A new way of looking at Embedded Systems

Started by Emanuel_Allely 2 years ago3 replieslatest reply 2 years ago102 views
Hi guys! My name is Manu,

I'm a new member of the forum, but happy to be here! Hope I'll be be helpful! :)

My friends (Nicolas Rabault and Simon Baudry) and I are working in Robotics/Electronics for a a long time (Aldebaran, Softbank Robotics, Mediatek, INRIA, Makingbot, Pollen Robotics). We are completely passionate about Embedded Systems!

During these years, we coped with a lot of technical struggles due to the complexity of all the ecosystems and because of a lack of modularity/universality. We got the feeling that we lose a lot of time re-building existing stuff or redeveloping something already existing...

That's why we are currently building a new way of designing embedded systems: Luos! We want to duplicate what microservices and API allow in the software world, in the Electronics world. Luos is like Docker for Electronics. It encapsulates drivers and embedded applications as modules and creates a network between all of them across all boards of a device.

Besides topology detection, hot plug and play, scalability and universality, what makes Luos different is that it creates a network between apps and functionalities, not between boards!

We just released a version of our libraries with examples, templates and documentation. To make this technology evolve and adapt to a bunch of situations, it needs to be used in a lot of projects and devices! We need your help because we need feedback! We would be really happy to discuss with you about our technology and project. If you're interested have a look to our repo on Github.

Thanks for your feedback!
See you,

[ - ]
Reply by MichaelKellettApril 10, 2020

If we filter out the hype and cut to the essence of this it's a set of Python libraries and some trivia hardware modules (like a pot on a little board with a connector).

As far as I am concerned Python and embedded are mutually exclusive. Hanging a load of cheap modules on an RPi is so NOT how decent embedded systems are designed.

This is the kind of approach which puts a linux computer inside a security camera and then spends the next 10 years trying to make it hack proof.

Some one, somewhere will love it, but not me.



[ - ]
Reply by Emanuel_AllelyApril 10, 2020

Hi Mickael Kellet,

Thank you for your interesting reply. I agree with you, decent embedded system can't be based on python. People using our high level (python) way are most of the time people making prototypes or really complex machines without strict real time constraints. This way interests people who are not into embedded.

All the boards we have are just voluntarily simple examples to demonstrate or technology.

We have another way of using Luos for people who are into embedded using only C/C++. Python is completely optional and not needed, you can use it with complete bare metal C coding. In our documentation we call it "[low level](". If you haven't already read this part it could interest you.


[ - ]
Reply by olzekeApril 12, 2020

I hope you all have lots of stamina.  Universal anything is hard to accomplish.

Your goal appears to be a mix of Arduino/mbed/++ .  Communications standards will

be a tough one to crack.

I'll give it a shot; as I do some ARM beta software testing for

I have wanted to do some Arduino stuff in a Home Automation scenario.

The Home Automation scene is fairly locked in, from my POV.

Good Luck to y'all

Gary Olzeke51