Autonomous safety decision-making in intelligent robotic systems in the uncertain environments

Abstract

This research paper presents a method to integrate safety within the decision making process of a mobile robot. For this development, RCS-RMA (real-time control system- reference model architecture) [1] is employed as an architectural framework. POMDP (partially observable Markov decision processes) model is employed at a decision making level of RCS-RMA to ensure safety from within the action selection process of an Intelligent System in the presence of uncertainty. The basic reasons for adopting this representation are initially outlined. Following this, fundamental model aspects are given, outlining the basis for the applied computational model. Experimental results and evaluation are presented in last section.