Pareto-Optimal Model Selection for Low-Cost, Single-Lead EMG Control in Embedded Systems

TL;DR

This paper evaluates various machine learning models for low-cost, single-lead EMG control on embedded devices, highlighting the Pareto-optimal solution balancing accuracy, latency, and memory constraints.

Contribution

It introduces a comprehensive evaluation of 18 models, including a novel hybrid network, and identifies Random Forest as the Pareto-optimal choice for resource-limited embedded EMG control.

Findings

Random Forest achieves 74% accuracy with optimal resource use.

Deep learning models like MaxCRNN reach 99% precision but are less resource-efficient.

Reliable EMG control is feasible on low-cost, embedded hardware.

Abstract

Consumer-grade biosensors offer a cost-effective alternative to medical-grade electromyography (EMG) systems, reducing hardware costs from thousands of dollars to approximately $13. However, these low-cost sensors introduce significant signal instability and motion artifacts. Deploying machine learning models on resource-constrained edge devices like the ESP32 presents a challenge: balancing classification accuracy with strict latency (<100ms) and memory (<320KB) constraints. Using a single-subject dataset comprising 1,540 seconds of raw data (1.54M data points, segmented into ~1,300 one-second windows), I evaluate 18 model architectures, ranging from statistical heuristics to deep transfer learning (ResNet50) and custom hybrid networks (MaxCRNN). While my custom "MaxCRNN" (Inception + Bi-LSTM + Attention) achieved the highest safety (99% Precision) and robustness, I identify Random Forest (74% accuracy) as the Pareto-optimal solution for embedded control on legacy microcontrollers. I demonstrate that reliable, low-latency EMG control is feasible on commodity hardware, with Deep Learning offering a path to near-perfect reliability on modern Edge AI accelerators.