On the Pose Estimation of an Untethered Microrobot via Magnetic Localization and Optical Tracking

The field of medical microrobotics has emerged as a promising approach for delicate medical interventions in anatomical regions where conventional approaches lack efficiency. Due to the fact that constrained anatomical workspaces offer limited visibility, in order to fulfill desired tasks, the precise tracking of untethered microrobots becomes a vital necessity and a critical challenge. Although there are studies in the literature to overcome said challenge, drawbacks include but are not limited to prolonged exposure to high magnetic fields, limited tracking workspace, and robustness. In light of this, the current paper proposes a hybrid tracking system utilizing both optical and magnetic tools to provide real-time pose estimation of an untethered microrobot operated within a closed workspace. The proposed system utilizes a custom-made Hall-effect sensor array apparatus that is capable of single-degree-of-freedom actuation to execute a planar sweeping motion and three low-cost motion capture cameras to track the apparatus pose in real time. In combining precise optical tracking, magnetic sensing, and the proposed artificial neural network (ANN)-based estimation approach, the system overcomes visibility limitations of optical motion tracking systems and workspace limitations in magnetic sensing systems. Hardware verification studies of the constructed setup are carried out utilizing a KUKA KR6 R900 SIX serial robot manipulator to simulate untethered microrobot actuation. During verification procedures, the sensor apparatus motion, end effector of the manipulator, and microrobot are streamed into a 3D Slicer software virtual environment to enhance visual feedback with an efficient surgical navigation approach. The proposed approach shows efficient results and sub-millimeter accuracy in microrobot tracking.