Characterisation and State Estimation of Magnetic Soft Continuum Robots

Minimally invasive surgery has become more popular as it leads to less bleeding, scarring, pain, and shorter recovery time. However, this has come with counter-intuitive devices and steep surgeon learning curves. Magnetically actuated Soft Continuum Robots (SCR) have the potential to replace these devices, providing high dexterity together with the ability to conform to complex environments and safe human interactions without the cognitive burden for the clinician. Despite considerable progress in the past decade in their development, several challenges still plague SCR hindering their full realisation. This thesis aims at improving magnetically actuated SCR by addressing some of these challenges, such as material characterisation and modelling, and sensing feedback and localisation. Material characterisation for SCR is essential for understanding their behaviour and designing effective modelling and simulation strategies. In this work, the material properties of commonly employed materials in magnetically actuated SCR, such as elastic modulus, hyper-elastic model parameters, and magnetic moment were determined. Additionally, the effect these parameters have on modelling and simulating these devices was investigated. Due to the nature of magnetic actuation, localisation is of utmost importance to ensure accurate control and delivery of functionality. As such, two localisation strategies for magnetically actuated SCR were developed, one capable of estimating the full 6 degrees of freedom (DOFs) pose without any prior pose information, and another capable of accurately tracking the full 6-DOFs in real-time with positional errors lower than 4~mm. These will contribute to the development of autonomous navigation and closed-loop control of magnetically actuated SCR.