March 7, 2018

The students had not seen Minerva before, so I displayed Thrun's video to accompany the lecture Localization II.
Localization II had a bit strange build, with examples of different map data-structures and their effect on localization, but with in between Kalman and Markov localization. The main difference seemed to be continues vs discrete, but the difference is also single hypothesis vs multiple hypotheses.

February 1, 2018

Toto asked for earlier material. On Blackboard links to the localization presentations 5a-5e of ETHZ are given. These are the presentations from my course page latest updates are links to the 2016 presentations are given, although the course was also given at ETZH in 2017.
The PowerPoints from my C:-drive are mainly from 2012:
- 5a - Nov 2015
- 5b - Nov 2012
- 5c - Nov 2012
- 5d - May 2013
- 6 - Dec 2013
- 7 - Dec 2013
I presume that most of those presentation are based on the PowerPoints van Davide Scaramuzza (update till 2010) which accomponied the book. De last PowerPoints published from ethz (before the start publishing pdf) is 2013:

A lot of lessons at Robotics Academy, based on Robotics, Vision and Control from Corke.

The projects of the NanoDegree Robotics for term 1 seem to be:
1. The Introduction to Robotics project.
2. The Kinematics project.
3. The Perception project.
4. The Controls project.
5. The Deep Learning project.
So the projects of the NanoDegree for term 2 are not disclosed yet.
Started with conda env create -f environment.yml. Activated the environment with source activate RoboND.
Yet, as ROS-user I should read warning. The trick is to unset PYTHONPATH before source activate RoboND. For me that is anyway no problem, because ros is not in my bashrc, but always called manually.

Jana Kosecka's course has also some nice homework assignments for Siegwart's book.