iRobot and the Create3

alanmcd · December 5, 2024, 4:31pm

Early this year (January 2024), when the iRobot-Amazon deal fell through, iRobot laid off 31% of their workforce and announced they were “pausing development of all non-floorcare products”. There was a brief (2-3 weeks) absence by the iRobot Education team from the GitHub discussion pages, but after that break Create3 users were again getting great support. Really phenomenal direct from developers support with one to three day turn-around on questions.

On Nov 2, 2024 the “Shop Create3” link on the iRobot Education site became non-functional, and iRobot stated they were “migrating the website, and the link should again be functional shortly.” It remains non-functional.

On November 5th, 2024 iRobot announced layoff of 16% of their remaining workforce, and another “radio silence” began with the last support posting on Nov 7th.

I am not aware of any statement by iRobot, nor ClearPath, nor any 3rd party about the future of the Create3.

I believe the Create3 is the best featured, best documented, best supported, ROS 2 programmable robot available on the market, and am very concerned that the earth has opened and swallowed it (at least as the product we know). A very important issue with a “ROS 2 native robot” is the yearly “march” of ROS 2 (a ROS pun…) to the next version often with backward compatibility issues preventing mixed versions. Therefore iRobot should publically commit to continued support of the Create3, or commit to release an Ethernet-over-USB Python API so that we can maintain our own ROS 2 node above a “non-ROS-Create3” - the latter would be my preference.

It appears the Create3 has been pulled from 3rd party sellers.

If anyone knows anything about the future of the Create3, I would sleep better hearing either the good or the bad.

Alan

cbenson · December 5, 2024, 5:01pm

@alanmcd Welcome back. Honestly don’t know any details, so we’ll see if we can reach out to Clearpath to know if they’re aware of anything, and we’ll get back to you here if we have knews. No reply = no knews.

Indeed, the robot vacuum market is effectively “flooded” with so many different manufacturers trying to compete against each other with small differences and prices (and available everywhere) that even Neato, once one of the top manufacturers in the field couldn’t compete.

The question becomes will consumers / developers require more AI integration or is a platform specifically for ROS be suficient?

Internal ref: 1021399

cbenson · December 5, 2024, 7:50pm

Clearpath’s reply (they’ve always been great to work with):

“Yes, we’re committed to supporting the Create3 and we’ve secured enough units to continue production of the TurtleBot4. We also worked closely with iRobot on the new ROS Jazzy update to make sure the TurtleBot4 continues to work with the latest version and its users can get the most out of it. Hope this helps. If you have any other questions please let me know. If anything changes on our end, you’ll be the first to know.”

alanmcd · December 6, 2024, 3:01pm

Thank you for this clarity. It really saddens me to see the best programmable robot platform ever, pulled from the market.

Robots are “life” for me, but the reality is robots come from businesses that must turn a profit. Robots are complex and the market for educational robots is complex, leading to great ideas that can’t find enough traction.

Over the last 25 years, I have felt “separation anxiety” at the end of the RugWarriorPro robot, the Cye robot, the “narrowed scope” of the GoPiGo3 robot, and now the Create3.

The Create3 has the most accurate odometry of any educational robot, has self protection sensing, power management sensing and self docking. If it had offered the Python (non-ROS) API via Ethernet-over-USB the bot would have needed less support, no continuing update commitment, and perhaps continued to be the most capable educational robot on the market.

Learning to program robots starts out extremely simple, but quickly becomes complex. Everyday another student asks on Reddit, “How do I learn ROS quick?” to which the Create3 was usually my answer to them.

At this point it may fall to either the DuckieBot or the YahBoom ROSmaster series, as supported ROS 2 educational platforms. Neither of these have bumpers or self-docking which I feel are needed for a “state of the art” platform. (I also feel every “self-respecting” robot should come with a wall-following and safe-wander capability.)

I would buy the Amazon Astro (even at the current $1600) if it had a user accessible API. I wish Amazon would offer to sell all those recalled “Astro For Business” robots to the ROS community, with a SDK that exposes an API. There is a PD capable USB-C port that can power a Pi5 or Jetson. Not sure if Ethernet-over-USB is also available on that port - I suspect it is available for them to run diagnostics on the bots before shipment.

Alas, my thoughts are all over the floor, writhing in pain from this situation. There is some comfort in finally knowing what is happening with the Create3, so I thank you again for your help.

Sincerely,

Alan