MINES ParisTech CAS - Centre automatique et systèmes

Optical-Inertial Tracking System with High Bandwidth and Low Latency

Authors: Göntje C Claasen, Philippe Martin, Frédéric Picard, in Recent Advances in Robotics and Automation, pp 171-181, Springer Berlin Heidelberg, DOI: 10.1007/978-3-642-37387-9_13
We propose an optical-inertial tracking system for a servo-controlled handheld tool in a computer-assisted surgery system. We present a mathematical system description and a data fusion algorithm which integrates data from optical and inertial sensors. The algorithm is a right-invariant Extended Kalman Filter (EKF) which takes into account system symmetries to improve the filter convergence. The tracking system has a high bandwidth thanks to the inertial sensors and a low latency thanks to a direct approach where sensor data is used directly in the data fusion algorithm without previous computations. Experimental data show that the optical-inertial system can indeed track a moving object.
BibTeX:
@Incollection{,
author = {Göntje C Claasen, Philippe Martin, Frédéric Picard},
title = {Optical-Inertial Tracking System with High Bandwidth and Low Latency},
booktitle = {Recent Advances in Robotics and Automation},
editor = {},
publisher = {Springer Berlin Heidelberg},
address = {},
pages = {171-181},
year = {2013},
abstract = {We propose an optical-inertial tracking system for a servo-controlled handheld tool in a computer-assisted surgery system. We present a mathematical system description and a data fusion algorithm which integrates data from optical and inertial sensors. The algorithm is a right-invariant Extended Kalman Filter (EKF) which takes into account system symmetries to improve the filter convergence. The tracking system has a high bandwidth thanks to the inertial sensors and a low latency thanks to a direct approach where sensor data is used directly in the data fusion algorithm without previous computations. Experimental data show that the optical-inertial system can indeed track a moving object.},
keywords = {}}