High level assisted control mode based on SLAM for a remotely controlled robot

Abstract

One aim of ambient assistive technologies is to reduce long term hospitalization for elderly people, especially with pathologies such as Mild Cognitive Impairment (MCI). The smart environment assists these people and their families with safety and cognitive stimulation, so they stay as long as possible at home. The originality comes from using the robot in the elderly person's home. This robot is remote controlled by a distant user, a therapist or a relative, for determining alarming situations or for participating in stimulation exercises. Several modes are available for controlling the robot. This paper deals with an assisted control mode in which the remote user gives to the robot one goal and the robot reaches the goal by itself. During the robot movement, the user can dynamically change the current goal. An important hypothesis is that the robot has no a priori knowledge of its environment at the beginning. The knowledge will increase with time and the planned trajectory will be refreshed at two levels: a local one - faster but not always sufficient - and a global one - slower but which always finds a path if one exists. The idea is to work only with local information, using the robot sensors, the operator keeping the high level control. To assure that control, the remote operator uses video feedback and information from a laser range scanner.