An investigation of service degradation in long-term human-robot interaction with a particular reference to recharge behaviour
Deshmukh, Amol Arun
MetadataShow full item record
Autonomous long-term operation of social robots has always been a challenge in Human robot-interaction. Social mobile robots acting as companions or assistants will need to operate over a long-term period of time (days, weeks or even months) to perform daily tasks and interact with users. Therefore they should be capable of operating with a great degree of autonomy and will require sustainable social intelligence. Social robots are fallible and have their own limitations with the service they provide. One of the most important limitations of mobile robots is power constraints and the need for frequent recharging. Social mobile robots generally draw power from batteries carried on the robot in order to operate various sensors, actuators and perform tasks. However, batteries have a limited power life and take a long time to recharge via a power source. While the recharge behaviour is active, which may impede human-robot interaction and lead to service degradation. This thesis raises some important issues related to recharge behaviour of social mobile robots which appear to have been overlooked in social robotics research. This work investigated service degradation in long-term interaction due to recharge behaviour of autonomous social mobile robots and proposes an approach to manage service degradation due to recharge. First we performed a long-term study to investigate the service degradation caused by the recharging behaviour of a social robot. Second we conducted a more focused social study which helped to understand user’s attitudes towards a mobile robot with respect to recharge activity. We explored a social strategy by modifying the robot’s verbal behaviour to manage service degradation during recharge. The results obtained from our social study indicates the use of verbal strategies (transparency, apology, politeness) made the robot more acceptable to the users during recharge. We believe that social mobile robots should behave in a socially intelligent manner while managing service degradation. We also provide some recommendations for social mobile robots to manage their recharge behaviour in this thesis.