Karl Granström: Research

The references [J#], [C#], and [R#] below refer to the list of publications.

---

Interests

Karl Granström's main research interest is sensor fusion, especially the simultaneous localization and mapping (SLAM) problem and multiple target tracking (MTT) problem.

For the SLAM problem, Karl Granström has developed a method for loop closure detection. The method is based on laser range data, and is based on the machine learning method AdaBoost, see papers [C2,C4,J1]. The work has been performed in collaboration with Thomas B. Schön, and Fabio T. Ramos and Juan I. Nieto at the Australian Centre for Field Robotics.

For the MTT problem, Karl Granström has considered extended target. An extended target is a target that may generate any number of measurements per time step, as opposed to a point target which may generate at most one measurement per time step. For multiple extended target tracking, Karl Granström has developed two implementations of a PHD filter for extended targets, see [C3,J2,J3,R2], and he has developed a method for shape estimation for laser range sensors [C7]. For extended targets that are modelled using Koch's random matrix framework, Karl Granström has considered mixture reduction [C10], combination and spawning [J4], and prediction [J5,R1]. He has also developed a method for estimation of the measurement generating Poisson rate, see [C9]. The work has been performed in collaboration with Christian Lundquist, and Umut Orguner at the Department of Electrical and Electronics Engineering at the Middle East Technical University.

Karl Granström attended SLAM Summer School in January 2009, hosted by The Australian Centre for Field Robotics (ACFR). He has presented his research at the IEEE International Conference on Robotics and Automation (ICRA) in 2009 and 2012, at the IEEE/RSJ International Conference on Robots and Systems (IROS) in 2010, and at the International Conference on Information Fusion (FUSION) in 2010, 2011 and 2012.

---

Awards

Karl Granström is the winner of the following awards:

First Runner Up to the Best Student Paper Award

For the paper On the Reduction of Gaussian inverse Wishart Mixtures, presented at the 15th International Conference on Information Fusion in Singapore, 2012.
Award shared with Umut Orguner.

Third Best Conference Paper Award

For the paper Extended Target Tracking with a Cardinalized Probability Hypothesis Density Filter, presented at the 14th International Conference on Information Fusion in Chicago, IL, USA, 2011.
Award shared with Umut Orguner and Christian Lundquist.