NSF Project Ideas

Latest revision as of 05:32, 30 September 2008

1. Most generally I am looking at the analysis and design of distributed systems, as well as their comparison to centralized systems.
   1. Less generally I am looking at multilegged robot locomotor schemes
      1. Presently I am looking at coupled oscillators. Literature is emerging investigating them as robotic design elements. There is also quite a bit of literature analyzing them.
      2. Relevent researchers right now are Golubitsky, Stewart and Ijspeert, the salamander guy. The Japanese are good for applications.

Questions:

1. What will my work be?
2. What is meant by CPG now that there is literautre describing them as networks of coupled oscillators. How central is that?
3. In how literal a sense did the Golubitsky Collins work describe 'modules' and how has that insight been applied to robotic applications
4. In what way are coupled oscillators suggesting a design methodology?
5. How can I compare these schemes to centralized equivalents? Can I model them as their centralized equivalents (insofar as the limit cycle is an ideal around which they oscillate). Does that even mean anything?
6. What are outstanding problems with the application of the oscillators to design goals?
7. How do I fake group theory?

Leads:

Here are some papers that I wasn't able to get my hands on

Title: Periodic dynamics of coupled cell networks II: cyclic symmetry
Author(s): Stewart I, Parker M
Source: DYNAMICAL SYSTEMS-AN INTERNATIONAL JOURNAL   Volume: 23   Issue: 1    Pages: 17-41   Published: MAR 2008 

Title: Learning to move in modular robots using central pattern generators and online optimization
Author(s): Sproewitz A, Moeckel R, Maye J, et al.
Source: INTERNATIONAL JOURNAL OF ROBOTICS RESEARCH   Volume: 27   Issue: 3-4   Pages: 423-443   Published: MAR 2008