McMaster University
Convex Relaxation Methods For Optimal Electromagnetic Actuator Placement
Abstract
dc:description.abstractMicrorobots have significant potential for minimally invasive medicine, since they can access regions of the body that are difficult to reach using conventional instruments. Among the available actuation methods, magnetic actuation is especially attractive because it can remotely exert forces and torques on the microrobot, with electromagnetic systems often preferred for their controllability. In some systems, actuators are placed under the operating table to reduce obstruction of the workspace. However, determining how to arrange them remains a difficult challenge in this setting, since different arrangements can yield substantially different levels of actuation performance. This thesis studies the problem of optimally placing electromagnetic actuators for under-the-table systems. We formulate actuator placement as an E-optimal design problem within a Bayesian optimal design framework, where candidate actuator poses are evaluated through the spectrum of an actuation matrix constructed from magnetic field and field-gradient contributions. The resulting selection problem is relaxed to a continuous optimization problem, solved using Frank-Wolfe, and then converted into a feasible actuator arrangement using a constraint-aware adaptation of Madow sampling. Through a downstream simulation we show that improvements in the E-optimal objective correspond to improved actuation performance, and that our proposed approach offers a better alternative to placing actuators using naive methods or simple heuristics. The methods developed in this thesis are packaged into a customizable Python software library for optimal actuator arrangement.
Degree
thesis:*- Department dc:contributor.department
- Computing and Software
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Saad, Hussein
- Advisor dc:contributor.advisor
-
- Giamou, Matthew
Rights
- Language dc:language.iso
- en
Identifiers
dc:identifier.*- Identifier URI
- https://doi.org/10.71548/22419