Development of a novel angular sensor for medical tracking systems

Minimally Invasive Surgery procedures, which were, in the past, usually performed via manual intervention of human beings, are nowadays more often supported by robotic system, enabling to overcome the limitations that surgeons can encounter. Since the safety of these types of procedures strongly depends on their achievable precision, the sensors responsible for the tracking of the robotic system are required to be highly accurate and reliable. The MIRACLE (Minimally Invasive Robot-Assisted Computer-guided LaserosteotomE) project, which is conducted at the University of Basel, aims to develop a surgical platform including a robotic endoscope that allows contact-free bone surgery using laser light. Because state-of-the-art optical tracking systems require line-of-sight, these kinds of sensors are not a feasible option in the frame of the MIRACLE project. For this reason, part of the research of MIRACLE is focused on the development of a new tracking concept to control the position and the orientation of the laser end effector as well as the entire shape of the endoscope, inside the patient's body. A potential solution is based on an opto-mechanic angular sensor (shadow sensor), called ASTRAS (Angular Sensor for TRAcking System), which essentially consists of an image sensor, a shadow mask, a mirror, and an LED. Placing several ASTRAS at each joint of an articulated robot, and measuring the angles at these positions, allows to reconstruct the shape of the endoscope. However, ASTRAS is operation in the range of ±15°, limiting the proper utilization of the robot. Consequently, the topic of this thesis project was to design and characterize a new prototype of ASTRAS, extending its measurement range to 360°. The new version of ASTRAS is called ASTRAS360 and is based on the use of several mirrors simultaneously. Realized from sketch, using the 3D printing technology, this system uses color filters to encode the light and, therefore, allows to extend measuring range of the sensor. In parallel, the software responsible for the data analysis and angular estimation was also designed. To test the performance of ASTRAS360, a motorized characterization setup was built. The latter, consisting of a highly accurate optical rotary encoder, was used as calibration standard. ASTRAS360 resulted to be a highly precise absolute encoder system with a linearity error in the range of 10 arcsec.