Abstract: Factor graphs have become a popular tool for modeling robot perception problems. Not only can they model the bipartite relationship between sensor measurements and variables of interest for inference, but they have also been instrumental in devising novel inference algorithms that exploit the spatial and temporal structure inherent in these problems. I will overview some of the inference algorithms and present two specific applications: Simultaneous localization and mapping for underwater robots and state estimation for aerial robots. For state estimation I will introduce a novel fixed-lag smoother for visual-inertial odometry. I will also give a brief overview of factor graphs in the context of other robot perception problems.

Bio: Michael Kaess is an Assistant Research Professor in the Robotics Institute at Carnegie Mellon University (CMU). His research focuses on probabilistic methods for robot perception, in particular efficient algorithms for navigation, mapping and localization. Prior to joining CMU, he was a Research Scientist and a Postdoctoral Associate in the Computer Science and Artificial Intelligence Laboratory (CSAIL) at the Massachusetts Institute of Technology (MIT). He received the Ph.D. and M.S. degrees in computer science from the Georgia Institute of Technology. He was one of the two runner-ups for the 2012 Volz dissertation award for the best U.S. Ph.D. thesis in robotics and automation, and also received three runner-up best paper awards (ICRA 2011, 2013, 2016). He serves as Associate Editor for IEEE Transactions on Robotics.

Host: Katharina Muelling

Contact: Stephanie Matvey (smatvey@andrew.cmu.edu)