets, but gradient communication overhead constrains deployment on bandwidth-limited IoT nodes. This paper investigates the impact of symmetric uniform quantization ($b \in \{32,8,4,2\}$ bits) on the accuracy--efficiency trade-off of a custom-designed lightweight 1-D convolutional model (AeroConv1D, 9\,697 parameters) trained via FL on the NASA C-MAPSS benchmark under a realistic Non-IID client partition. Using a rigorous multi-seed evaluation ($N=10$ seeds), we show that INT4 achieves accuracy \emph{statistically indistinguishable} from FP32 on both FD001 ($p=0.341$) and FD002 ($p=0.264$ MAE, $p=0.534$ NASA score) while delivering an $8\times$ reduction in gradient communication cost (37.88~KiB $\to$ 4.73~KiB per round). A key methodological finding is that na\"ive IID client partitioning artificially suppresses variance; correct Non-IID evaluation reveals the true operational instability of extreme quantization, demonstrated via a direct empirical IID vs.\ Non-IID comparison. INT2 is empirically characterized as unsuitable: while it achieves lower MAE on FD002 through extreme quantization-induced over-regularization, this apparent gain is accompanied by catastrophic NASA score instability (CV\,=\,45.8\% vs.\ 22.3\% for FP32), confirming non-reproducibility under heterogeneous operating conditions. Analytical FPGA resource projections on the Xilinx ZCU102 confirm that INT4 fits within hardware constraints (85.5\% DSP utilization), potentially enabling a complete FL pipeline on a single SoC. The full simulation codebase and FPGA estimation scripts are publicly available at https://github.com/therealdeadbeef/aerospace-fl-quantization.