We will need a model for claw like attachment and sticky pad type attachments, unless the model is flexible to handle both types.

The foot attachment model may need to include time-dependent information. The adhesion and frictional properties of viscoelastic type materials probably changes significantly as a function of time. Perhaps as a starting point just being able to know the state a particular previous state a fixed time previously would be enough.

Depending on the substrate, the friction model may need to directionaly sensative. For non-isomorphic surfaces the orientation of the foot is going to be very important. Perhaps we will need to specify a major and minor axis with high and low frictional properties.

At some point we will need to have a model that permits a given amout of adhesion. Again, this may be time dependent. The amount of adhesion may or may not be a function of shear force. E.g. a claw when hooked may be able to apply a large adhesion force independent of shear.

Removing a claw or a Gecko seate is much easier when peeled off at the right angle. Experiments with Geckos suggest that at an angle of about 30 degrees the seate begin to pull off easily.

The frictional forces in the plane are not independent of each other. I.e. friction circle instead of a friction square. In addition in-plane torques also add, and we may need to think of it as a 'force hemispere.'

-- JonathanClark? - 23 Sep 2003