The Robotis TurtleBot3

[Gort]

Looks like Intel is getting out of the IoT business! Below is the road path for the Galileo, Galileo 2 and Joule.

• June 16, 2017 – Product Discontinuance Program Support Begins
• July 16, 2017 – Product Discontinuance Demand To Local Intel Representative
• September 16, 2017 – Last Product Discontinuance Order Date
• September 16, 2017 – Orders are Non-Cancel-able and Non-Returnable After
• December 16, 2017 – Last Product Discontinuance Shipment Date

Why would anyone buy the Robotis TurtleBot 3 Waffle and Trossen TurtleBot 2i now?

http://mike-ibioloid.blogspot.com/

[tician]

Doubt it means much of anything other than Intel finally realizing they got into the mobile/embedded/IoT world too late and too half-assed. If they skipped the Edison and instead released the Joule modules a couple years ago, then they might have managed to compete with the RPi. There are plenty of ARM boards that compare relatively well to the Joule 570, and most already have all the USB, A/V, ethernet, and storage connectors already populated instead of needing custom shields through tiny board-to-board connectors. If a full x86-64 platform is needed to avoid ROS on ARM, there are still plenty of small PCs that can work with the Kobuki and might even be able to fit into the burger.

[jwatte]

This is sad for the people who invested in the Intel offerings, for sure. Trossen and Robotis are probably not happy right now.
I suggest that the Raspberry Pi 3 is probably the best bet for low-end robotics; the price is right, and the board is quite powerful enough for most robotics algorithms, including visualization and SLAM.

[tician]

Of course, the RPi3 does not have USB3 and only 1GB of RAM, which might cause trouble with the Realsense cameras. Assuming they do not get discontinued as well.

[KurtEck]

There are also several other Intel based products out there... Like UP... (As I have mentioned I already have an UP, sometime next month my UP2 should arrive. Plus I currently am supporting the kickstarter for UP core (https://www.kickstarter.com/projects...project_update)

The Edison was an interesting platform, the problem I ran into was they had did not do a very good job of supporting it software wise as well as add on hardware support.

[tician]

I really wonder who at Intel thought the Edison would ever do well. The low-power mobile/android/linux with wireless connectivity market has been dominated for many, many years by a huge ARM ecosystem, so there was no real chance of gaining influence there. The high-performance mobile/android/linux with wireless connectivity still had plenty of room for x86 processors a couple years ago, but not so much now with quad/octo-core ARM in 64-bit, built-in peripheral hardware, and greater software support. Guessing a bit of it is Intel being a fab instead of just licensing the platform to others like ARM does.

[Gort]

The TurtleBot3 Burger will be fine because it uses the Raspberry Pi 3.

[jwatte]

there was no real chance of gaining influence there

I think there could have been, if they had:
1) Been price competitively.
2) Interfaced natively with common standards (3.3V/5V I/O, USB, display, at a minimum, plus on-board wireless)
3) Been power competitive at equal or better performance..
4) Commitment to skilled software engineers supporting it in an ongoing fashion post-sale.

1) and 2) are important for hobbyists, 1) and 3) are important for actual system wins, 4) is something that Intel just hasn't learned how to do. They're a hardware vendor; the support ends when the hardware ships; the hardware is considered obsolete when the manufacturing lines stop making the chips. That world may have worked 30 years ago, but it will be the undoing of Intel in the modern, software-defined, connected-forever world.

[KurtEck]

As I mentioned I thought Edison was interesting... I started to play with them when HR-OS1 mentioned that was the main processor.

It has Wifi and BT built in. It was nice programming the board over wireless and the like.

Yes you needed their Arduino board to be able to do 3.3v or 5v and the IO was slow. It was really slow to change the IO pins directions as it required a couple of (trying to remember if it was I2C or SPI) changes to external chips to change the configuration...
But I could live with that.

What killed it for me was 4)

Examples: As I mentioned things like pinMode() and digitalWrite() were real slow... I found ways to speed some of these up a lot (like half the time maybe less) and submitted it to them... It took maybe 6 months to a year before anything showed up in a build...

SPI - The spec showed that it could run at a very high speed, but actually an Arduino did circles around it... And each release the broke it worse...

But really killed it for me, was their support people were basically not allowed to tell the customers anything. Like they might say the engineers tell me that have fixed the SPI problem (not which one) locally. But they can not tell you if it will be released with the next release nor when they expect the next release...

Could go on, but as I said 4) :8

[Gort]