High fidelity remote manipulation of microrobots by a new generation of electromagnetic control system

The proposed research intends to develop a generalized approach for the remote control of untethered effectors with less feedback required for stable control. Effectively, this approach doesn’t attempt to control effectors directly, instead aiming to control a weighted distribution of possible locations. In situations where imaging bandwidth is sufficient for stable control, the controlled distribution collapses to a single point, instead serving as a safety net in the event that an effector gets “lost” or otherwise cannot be resolved against the surrounding tissue. This approach translates seamlessly to controlling swarms of effectors, which are effectively weighted distributions of effectors themselves. To achieve this, a simplified polynomial form has been developed to express the controlling electromagnetic potentials, which serves as a simplified, universal interface across different controllers, effectors and designers. This inherent portability is a crucial step towards extending Machine Learning and AI into nanorobot design and control, allowing an initial human-designed model of field-effector and effector-environment interactions to be progressively refined by identifying patterns in the model’s random diffusion term.

Faculty Supervisor:

Hamed Shahsavan

Student:

Partner:

Universität Stuttgart

Discipline:

Engineering

Sector:

Education

University:

University of Waterloo

Program: