Thanks for intelligent comment - it is so rare to recieve it here...

The problem is with the too limited intelligence and capabilities of the systems NASA is working on... and yes, my definition is more than a little different, and extends far beyond simple colllision avaoidance...

If I remember correctly, Sojourner "ran aground" twice, once by being given a plot over a rock which was too large, and once by jamiing itself up against a rock it was directed to sample. Ground controllers set up a VR landscape, and tried using VR techniques to "practice" instructions, but none of it worked very well...

The human interface generated problems, and for the MER's NASA has spent a lot of money on the human side of things- controller training, staffing, design of the human control system facility. If I remember the PR releases correctly, the best vision systems that NASA could come up with for the MERS has been the systems from Cornell, and these are limited... and of course way less than those available to Japan, France, Italy, and Russia...

Based on the Soviet experience, where work with the lunakohds led to an examination of this problem in detail, my thinking is that we can only expect a detailed analysis after operating the MERs reveals fully the problems with remote control of distant rovers... In the meantime, we'll get a unnecessarily limited return on our $800 million investment for these rovers..

ep