The Development of a Flexible Sensor for Continuum Soft-Bodied Robots

In this thesis, we investigate, develop, and verify an approach to sense soft and flexible materials based on the use of a tomographic technique known as Electrical Impedance Tomography. The technique has been already used in different application domains to sense rigid, or semi-rigid structures. In this work we wanted to expand its application to situations that involve deformable domains in which the the underlying geometry can change shape. The use of such capability can be beneficial in applications related to the fields of robotics, rehabilitation, costumer electronic devices, but also in the new field of soft robotics. The technique used as transduction method allows to infer the internal structure of the domain under study by reconstructing its conductivity map. The use of EIT as sensing mechanism allows to reduce the number of rigid components in the systems, and enables the use of this sensing technique for large surfaces without the need of sensing based on “matrix” structures (e.g. two-dimensional capacitive or resistive sensors). By applying the technique to a material that changes its resistivity according to applied forces, it is possible to identify these changes and then localise the area where the force was applied. In order to fully control all the parameters during the measurements, we developed a low-cost driving/measurement electronics and an inverse solver software. We then tested the system in both two– and three-dimensional scenarios. For the two-dimensional case, we used a conductive textile as domain under study and we evaluate the sensor capabilities to detect forces applied over a single point, over multiple points, and changes in the underlying geometry. For the three-dimensional case, we developed a silicon-rubber conductive substrate by creating a mixture of silicon rubber and dissolved carbon fibres. The resulting volume was than used to simulate the elongation and the bending of a soft robotic arm, and to prove the paradigm “the robot body is the sensor”. The main contribution of the work is to introduce Electrical Impedance Tomography to the field of deformable structure. The results are all promising, and open the way for the application of such sensors in different robotic contexts where there are strong requirements on the deformability and adaptability of the technology used to develop these robotic devices.