An Original Approach for a Better Remote Control of an Assistive Robot

Abstract

Many researches have been done in the field of assistive robotics in the last few years. The first application field was helping with the disabled people’s assistance. Different works have been performed on robotic arms in three kinds of situations. In the first case, static arm, the arm was principally dedicated to office tasks like telephone, fax... Several autonomous modes exist which need to know the precise position of objects. In the second configuration, the arm is mounted on a wheelchair. It follows the person who can employ it in more use cases. But if the person must stay in her/his bed, the arm is no more useful. In a third configuration, the arm is mounted on a separate platform. This configuration allows the largest number of use cases but also poses more difficulties for piloting the robot. The second application field of assistive robotics deals with the assistance at home of people losing their autonomy, for example a person with cognitive impairment. In this case, the assistance deals with two main points: security and cognitive stimulation. In order to ensure the safety of the person at home, different kinds of sensors can be used to detect alarming situations (falls, low cardiac pulse rate...). For assisting a distant operator in alarm detection, the idea is to give him the possibility to have complementary information from a mobile robot about the person’s activity at home and to be in contact with the person. Cognitive stimulation is one of the therapeutic means used to maintain as long as possible the maximum of the cognitive capacities of the person. In this case, the robot can be used to bring to the person cognitive stimulation exercises and stimulate the person to perform them. To perform these tasks, it is very difficult to have a totally autonomous robot. In the case of disabled people assistance, it is even not the will of the persons who want to act by themselves. The idea is to develop a semi-autonomous robot that a remote operator can manually pilot with some driving assistances. This is a realistic and somehow desired solution. To achieve that, several scientific problems have to be studied. The first one is human-machine-cooperation. How a remote human operator can control a robot to perform a desired task? One of the key points is to permit the user to understand clearly the way the robot works. Our original approach is to analyse this understanding through appropriation concept introduced by Piaget in 1936. As the robot must have capacities of perception 11 Remote and Telerobotics 192 decision and action, the second scientific point to address is the robot capacities of autonomy (obstacle avoidance, localisation, path planning...). These two points lead to propose different control modes of the robot by a remote operator, from a totally manual mode to a totally autonomous mode. The most interesting modes are the shared control modes in which the control of the degrees of freedom is shared between the human operator and the robot. The third point is to deal with delay. Indeed, the distance between the remote operator and the robot induces communication delays that must be taken into account in terms of feedback information to the user. We will conclude this study with several evaluations to validate our approach.