Real-Time Multi-Object Magnetic Tracking for Multi-Arm Continuum Robots

Abstract

Multi-arm wire-driven continuum robots can provide dexterous maneuver during minimally-invasive surgeries (MIS). To ensure safety during MIS, it is critical to obtain feedback information such as poses of end effectors on all arms. However, precise and real-time pose information of multi-object cannot be obtained within human body due to high computational cost in current optimization approach of magnetic tracking systems. In this paper, we propose to track poses of multiple end effectors of a multi-arm continuum robot by permanent magnetic tracking technology. A permanent magnet is mounted at each distal end of the flexible arms and a sensor array is used to sample the magnetic field signals. The extended Kalman filtering algorithm is utilized to fuse the poses estimated from the kinematic model and the magnetic field model to realize multi-target positioning for surgical navigation. The proposed method is validated by experiments. The mean position error is 1.32 mm and the computing time is 1.5 ms. As a result, the proposed method can effectively provide accurate tracking results and meanwhile decrease the tracking latency for real-time applications.