So, I got to a point where I needed to refactor the map planning in order to get live data projected on it. And that led me to a point where it made sense to put in the Visual Studio tabbed docking interface I knew I was eventually going to end up with.
I used a non-open source library I found after trying just about everything else except the $1000 ones (DevExpress, Telerik). AvalonDock, part of xceeds free and pay toolkit has horrible documentation and is awful to use.
DevZest WPF Docking
It went in way easier than I expected, but now learning the details with auto generated doc from code is going to take a while. I also started seriously looking at 2 other themed robot projects, that have business considerations, so things might slow down on RoboNUC more than I had expected, but in the end, a new outdoor version of RoboNUC would be the 3rd themed project, so it might also go to higher places than expected as well
Mike Hansow recently pointed me to a new bread board rig he saw;
It looks exactly like what I had been looking for. I had recently bought some logic ‘trainer’ off Amazon for 3 times the price and was so disappointed in it (malfunction out of the box, poor design, terrible build quality) I had dropped my intended project.
This Arduino form factor bread board has shipped and is on its way. As soon as I finish the other 3,123,423 projects that have since cropped up, I hope to get back to the original.
about open source, and GIT in particular;
“This page is not maintained anymore”
Yes, and apparently the one they point to is not either.
Open Source: Broken by default™
If you think about the personal computer’s history, there was a period where nerds soldered their own components and wired their own PC with switches and LEDs on the front panel – and that eventually grew into the commodity personal PC, Smart Phone, Tablets etc explosion of today.
Robots should be on the same path. I do not think there is question in anyone’s mind of a future with commodity robots performing factory labor, domestic labor and new tasks unimaginable today.
But there was a step in the middle for PCs. There was a time when components could be purchased and added to your PC without the need for a soldering iron. And that is when the PC explosion started. Even today, we usually plug in a USB cable, and thats it.
We need to push robotics to the same place.
Put away the breadboards and soldering irons. Allow the general public to concentrate on software, not common hardware. But allow an ample market for hardware innovators who are willing to fit in. Robotics no longer requires hardware knowledge, nor should it. The foundation has been poured, now it is time to actually get something done.
I put together a video and some code that demos a cheap IMU ($6) on an Arduino. Amaze your friends, have fun.
MPU2 <- code
How many chances do you have to get something reversed when building the drive system of a robot?
Answer: A lot.
Here are the slides I will show at the EC Robotics meetup tomorrow;
When WPF was killed at Microsoft, I said it was one of the stupidest tech decisions ever made. Well they fired the morons, and it’s back. WPF will be enhanced, and offered on other platforms going forward. Since WPF is my strongest skillset, this is great news to me.
Ok, I will be honest; this one hurt.
Do not follow the code comments exactly. They are there to trick you.
So anyhow, I now have the algorithms to do a super simple ICP Slam, and I think the next step is to convert to C#, do some optimization (LINQ) and try it on the real thing. And include what I have so far into the UI. What you are looking at is the lecture supplied ‘viewer’ application. It is showing raw scan data from the LIDAR. And corrected wall and landmark points, derived from calculating least squares transformations, 40 times each iteratively. I’m not a rocket scientist, but it seems to me like the raw scan data is pretty close itself, without all the fancy math. I assume this is headed somewhere I do not understand yet (it looks like a better box than the results do).
The iterative part is slow as hell. It will be interesting to see how C# compares.
But this has to be the most hilarious picture I have ever seen;
Never has failure been better captured.
Last time, they could not stream without problems. Kind of ironic for the brightest in a tech field don’t you think?
Team SCHAFT, who won the preliminary competition in 2013, has withdrawn to pursue commercial interests. One team, KAIST, out of 11 remain self funded.