Human-in-the-loop Bayesian optimization of wearable device parameters

• Published in: PloS one Vol. 12; no. 9; p. e0184054
• Main Authors: Kim, Myunghee, Ding, Ye, Malcolm, Philippe, Speeckaert, Jozefien, Siviy, Christoper J., Walsh, Conor J., Kuindersma, Scott
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 19.09.2017; Public Library of Science (PLoS)
• Subjects: Adult; Algorithms; Ankle; Artificial intelligence; Bayes Theorem; Bayesian analysis; Biology and Life Sciences; Biomechanics; Computer and Information Sciences; Convergence; Energy expenditure; Energy Metabolism; Engineering schools; Exoskeletons; Female; Global optimization; Humans; Identification methods; Male; Medicine and Health Sciences; Metabolism; Methods; Noise; Optimal control; Optimization; Parameter identification; Physical Sciences; Physiology; Prostheses; Prosthetics; Research and Analysis Methods; Robotics - instrumentation; Scholarships & fellowships; Signal processing; Signal-To-Noise Ratio; Walking; Wearable technology; Young Adult; United States > US
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0184054

The increasing capabilities of exoskeletons and powered prosthetics for walking assistance have paved the way for more sophisticated and individualized control strategies. In response to this opportunity, recent work on human-in-the-loop optimization has considered the problem of automatically tuning control parameters based on realtime physiological measurements. However, the common use of metabolic cost as a performance metric creates significant experimental challenges due to its long measurement times and low signal-to-noise ratio. We evaluate the use of Bayesian optimization-a family of sample-efficient, noise-tolerant, and global optimization methods-for quickly identifying near-optimal control parameters. To manage experimental complexity and provide comparisons against related work, we consider the task of minimizing metabolic cost by optimizing walking step frequencies in unaided human subjects. Compared to an existing approach based on gradient descent, Bayesian optimization identified a near-optimal step frequency with a faster time to convergence (12 minutes, p < 0.01), smaller inter-subject variability in convergence time (± 2 minutes, p < 0.01), and lower overall energy expenditure (p < 0.01).