Do you have an idea for a new feature for the Bento Arm? Post it up here. I will maintain a list of your requested features along with a list of the features currently in development and their completion percentage. If you see a feature on the list that you would like developed please reply to this thread with the feature name and an indication of your vote. We are going to use your input to help prioritize the order we work on the features to make sure we get the most important ones to you the fastest. Once we have a decent amount of votes for a feature we will move it into the active queue and start trying to gather resources and team members to start pushing the feature towards open source release.
How much trouble would it be to add more degrees of freedom to the shoulder joint? It seems to me adding a few more servos in the appropriate orientations would do the trick, but would probably make the shoulder very bulky...
I think it would be possible to add more degrees of freedom to the shoulder joint, but as you suspected I think it would add quite a bit of bulk to the shoulder especially if you wanted both abduction and flexion. It would be tricky to maintain the anthropomorphic feature of the arm with the extra servos. Another tradeoff is that the extra servos might possibly reduce the payload capacity of the arm. In any case it is an interesting proposition and I will add it to the feature request list.
I concur to the "more DOFs in the shoulder" idea. especially interesting: the abduction - and yes I guess you'd need much more power to enforce it... well, just my 2 cents.
by the way, are you really going to designa clinical version of the bento arm? eagerly awating that.
last request: yes, control via UDP / TCP-IP. well, that is a little biased... it's exactly how we do it right now
@drwho72 Thanks for your feedback! I have updated the votes on the feature lists.
I think it may be possible to add in more shoulder DOFs, but we would need to maybe axe the shoulder bearing in order to make more space for servos and as you mentioned use servos with more power. For stronger servos it may be possible to use a higher torque dynamixel PRO servo (would need to be controlled and powered separately) or possibly use an MX-106T with a small transmission to amplify the torque.
The clinical version of the Bento Arm is well on its way and we are hoping to have a working system installed in our local rehabilitation hospital by midsummer. I am investigating increasing the payload capacity by upgrading the existing MX-28AT wrist flexion servo to one of the ones from the new dynamixel X series (probably XM430-W350-T). They are actually a bit smaller than the MX-28AT model, but provide significantly more torque.
As for the communication options for brachI/Oplexus -- we have already begun using these protocols in experimental builds and will hopefully have TCP/IP and UDP modules available in an upcoming release.