Initial brainstorming

• 1st brainstorming whiteboard
• 2nd brainstorming whiteboard

Proposal notes

• WhyPerchingIsCyberPhysical -- why what we propose is a good fit with the program
• Cyberphysical proposal notes
• PerchingResearchPlan
  • Need tasks/milestones for 3 years. (Think where we want to be in 3 years and what has to happen at 1, 2 years to make that possible.)
  • CPS-timeline.xls: a first spreadsheet of tasks and times
• NSF-ese: Need to articulate the key science Questions and the Methods that will be used to answer them

Background information:

• ICRA 2010 submission with some new material on powered landing and take-off
• Alexis' CS project on learning trajectories for vertical surface landings
• The public Perching Page at Stanford, including whitepaper
  • ISRR perching paper
  • UAV '09 paper

At right we see Tedrake funnels (from R. Tedrake "LQR-Trees: Feedback motion planning on sparse randomized trees") based on an earlier concept from Matt Mason. What I'd like to do is augment this image in two ways: you plan a nominal trajectory that takes from a start location to the centroid of the desired exit state. But due to uncertainty (e.g. wind), you have a widening cone of states as you progress along the path. (This is the opposite of a funnel.) The trick is to make sure it does not widen so much that you fall outside the funnel. but if you do fall outside the funnel, there is a wider encompassing funnel, which is the "bailout funnel" You can switch to safe bailout (grafeful failure). You abandon the initial plan and go to the bailout plan.

What is a Cyberphysical system? One candidate is a biological or bio-inspired robotic system as defined by Pfeifer07 and illustrated at left. This looks like the embodiment of cyber+physical intelligence. The ideas are similar to the concept of "preflexes" as found in the biology literature (R.J. Full and others).

-- MarkCutkosky - Started 15 Oct 2009