Abstract

This thesis investigates human sensorimotor control using three complementary sets of experiments. These examine contextual motor learning, aging-related motor performance, and human balance.First, a series of experiments examined how different sensory modalities influence motor memory formation during dynamic learning. Using a robotic manipulandum to implement an interference task, we assessed how the contextual effect of visual and passive lead-in movements decays over time, specifically as a function of dwell time between the movement cue and the required action onset. Results revealed a modality-dependent pattern of decay: while both cue types enabled dual adaptation to opposing force fields, passive cues preserved contextual influence at significantly longer dwell times than achieved using visual cues. These findings underscore fundamental differences in how the motor system integrates proprioceptive versus visual information to shape predictive control strategies.Second, we explored age-related differences in sensorimotor performance through a set of tasks using the vBOT robotic system. Healthy younger and older adults completed bimanual coordination and unimanual reaching tasks under varying dynamic conditions. Key performance metrics, including movement duration, force generation, and response onset time, reveal consistent age-related declines, but with apparent individual variability.Finally, a series of studies was conducted to investigate the dynamics of inverted pendulum systems and the control mechanisms applied to balance this inherently unstable configuration. A novel inverted pendulum system was developed and stabilized using a dual-mode gain-scheduled state feedback controller. The hardware platform featured modular mechanical components, adjustable pendulum lengths, and an automated tapping mechanism for controlled perturbations. Distinct state-space models were derived for position and velocity control, using linear quadratic regulators (LQR) to compute optimal gains for each mode. A DIN rail panel unit was custom built to implement the controller. Experimental results demonstrated robust stabilization and task-dependent adaptability, validating the control approach for real-world balancing tasks. To explore how humans perform such balancing, participants were asked to manually balance similar physical pendulums of three different lengths for up to ten seconds. Results indicated that participants achieved greater stability and longer balancing durations with longer pendulums, highlighting significant behavioral differences across the three configurations.Together, these studies contribute a unified perspective on sensorimotor control, ranging from engineered systems to human adaptation of dynamical systems, and their limitations. As such, the work advances our understanding of real-time control, context-dependent learning, and aging, and offers validated tools and methodologies for future research in human-robot interaction and neurorehabilitation.

Awarding Institution(s)

University of Plymouth

Award Sponsors

UKRI‐ESRC

Supervisor

Ian Howard, Gunnar Schmidtmann, Jonathan Marsden

Keywords

Sensorimotor control, Motor learning, lead-in movements, Motor memory, Force field adaptation, Robotic manipulandum, age-related motor decline, Aging, Inverted pendulum, state feedback control

Document Type

Thesis

Publication Date

2026

Embargo Period

2026-06-01

Deposit Date

June 2026

Creative Commons License

Creative Commons Attribution-NonCommercial 4.0 International License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License

Download

Share

COinS