I have a project idea which would make use of a robotic arm. But since I have no experience with robots I’m not sure what is possible. I’ve done some research and will do more but thought I would stop and ask some questions as this board seems pretty active.
Basically I want to build an arm which can copy a motion the arm is taught by manualling moving it. For example, lets say I outfit the tip of an arm with a pencil. I instruct the arm to go into “learning” mode, then grab the arm and draw a simple shape. I take the arm out of “learning” mode and instruct it to repeat it.
Is this possible? OK, just about everything is possible with a sufficient application of time and money, but is it doable with common of the shelf parts and some custom code?
What would I need? I’m not looking for a parts list, but more of a general guidance.
I have a BSEE, but have spent the last 12 years in the software/telecom industry. Can you recommend any good books or web sites to get me up to speed in the world of robots?
Thanks for reading and thanks in advance for any guidance you can give.
I havent looked around on these forums for many other answers, but I would look into a fairly cheap way of doing this with parts already out there. For example using the Robotis Bioloid Servos (just google robotis or bioloid) you can communicate at 1 megabaud to the servos so you easily can read a position while keeping a torque on the servos stiff enough to sit in place while you think… And as for a pencil and drawing, why stick with simple shapes?
The RoboNova has the capability that you describe.
I assume that if you purchased the RN-1 controller, RoboBasic software , and some RN-1 servos, you could make an arm that would have those functions.
You also might want to look into OpenServo as a much cheaper and much more unrestricted way of doing it (you swap the internals of a cheap analog servo).
Last I heard, they were testing them in a Biped Scout.
This is why Open Servo may be the way for them to go. The cost is pretty much controlled by the kind of servo hardware you want to use. People are using HS-475s and some are experiementing with HS-645s now.
I want to get involved with Open Servo for sure, but I am waiting for somebody to come up with assembled boards. There is no way I want to try and solder surface mount components. Having position and current sense feedback in servos is very attractive. Check the Open Servo site for more information and they have message boards too.
An open servo probably wouldn’t work either as servos are not made to be rotated by the control horn. The arm would probably immediately start to bend and finally break under the stress of being externally moved.
That is true for some micro servos such as the HS-85. But all standard Hitec servos will not be damaged by manually rotating the final gear. Actually all he needs to do is open the servo and attach two wires to the pot. One to the wiper (center) and one to the ground. Then attach this signal to a high impedance input A to D converter. Then the A to D can easily read the servos position at any time. Besides the hassle of opening each servo and routing the two extra wires out of the enclosure, it’s relatively easy. Much easier and cheaper then replacing the control board with an open servo one.
Hmm. I did not know this. One could multiplex one ADC amongst several servos. Otherwise it would require one ADC per servo, which sure wouldn’t be cost effective. With 16 input selectors readily available, only two ADCs would be required for 32 servos.
You eeese da schmat dood!
I will be filing this for future reference and my purchase an inexpensive (maybe an HS422 or similar) servo to try this on soon. I think I want to keep my experimentation with servos to the Hitec stuff.
Attaching across the pot should also work for normally operating servos. One could also monitor for high servo current conditions (stalled servo) by measuring the relative current being used by each of the servos.
