DocumentCode :
574412
Title :
Model estimation within planning and learning
Author :
Geramifard, Alborz ; Redding, J.D. ; Joseph, Jayaraj ; Roy, Nicholas ; How, Jonathan P.
Author_Institution :
Aerosp. Controls Lab., MIT Cambridge, Cambridge, MA, USA
fYear :
2012
fDate :
27-29 June 2012
Firstpage :
793
Lastpage :
799
Abstract :
Risk and reward are fundamental concepts in the cooperative control of unmanned systems. In this research, we focus on developing a constructive relationship between cooperative planning and learning algorithms to mitigate the learning risk, while boosting system (planner & learner) asymptotic performance and guaranteeing the safety of agent behavior. Our framework is an instance of the intelligent cooperative control architecture (iCCA) where the learner incrementally improves on the output of a baseline planner through interaction and constrained exploration. We extend previous work by extracting the embedded parameterized transition model from within the cooperative planner and making it adaptable and accessible to all iCCA modules. We empirically demonstrate the advantage of using an adaptive model over a static model and pure learning approaches in an example GridWorld problem and a UAV mission planning scenario with 200 million possibilities. Finally we discuss two extensions to our approach to handle cases where the true model can not be captured exactly through the presumed functional form.
Keywords :
aerospace computing; autonomous aerial vehicles; control engineering computing; cooperative systems; intelligent control; learning (artificial intelligence); planning; GridWorld problem; UAV mission planning; adaptive model; agent behavior safety; asymptotic performance; constrained exploration; cooperative planner; cooperative planning; embedded parameterized transition model; iCCA module; intelligent cooperative control architecture; learning algorithm; learning risk; model estimation; reward; static model; unmanned system; Adaptation models; Algorithm design and analysis; Analytical models; Noise; Planning; Risk analysis; Trajectory;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
American Control Conference (ACC), 2012
Conference_Location :
Montreal, QC
ISSN :
0743-1619
Print_ISBN :
978-1-4577-1095-7
Electronic_ISBN :
0743-1619
Type :
conf
DOI :
10.1109/ACC.2012.6314997
Filename :
6314997
Link To Document :
بازگشت