I am a Research Scientist in the Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT. My research is focused on probabilistic methods for localization, tracking and mapping. In particular, I am the perception lead on MIT’s Darpa Robotics Challenge team – a multi-year multi-PI competition developing technologies for semi-autonomous humanoid robotic exploration and manipulation led by Russ Tedrake.
If you’re interested in my research you can find more information here.
- January 2015: I will take up a Lectureship in the Dept of Informatics in the University of Edinburgh. (Equivalent to Assistant Professor in the US).
- November 2014: We’ve open sourced the MIT state estimator – now called Pronto. Its not specific to humanoids and builds upon a very reliable IMU process model EKF. More details here.
- November 2014: In paper presented at the Humanoid conference, I described the state estimator we use on MIT’s Atlas robot. It achieves very low rate kinematic drift (2cm per 10 steps) and with LIDAR integration it avoids all drift. More details here.
- July 2014: Scott Kuindersma and I organised a workshop on the DRC at the Robotics: Science and Systems conference in Berkeley. The talks have been archived here.
- July 2014: Our submission to the JFR special issue on the DRC Trails has been accepted. The original submission can be found here.
- Dec 2013: We just competed in the 2nd stage of the Darpa Robotics Challenge in Homestead in Florida. After a grueling week under the sun we finished 4th out of a field of 16 from around the world. The competition demonstrated significant advances in the capability of humanoid robots to do the work of humans. There’s lots of details about the competition here and about our team here. This is the MIT team leads and I getting our award from DARPA PM Gill Pratt.
- June 2013: We finished 3rd in the simulation stage of the Darpa Robotics Challenge. As a result we will receive Boston Dynamic’s amazing Atlas robot and use it to compete in the physical stage in Miami, Florida in December 2013.
- May 2013: Hordur Johannsson‘s PhD research was nominated for a best student award at ICRA 2013. The paper was entitled “Temporally Scalable Visual SLAM using a Reduced Pose Graph” and is works with a dataset about an order of magnitude larger than previous work and in doing so solves some important scalability problems.
I was a PhD student in the Signal Processing Group at the Department of Engineering of the University of Cambridge. Simon Godsill was my supervisor with Andrew Blake as my industrial supervisor at Microsoft Research – who kindly funded my PhD studies. I completed my undergraduate degree in Electronic Engineering (BEng) in University College Dublin (Ireland) in June 2004 where I worked with Scott Rickard.