A Model for Autonomous Safety Management in a Mobile Robot

Abstract

This paper deals with the problem of embedding safety into an autonomous mobile robot's behaviour. Although robot safety has long been an issue of concern, there has been little work on the contribution that autonomous behaviour can make to a robot's safety management. An approach is hereby explored for explicit safety reasoning within the robot's decision-making process, based on the assumption that the robot, at best, can have only partial knowledge of the robot and environment current and future states. A probabilistic model is defined, based on a robotic system's safety analysis outcome. The model essentially represents a process of autonomous risk assessment and management of the robot's actions