Today we have a special edition of the newsletter, in which I will try to highlight all meetups we’ve had in this season. Sine February we’ve had some amazing presenters share their knowledge and experiences. In every single meetup I’ve learned something interesting. Some of the conversations we’ve had in the Q&A parts were also very captivating. I’ll be looking forward to continuing with these meetups in the future. Whether you did or did not watch any of the meetups I would highly appreciate your feedback on them, so we can make them as useful as possible when we return to organizing them in Autumn.
Apologies for no audio here, both my main and secondary recordings did not capture sound. In this presentation Kenneth showed us his iOS app for crating RGB-D datasets. The app has since been released and Kenneth had wrote an in-depth blog post about it.
The 2nd meetup was special - it’s the only meetup we’ve had two presentations. In the first presentation Ignat Georgiev and Adrian Brandemuehl shared their experience working in Formula Student Driverless and Josh Whitely presented the work of Autoware Foundation. In both presentations there are some very good points made on processing huge amounts of data, lattice planners, importance on simulation, heat breaking your computers and more.
Looking at the video statistics this is the most popular talk. In this blog post, Victor Klemm shared insights on Ascento, a wheeled-bipedal robot Victor has been working on with his team. The presentation contains some very good lessons on robot making, and Victor unveils some hints on the future iteration of this robot.
In the 4th meetup, I did an AMA about my work as a robotics consultant. I’m not sure you will find something interesting there unless you are thinking about going solo or hiring some extra help for your projects.
In this meetup, Darko showed us the current state of Webots. The bits that I found particularly interesting is the support of GitHub Actions and Robot Benchmark that in itself can be an amazing learning tool for robotics programming in the classroom.
Here Loy van Beek presents his experience taking part at RoboCup@Home with TechUnited team. My highlights from this talk include working with state machines, object recognition, 3D world model and constraint-based navigation.
In this talk, Josh introduced mjbots, his project where he develops some open-source hardware for robotics such as BLDC controllers, power distribution boards and even massive servos. All this hardware can be used for building quad A1 quadruped robot. Josh is updating his blog quite frequently, so if you are interested in his work, I recommend following it.
In this meetup, James talked about managing and running robot fleets and how Formant helps with that. James did a really good job presenting the system and showing us some demos, including the features that recently enabled remote users to play with Spot. James had lots of interesting robotics business insights to share, and I appreciate the quality of advice he gave.
Steve Macenski gave us an intro to Nav2 project and explained how the new navigation stack for ROS2 is structured. The key takeaway for me: using BehaviorTrees for structuring the complex navigation behaviours is amazing. If you feel like you are missing out by not actively using Nav2 then welcome to the club!
In May Roland took us back to thinking about self-driving cars and delivered the most interactive presentation up to date. I enjoyed guessing the self-driving levels of cars and the discussions we had about safety.
Davide Faconti delivered a very good presentation on writing quality software and presented his recommended workflow for optimizing C++ code. It’s amazing how by just changing a single lince of code you can hugely improve your software’s speed. In the Q&A we’ve gone a bit off-topic and started chatting about BehaviorTree.CPP - Davide’s BT implementation. If you enjoy Davide’s work, I highly recommend supporting him on GitHub. I think we need to do a better job supporting open source developers.
This one boggles my mind. Why Achille seems to be the only person outside of academia talking about cuspidal manipulators? To me, this sounds like a big deal, and if I’m ever on a market for a robot arm checking for cuspidality will be very high on my list of things to check.