What is missing in MyzharBot to be a perfect outdoor robot?

MyzharBot wants to be autonomous, but mostly it wants to be an autonomous OUTDOOR robot… so what is missing in MyzharBot to be a perfect outdoor Robot?

The reply is really easy… MyzharBot needs two main improvements:

  • sensors that can allow it to see in 3D in full sunlight
  • a closed chassis capable to protect its electronics and motors

While the first improvement is not easy to solve (stereovision? outdoor laser? they are really expensive…), the second improvement is under construction and it will be presented at MakerFaire European Edition 2015 in october in Rome, if MyzharBot will be chosen as project to be exposed.

The chassis is under study and a 3D virtual model is growing. The new chassis will be closed, the tracks will not be directly connected to the main bases and, really important, the electronics will be placed on anti-vibration dampers. In the current MyzharBot v3 the tracks are really noisy and the vibrations generated at full speed can cause electronic failures, so I decided to use anti-vibration damper and I’m going to search for different tracks with a “more continuous” elements.

Naturally MyzharBot v4 will be powered by the Nvidia Jetson TK1, the only embedded ARM board that allows to perform real time computer vision elaboration thanks to CUDA… unless Nvidia will not release a new embedded board based on the new Tegra X1 processor…………….

Actually the 3D virtual model looks like the one in the picture… but the things are quickly changing.

If you are interested you can discuss about the news on MyzharBot at Officine Robotiche 2015 in Rome the next weekend. MyzharBot v3 will be exposed, its current capabilities will be shown in a few demos and I will have a talk about “SLAM – Simultaneous Localization and Mapping”, to explain with really simple word what is SLAM problem, how it can be solved and why it is really important for autonomous robots.

So lets see at Officine Robotiche 2015… I’m waiting for you in Rome.

PS to be honest there is a good 3D sensor that can solve also the problems for the first improvement. Since GTC2015 I’m following the Zed stereo camera made by StereoLabs… it is a really promising sensor and I can’t wait to see it in action. Its performances are really interesting and the Nvidia Jetson TK1 can natively manage the data provided by this new sensor.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.