🚀 go-pugleaf

Heres my 2c worth to the discussion:

1) Dante's (the Robot that almost went down the volcano) purpose is 

      a) Not to go down Mount Erebus and do a jig
      b) Not to test teleoperation as suggested, but it so happened that it
         was required.

   The primary purpose of Dante was not to build a Robot but to fulfill
   a mission that was till now not possible (see below). That in itself
   was a major step for mobile robotics research at CMU (and probably 
   elsewhere).
  
2) Mount Erebus is probably the only unvisited surface feature on this planet.
   Snippets 

         . A while ago a manned attempt to go there and collect scientific
           data was attempted and a minor eruption almost cost a life and
           further missions were FORBIDDEN.

         . The glass panes at an observatory on the tip of Erebus got corroded
           in a few months (glass is supposed to be pretty inert)

         . Geologists think that this volcano which is almost continuously
           spewing tons of gases into the atmosphere may have a significant
           impact on the environment. Data on these gases and other relevant
           data is deemed of high value

  Conclusion (at least of the people interested) : Exploration of Mt. Erebus is
  of high scientific value.


3) Science projects aboard Dante: 
      a) Stereo camera feedback.
      b) Gas chromotograph and other sensors to determine gas composition.
      c) Other stuff i may be unaware of.      

4) Solutions:

     a) Dante (cost 2-3 mil), compare with costs of other envirnmental projects.

     b) ? Last year when this project was up for grabs, there was apparently no
        microrover proposal. Please correct me on this, i could
        be wrong.


5) What did the `failed' Dante accomplish?

   a) Proximity sensors for the feet which are apparently reliable.
   b) A unique polar laser range finder which is much lighter and the
      first of its kind designed to suit the weight and sensing requirements
      of a mobile robot.
   c) Teleoperation (enough said about that)
   d) Software for robot simulation in terrain, mission monitering and 
      teleoperation software for operation at various levels of autonomy.
   e) A science package (which costs a lot) assembled for such a mission.
   f) A new kind of walking mechanism which exploits the mechnical design
      to generate a natural gait (much like the ackerman steering reduce
      degrees of freedom).

6) Questions:

     a) Is reliability the issue here or viability?
         
      Aspects that may hinder a microrover:
      . Dante required protection and heating for its electronics.
      . Sensors like foot proximity sensors, stereo vision, laser range finders
        are not scalable (though they may not be required for the microrover
        concept, some sensing is required).
      . The science package is probably not scalable too. 

     b) The microrover concept seems to be tied in with large numbers and low
        cost. Given the same mission or objective, i am interested in knowing
        how the cost scales since that is a major concern for planetary 
        exploration. The question is planetary exploration for what (and why)?

     c) Continuouing the previous question, the JPL microrover slated to go to
        mars will travel about 10 m or so. That clearly demarkates it from a 
        `robot' concept which is capable of locomoting very long distances, since
        it carries all communications and computation on board. 
        The question that arises for mars exploration is, will not an army of
        micro landers do the same job? Why do we need a micro rover which adds to
        weight and cost? (for example, a mini arm attached to the lander may be
        able to do the required exploration?

        Note that the question above is rhetoric. In a sense the `required' 
        exploration is tailored to meet the budget, based on a priority of
        science missions. Hence if a particular objective can be accomplished
        with a lower cost, it may get a higher priority.


     d) How does Dante help the debate for or against any kind of robot 
        philisophy? Can we generalize from an instance where the failure
        was not due to a flaw in the robot design.


7) You probably guessed my personal bias by now. I think that mobile robots
   for exploration will not be monsters like the Ambler but not microrovers
   simply because of the sensing, computation, power and exploration 
   requirements prevent a micro size (today). The robots that do exploration
   might very well be the size of pet animals wieghing upto 150kg or so and
   able to accomplish long, autonomous missions with almost any kind of science
   required. This seems like predicting a cray on every desk in 1975 or even 1980.
   But commercialization of sensors and actuators might make it possible soon;
   how soon is up for grabs.

cheers,
lalit