ROSAIA explores how a flexible, continuously shaped robot can sense, learn and act safely in complex physical environments. The project brings together mechanics, visual feedback, control and artificial intelligence in one platform.
The challenge
Continuum robots can reach and interact in ways that rigid mechanisms cannot, but their flexibility makes modelling, calibration and control difficult. The system has to account for coupled actuation, changing configurations and uncertain interaction with the environment.
What I work on
My contribution spans robot control, visual sensing, learning-based methods and the integration of software with the physical ROSAIA platform. Depending on the public project scope, this includes shape control, reinforcement-learning experiments, multi-arm identification and calibration.
Demonstration
The following video shows the ROSAIA four-arm platform and portable two-arm demonstration.
Why it matters
The long-term aim is to make flexible robots more useful and predictable in applications where access, compliance and safe interaction matter. ROSAIA is a testbed for studying the connection between intelligent algorithms and real mechanical behaviour.
This page presents publicly shareable work from the ROSAIA project.