Standardized Control Interface for Prosthetics (SCIP), a CAN-based Limb Control Protocol
I've spoken before about the need for a standard for prosthetic/robotic arm (or leg) control, preferably one also adopted by a larger industry. University of New Brunswick Fredericton (UNB)'s AIF UNB Hand Project has settled on a CAN-based communications system, using the established standard for automotive communication. Under the name of a consortium that has long sought to standardize upper extremity prosthetic control, the Standard Control Interface for Prosthetics (SCIP), UNB's Yves Losier has posted a draft of the SCIP CAN-based standard for prosthetic control.
Because both of the current DARPA prosthetic arm efforts both use the CAN bus, but are not harmonized, this has been much discussed as a possible basis for a standard. While one of the DARPA arm groups, JHU-APL, had announced some months ago that they were planning on publishing the RP2009 protocol as an open standard, they have thus far not done so. The other group, DEKA Research, has repeatedly declined to discuss a common standard, but have not said that they will never participate. Several European research groups have settled on using the CAN bus, and the publication of the SCIP document represents a great start to the discussion. As all of these projects generate inertia, it will only become more difficult to introduce changes in architecture. It seems to me that the sooner we begin a dialog on these issues, the better.
Thanks to UNB and Yves for putting up this draft, which looks like it gives us a pretty general basis for beginning a discussion of a possible standard. Obviously, there is detail missing from this document that would allow the control of the AIF UNB hand with a third party controller--through the Device Profile Layer in the document. This level of detail (or a companion Application Programming Interface or API from which the appropriate Device Protocol could be generated) is what would be required to achieve true compatibility and interchangeability across systems by different manufacturers. As hands with varying levels of articulation become available for research and development, I hope that this level of detail emerges for all terminal devices for with input from all stakeholders. It would be great to see a sample Device Protocol from someone. I'm interested as well in the sensor module, and if anyone would be interested in a standard sensor bus.
Ultimately, my hope is that any manufacturer or research group could use the published standard and related APIs or published source code to ensure compatibility with any arm joint, hand, or intrinsic hand joint. I believe that external control of intrinsic hand joints is key to this standard, for several reasons. For a highly articulated hand in a transradial application (the vast majority of cases we are talking about), the hand contains the majority of joints in a complete arm system. The effective control of these joints for true dexterity rather than grasping is both the ultimate clinical goal, as well as being the most interesting and challenging research goal in this area. The reduction of the full range of articulated hand control to a subset of degrees of freedom smaller than the set of joints available stands to limit the research possibilities available with any hand for which this is done, and will force researchers to duplicate effort in order to conduct meaningful research.
A common infrastructure as I have described would ensure that the best of any single electrical mechanical component could be used with any other. This would ensure that the marginal cost of innovation remains low, and we can all focus on solving the problems that no one has solved, rather than solving the same problems over and over again, in many cases using US Government money to do it each time. While not part of the proposed standard, Open Prosthetics intends to create an open source software platform to make it easier for researchers to configure multiple limb components and to test and share different software algorithms. Please contact me if you have an interest in this project. We have some code published already (thanks to Blair Locke, UNB, Tim Hanson and Peter Truskey).
Please consider joining the Google Group and check out the wiki for this project. I look forward to your input into this open standard as we develop it.
Sincerely,
Jon
Because both of the current DARPA prosthetic arm efforts both use the CAN bus, but are not harmonized, this has been much discussed as a possible basis for a standard. While one of the DARPA arm groups, JHU-APL, had announced some months ago that they were planning on publishing the RP2009 protocol as an open standard, they have thus far not done so. The other group, DEKA Research, has repeatedly declined to discuss a common standard, but have not said that they will never participate. Several European research groups have settled on using the CAN bus, and the publication of the SCIP document represents a great start to the discussion. As all of these projects generate inertia, it will only become more difficult to introduce changes in architecture. It seems to me that the sooner we begin a dialog on these issues, the better.
Thanks to UNB and Yves for putting up this draft, which looks like it gives us a pretty general basis for beginning a discussion of a possible standard. Obviously, there is detail missing from this document that would allow the control of the AIF UNB hand with a third party controller--through the Device Profile Layer in the document. This level of detail (or a companion Application Programming Interface or API from which the appropriate Device Protocol could be generated) is what would be required to achieve true compatibility and interchangeability across systems by different manufacturers. As hands with varying levels of articulation become available for research and development, I hope that this level of detail emerges for all terminal devices for with input from all stakeholders. It would be great to see a sample Device Protocol from someone. I'm interested as well in the sensor module, and if anyone would be interested in a standard sensor bus.
Ultimately, my hope is that any manufacturer or research group could use the published standard and related APIs or published source code to ensure compatibility with any arm joint, hand, or intrinsic hand joint. I believe that external control of intrinsic hand joints is key to this standard, for several reasons. For a highly articulated hand in a transradial application (the vast majority of cases we are talking about), the hand contains the majority of joints in a complete arm system. The effective control of these joints for true dexterity rather than grasping is both the ultimate clinical goal, as well as being the most interesting and challenging research goal in this area. The reduction of the full range of articulated hand control to a subset of degrees of freedom smaller than the set of joints available stands to limit the research possibilities available with any hand for which this is done, and will force researchers to duplicate effort in order to conduct meaningful research.
A common infrastructure as I have described would ensure that the best of any single electrical mechanical component could be used with any other. This would ensure that the marginal cost of innovation remains low, and we can all focus on solving the problems that no one has solved, rather than solving the same problems over and over again, in many cases using US Government money to do it each time. While not part of the proposed standard, Open Prosthetics intends to create an open source software platform to make it easier for researchers to configure multiple limb components and to test and share different software algorithms. Please contact me if you have an interest in this project. We have some code published already (thanks to Blair Locke, UNB, Tim Hanson and Peter Truskey).
Please consider joining the Google Group and check out the wiki for this project. I look forward to your input into this open standard as we develop it.
Sincerely,
Jon
Comment
About
Badge
© 2009 Created by Jon Kuniholm on Ning. Create a Ning Network!
You need to be a member of OpenProsthetics.org to add comments!
Join this Ning Network