Research

Our research addresses various aspects to facilitate human-centred deployment of collaborative robots. Firstly, we want to ensure robots can be easily programmed—via demonstration or other natural interaction methods. Secondly, we want the resulting behaviours to be understandable to human users. Lastly, we aim to ensure all systems are safe and certifiable   

Our methods focus on programming systems without relying on data-driven simulation models, leveraging human priors instead. Moreover, we examine deployment aspects of these systems—from user-interface comparisons to acceptance—according to the current ISO standard, ISO 10218-2:2025 

Our research focuses on modeling aspects of serial and closed-chain robot manipulators. The aim is to support deployment of systems in varied environments—especially highly cluttered ones—and understand how their capabilities are affected.  

One current topic involves using polytopes to represent the velocity and wrench capabilities of a kinematic chain—both mathematically and visually for human users. We’ve developed methods to embed various constraints and ensure calculations remain continuous and locally differentiable.

While robot manipulators have typically converged on an anthropomorphic design resembling a human arm, alternative morphologies—such as humanoids or systems tailored for large-scale tasks—can be better suited to diverse environments. 

This research focuses on the design, analysis, and control of different robot morphologies involving multiple kinematic chains to optimise specific performance criteria.  

One such family of systems is cable-driven parallel robots. These systems can scale to cover large distances since the motors are grounded, while the cable arrangement maintains stiffness and full controllability. Our research focuses on modeling, design, and control of these systems, with applications in aerospace, construction, and infrastructure maintenance.