Relating Autonomy To a Task: Can It Be Done?
Bruce T. Clough · 1st UAV Conference · 2002
There are currently countless methods touting being the best at autonomous control. whether it is emergent behavior or dynamic programming, fuzzy logic or Bayesian Belief Networks, each technology has its phalanx of zealots that proclaim it to be superior for the job and the rest rotten, or at least trivialized. In this paper, we step back and look at the big picture in autonomous control, developing the idea that what will drive the choice of the particular technology is the task, or tasks, that the autonomous system must do, and not the merits of any particular technology by itself. The optimal technology for any task in inexorably intertwined with the task - they cannot be separated.