Towards Safety in Human Robot Interaction

Many new application areas are emerging where robots will be required to co-operate with or assist humans. This means that they are sharing the same work-space or they are working in our ‘personal’ space. The fundamental resultant issue of working in co-located space is that of safety. In service and social robotics, a major focus has therefore to be safe and reliable operation. It has been identified by some researchers that safety might be expressed in two inter-related modes, namely: physical and behavioural. Physical Safety is quite well understood and refers to potential failures in such areas as torque control and sensor failure as well as exploiting such phenomena as passive compliance. Behavioural Safety is a relatively new research theme, which aims to exploit the bi-directional multi-modal communication channels employed by humans such as: gesture, body pose, facial expressions, back-channel utterances and speech. This further requires individual and shared cognitive models between robot and human to provide the basis for models of shared goals and shared intentions. The key hypothesis is that safe interaction between human and robot can be engineered physically and cognitively for joint physical tasks requiring co-operative manipulation of real world objects. The advancement of actuator technology, mechanical systems design, sensor systems, computational power, control technology and strong links with disciplines such as cognitive psychology has lead us to this assumption to be true in