The Spine - Surface interaction model should start with some simple geometric considerations I still have somewhere on a scrap of paper and never put on Twiki with some rigour. The most interesting and very useful results can be summarized as follows.

• If we consider both an ideal spherical tip (for a spine/claw) or an ideal sharp tip (real cases are in between)

• in both cases the interaction with a surface depends on the Coulomb friction coefficient

• (of course we are neglecting other forces, which have a lower order of magnitude)

• the hooking force can be simply determined from just the friction for the interaction area.

• The actual interaction area can also be simply calculated, from the actual spine/claw tip geometry (see also ClawTipMeasures? ) and from the surface geometry (see also SurfaceProfileAcquisition? ).

• It would be useful to gather statistical data about surface roughness from buildings of different types (stucco, stone walls, bricks etc.), but probably architects/engineers are not interested in those kind of information, so those measures are not available.

• To further simplify the calculation (also considering the simple initial assumptions) a spherical tip of average geometrical size can be considered and some general parameters from the surface can be used.

• In conclusion, by the actual distribution of asperities on a real surface (which can be measured) we get the probability distribution of hooking and many other useful statistical information.

This model is still under study and will appear soon in a paper.

The profile analysis is carried out in the attached m-files. Acquired surface profiles are attached to this page as a large zip file in different batches.

• LabJackStreamSlow1.zip: First batch of reliable profiles (87x8MB)
• LabJackStreamSlow2.zip: Second batch of reliable profiles (17x3MB)
  • .dat files are the original acquired file (Excel)
  • .xls are cleaned profiles also displayed as a graph
  • .m are cleaned profiles ready for import in Matlab
    • all the above share the same name for each type of material/acquisition
• M-files.zip: Matlab files (20x100KB) with profile input and analysis
  • ProfXX.m are profile analysis file (without actual profile input)
  • MdaqXX.m are tentative matlab files for direct acquisition with the labjack card

The profiles measured and attached come from different materials and textures. Two concrete bricks are staying in the lab and should be kept safe (not broken or discarded) because they can be used to assess the performance of different feet configurations on non-penetrable surfaces where the actual surface geometry is known!!!

Unfortunately we do not have a portable surface profilometer which can be used on the vertical external surfaces where we use to test the climbing robots. And also, it is much easier to test spine/claws on a table in the lab than on the outside walls (especially in winter).

"Maybe one day some bricks will fall out from those walls for unknown reasons..." -- MarkCutkosky