Keywords

Brain Machine Interfaces, Electromyography, Texture Recognition, Unary Computing, Alignment

Subject Categories

Biomedical | Biomedical Engineering and Bioengineering | Electrical and Computer Engineering

Abstract

Human-machine interfaces (HMIs) are poised to transform rehabilitation and assistive technologies, including prosthetic control, robot teleoperation, and immersive virtual and augmented reality. Despite substantial advances, the realization of robust and adaptive closed-loop systems capable of natural interaction remains an open challenge. Contemporary HMIs are constrained by two fundamental barriers: unreliable decoding of user intent and computationally expensive sensory feedback. Biosignals, including surface electromyography and ultrasound imaging, have emerged as promising modalities for decoding movement intent. However, their reliability often deteriorates under real-world conditions due to limb position variation, electrode shift, cross-day variability, and individual differences. Most existing work addressing these limitations has been validated on individuals without limb differences, limiting clinical translation. Advances in tactile sensing have enabled the acquisition of high-dimensional sensory information, introducing substantial computational, communication, and energy bottlenecks for real-time deployment on wearable and edge devices. This dissertation addresses both barriers by developing methods for robust intent decoding and edge-effcient sensory processing, grounded in two hypotheses: that motor intent is encoded as an invariant latent representation that persists across variable biosignal acquisition conditions; and that accurate sensory perception does not necessitate high numerical precision provided that the latent structure of sensory features is preserved. We develop a one-shot alignment framework based on Multi-set Canonical Correlation Analysis to identify a shared latent structure across biosignal modalities, acquisition conditions, and individuals with upper limb differences. We introduce unary-computing-based stochastic edge neuromorphic sensing (uSense), a fast Fourier transform (FFT) framework based on unary computing for near-sensor tactile sensing. These contributions offer a principled path toward closed-loop HMIs in smart wearables and next-generation assistive technologies.

Completion Date

2026

Semester

Summer

Committee Chair

Dr. Mohsen Rakhshan

Degree

Doctor of Philosophy (Ph.D.)

College

College of Engineering and Computer Science

Department

Electrical and Computer Engineering

Format

PDF

Document Type

Dissertation

Language

English

Release Date

8-15-2028

Download

Available for download on Tuesday, August 15, 2028

Share

COinS
 

Accessibility Statement

This item was created or digitized prior to April 24, 2027, or is a reproduction of legacy media created before that date. It is preserved in its original, unmodified state specifically for research, reference, or historical recordkeeping. In accordance with the ADA Title II Final Rule, the University Libraries provides accessible versions of archival materials upon request. To request an accommodation for this item, please submit an accessibility request form.