MSc Thesis: Surface-Adaptive Anisotropic Force Control
Jan 2026 – present · ongoing
Moog HapticMaster, Python, admittance control, online surface estimation
Ongoing MSc thesis (BodyHaptics group, TU Delft, Dr. Arno Stienen): contact-force regulation for assistive robotic tools in industrial work — hold a set contact force on an unknown curved surface while making tangential sliding effortless, estimating the surface by feel alone. Defense: Oct 2026.
What it is
A worker holds a robot arm; the robot arm holds a sanding tool against a curved surface. The control strategy must, per the research question published in the systematic review preprint: “maintain desired contact force along the surface normal while providing power-assist in tangential directions, when surface geometry is unknown and must be estimated online from force-motion interaction, to reduce operator muscular load during tool-handling tasks.”
The review decomposed this into five components — surface estimation, physical human-robot interaction, shared control, anisotropic force control, ergonomic validation — and found that 0 of 71 papers combine all five with genuine online surface estimation. This thesis builds in that gap: experimental work runs on a Moog HapticMaster (admittance-controlled), with a custom Python control loop implementing contact-force recovery and online surface-normal estimation.
It is the direct continuation of the Bilfinger/Shell field study that defined the problem space, and of the systematic literature review that mapped the control-strategy gap. Defense: October 2026.
Results (all numbers live here, verbatim)
- Published so far: the systematic review preprint (engrXiv, DOI 10.31224/7827) — its numbers live in the contact-force-review entry.
- Experimental thesis results are unpublished until the defense (October 2026). Quantified progress findings (estimator accuracy improvements, damping-compensation discovery) are recorded in thesis-progress-2026-04-08.pdf in this folder.