- Main
Optimal tasking of mobile autonomous sensing assets in uncertain adversarial settings
Abstract
In this study we consider the sensor tasking problem, with the case of military reconnaissance as a motivating example. In order to determine the optimal control set, we develop a mathematical formulation for the value of information. This value is inherited from the expected payoff for the activity which uses the information. We address the problem of optimal control of the sensing assets given this inherited payoff through both open loop and state feedback control approaches. A micro-UAV deployment scenario for a three stage urban operation was analyzed with the open-loop approach. For the feedback case, it was found that a particular generic form of the value function is preserved under backward dynamic programming. This form is exploited to develop an idempotent-based numerical approach. Also, efficient refining and pruning methods were developed for attenuating the curse-of-complexity associated with this class of methods
Main Content
Enter the password to open this PDF file:
-
-
-
-
-
-
-
-
-
-
-
-
-
-