-- MarkCutkosky - 01 Aug 2003 I had written my own notes and was about to post them when I received Dan's so I've decided instead to intersperse mine with his

From: "Daniel Koditschek" <kod@umich.edu>
Date: Fri Aug 1, 2003  9:41:38 AM US/Pacific
Subject: Notes: 8/1 RiSE Telecon 

To: "Ron Fearing" <ronf@robotics.EECS.Berkeley.EDU>, "Alfred Rizzi" <arizzi@benedict.msl.ri.cmu.edu>, "Dan Koditschek" <kod@umich.edu>, "Kellar Autumn" <autumn@lclark.edu>, "Mark Cutkosky" <cutkosky@cdr.stanford.edu>, "Martin Buehler" <buehler@cim.mcgill.ca>, "Robert Full" <rjfull@socrates.berkeley.edu>

I) What's been done over the last month:

MarkCutkosky: Stanford team is exploring different ideas while

• educating selves about approaches in nature (BioClimbing -- includes GeckoPads, WetAdhesion, GraspingToes, ClawGeometry, etc.)
• making some crude prototypes for intuition. (see UnderActuatedFrontLeg, FootDesignIdeas, ClimbingConcepts)
• doing some simple modeling & simulation: (http://dart.stanford.edu:88/Get/File-4746/LegDesign.zip) and (RobotAnalysis).

Early observations: can't just create gecko material without paying attention to the conformability of substrate. Options include toes, gel pads, foam, etc. The same issues are important for claws and spines (without compliant substrate they don't scale beyond individual point contacts).

Questions:

• Does this compliance need to be active? (In part, yes)
• Does the initialization of contact make a difference? (Yes)
• What is a "foot" and how many DOF does it need?
• Where do various parts of the control and actuation reside? body versus leg (e.g., gecko vs crab)
• What are the force constraints at each limb? -- this can be answered more abstractly, which leads to the notion (see below) of distinguishing, formally, between Physical and Functional specifications.
• How to structure the inteface between the systems - body, legs, feet, compliant?
• What are the right axes for comparison, or what is the right decomposition?

II) Toward pinning down the axes of interface specs./dimensions:

Three working subcommittees to form over the weekend, develop initial crude "shopping list" of criteria/dimensions/specs and document same in time for general group review prior to next (8/8) telecon such that there can be a critical appraisal of how to take the next finer cut during the 8/8 meeting.

1. MarkCutkosky leading group to outline FunctionalDecomposition? , concentrating on functional requirements
   e.g. need-to-climb => constraints on forces/speed/payload; reliability; comms.; computation; power; etc. all independent of the "how" This group should include Ron.
2. AlRizzi leading group to outline "functional decomposition" concentrating on computational/communications/sensing aspects. This group should include ShaiRevzen from Polypedal lab and Trey (AMcClung) from Stanford Lab
3. Martin will lead initial exercise in a physical spec (need to put a stake in the ground somewhere to get design off the ground): eg.,

assume machine will have 4 legs 3 actuated dof per leg ==> many consequences assuming 2kg machine can travel vertically at 1 bl/s; (e.g. also require for quasistatic (deliberate) climbing/walking dynamically need to maintain a certain support pattern that further imposes constraints, etc, etc.
This group will meet by telecon Monday 8/4 1pm. Includes Kod & Pei-Chun, and Mark.

III) Miscellaneous Notes

Early morphology decisions:

• How many legs? (affects probability of foot stick)
• Can legs pull in? or is it RHex like? NO almost surely will need dexterity => 3dof actuated.
• Bending back? Can substitute dexterous legs for bendable back?

(And if you can, do you want to? The functional analysis tells us that horizonta/vertical transitions occur much less frequently than periodic leg motions and can therefore occur at lower speed. And therefore can be low power "shape changing" actuators... -mrc)

All agree on an imperfect developing interface definition to facilitate individual groups forging ahead with a little bit of autonomy. Need a working prototype by 6 months, each iteration costs ~ 2 months (parts ordering and delivery etc) so time is very tight!

Need to think about what we'll show/do/present at 8/28 DAPRA Biodynotics meeting (Kod sent out the abstract page to all the pis during this meeting: please look it over and think about what to do with it).

No "check-in-mail" message from SPAWAR yet, but Kod thinks "any day now" In any case, we should feel as though the project has officially started as of today.