Robot Operating System To Officially Support ARM Processors 33
DeviceGuru writes: The Open Source Robotics Foundation (OSRF), which maintains the open source Robot Operating System (ROS), has announced its first formal support for an ARM target. The organization will add support for the Qualcomm Snapdragon 600, a smartphone-oriented, quad-core, Cortex-A15-like system-on-chip running up to 1.7GHz. The Linux version of ROS for Snapdragon 600 will be available in Q4 of this year, with the Android version due in the first half of 2015. The OSRF will test, refine, and fully integrate support for the ARM instruction set architecture into ROS development efforts. OSRF will also perform ongoing maintenance to support ROS on the Snapdragon 600.
Linux version first? (Score:2)
Surely the Android version would be the more appropriate.
Re: (Score:2)
If ROS is the operating system (I'm assuming the userland from context, never heard of this before now) and Linux is the kernel, then what does an "android version" even mean? Its not evident from TFAs.
And what the heck have the robots been running on up till now? x86-windows?
Re: (Score:1)
ROS is not an OS, its just middleware that runs on Linux
Re: (Score:2)
In their introduction they use the term meta-operating system.
ROS is an open-source, meta-operating system for your robot. It provides the services you would expect from an operating system, including hardware abstraction, low-level device control, implementation of commonly-used functionality, message-passing between processes, and package management. It also provides tools and libraries for obtaining, building, writing, and running code across multiple computers. http://wiki.ros.org/ROS/Introd... [ros.org]
Re: (Score:1)
I don't see the need for a Robot Operating System anyway. Wouldn't it be more efficient to write everything in assembly since there's only one Data?
Re: (Score:3)
Android is Java. Although ROS supports multiple language bindings the core is written in cpp.
Google has a rosjava port. Multiple machines can run in a single ROS system, so the rosjava port is usually used to create Android UI applications to control or view information in a multi-machine ROS system.
ROS development is tied very closely to Ubuntu. Some things work on Mac OSX, but that's because it's Unix based and the built system was easier to adapt. Windows support is practically non-existent because to bu
Re: Linux version first? (Score:2)
How is that even possible? Android is essentially a Linux distribution. It's like saying you want it to work on Sprint before it works on a phone network. Though come to think of it Sprint might not be one given their spotty coverage.
We'll of course it will (Score:4, Funny)
Many robots have arms - the OS needs to be able to control their processes.
Re: (Score:2)
No legs? No problem (Score:2)
Re: (Score:2)
Hmmm, sounds like a commercial opportunity there. The LEG processor? It could be a hit!
rgb
Re: (Score:2)
You insensitive clod! (Score:2)
signed,
R2D2
Re: (Score:2)
A robot OS has to be able to control multiple tentacles simultaneously otherwise the Japanese won't use it.
Re: (Score:2)
Especially the potty tentacle.
Re: (Score:2)
Robots are up in ARMs about the new controllers for their arms.
Re: (Score:1)
That's boring; give them tentacles instead. Is there a TENTACLE chip?
Qualcomm must be funding it. (Score:2)
ROS has had portability issues for a long time, but those issues have been getting a lot of attention for at least a couple of years. The build system is much better, for one thing. It should be acknowledged that a lot of people (hobbyists, mainly) have been putting in signficant effort on making an ARM port possible for some time, Raspian on the RasPi being the main target. So while it is a good thing, on balance, that Qualcomm is putting in some money to make it happen, I'm disappointed that the work al
Re:Qualcomm must be funding it. (Score:5, Informative)
Heres the thing about R/T robotic control systems in industrial robotics. The tasks that they run are well defined repetitive tasks that don't change much. The kinds of robotics that we see in Hollywood block busters require significant more intelligence and perception that require a little looser definition of real-time computing.
What you're talking about is hard real-time requirements. This means that if we miss some timing we have a total systems failure. ROS does not guarantee timing, so you'll have to come up with a different way to handle these. One solution is to have specialized, verifiable hardware the handles those hard real-time requirements. Usually these are things that prevent the robot from damaging itself, property or others.
There are many applications of robotics that have firm or soft real-time requirements. These are real-time tasks that can be handled by ROS. Robotic perception is something that doesn't really fit into the hard real-time requirement, so we have to think about ways that these systems degrade and fail gracefully when deadlines are missed.
Re: (Score:2)
Qualcomm isn't necessarily in the business of making cell phones, they're in the business of selling chips.
nice CPU but with what boards? (Score:1)
This is nice CPU/SoC but what board are they targetting or thinking people can use? Or are they targeting real phones w/ this SoC (assuming it will run upper layers of "cognition" leaving real-time outside of it)?
Why does it need to be processor-specific? (Score:2)
Why isn't that whole code base in plain C/C++? Aside from hardware layers, it would seem that the higher lever functionality shouldn't be tied to a specific processor.
I'd like to see it in systemd (Score:1)