robos

Reading notes on Dellaert & Kaess, “Factor Graphs for Robot Perception”. Source notebook: rmath/factor_graphs_dellaert_kaess.ipynb — repo not public yet. My first exposure to SLAM was EKF and learned about cartographer during my internship at BNR over the summer. Got to know about GTSAM there while digging around the calibration stack. I mainly did the systemic error parameter calibration for the odometry. And here I was in my third semester digging into factor graphs when Im supposed to be writing a local planner for topological maps....

Source notebook: rmath/lie_algebra_sola.ipynb — repo not public yet, link will go live later. Micro Lie Theory - Sola et al I’ve always known how lie theory is used in SLAM. Had a pretty good understanding about how the $\boxplus$/$\oplus$ and $\boxminus$/$\ominus$(I like $o$ notation over $\Box$ notation ) operators are used. But neither of this led me to concretely explain the theory behind this. I came across this paper very recently....

Design notes for ranif, a from-scratch Rust Lie-group / factor-graph library. Source: docs/index.qmd. Yet another document I know, there have been a lot of documents specifically associated with just the Lie groups and the associated math required to do state estimation in robotics. But I feel nothing has been unnecessary and all the docs serve a specific purpose. In lieu of that this one is my internal walk through of how I went about doing this library....