To act or react, which is more crucial? Usually this question intends to elevate all our actions to the levels of conscious and awareness and missing the point that even in our aware actions we are still reacting to something prior to the action itself. Meaning that both acting and reacting are two inseparable concepts that co-exist in the same situation. As a result, the wheel of acting and reacting is eternally in motion leading to progress and evolvement in life. Therefore, acting and reacting in practice play a role in research studies. We will discuss theatre acting and audience reacting to it in a novel method that involves robots on stage within the context of HRI experiments.

Testing Human Robot Interaction

Within our research in the dynamics of living with robots and interactive companions, it is essential to consider actions of robots and reactions of humans for the near future scenarios. Testing similar scenarios during the present could approximate and anticipate the future. In LIREC and within the various studies conducted, a relevant study of sharing living spaces with robots in a home and social places was performed to investigate how this will affect social spaces relationships and proxmics with humans in order to shape how the interaction will have to be. For this reason the experiment has to involve working with medium or large human-scaled mobile robots in rather big spaces with considerable resources of hardware and human researchers and experimenters [1,2,3,4].

If we take the Wizard of OZ method for example, it involves specific settings that are not always easy to have with many particular limitations related to the number of participants on the one hand, and to the number of the experimenters running and administrating the robot on the other hand. It is essential however to have valid and reliable results in these kind of experiments. Thus a sufficient number of experiments have to be run with rather big number of participants.

Due to the restrictions and unaffordable resources to run such studies that would last for long periods of time; other methods of experimentation has to be used.

Added-values of Theatre

For all the above-mentioned restrictions in large-scale experiments, the pilot study of using theatre in studying robot-human interaction is a promising novel approach for initial user trials in HRI. This method uses theatre acts involving human actors and robots in Human-Robot interaction scenarios.

The added values of this method compared to the Live and Video HRI methods are many. It affords involving more than one individual user to test the technology in hand in an interactive effective way with the feasible use Wizard of Oz. Allowing running early stage experiments before employing many resources for implementing full autonomous robots. The experiment scenarios have a variety of options since they don't rely on the presence of one subject who is responsible of participating in action with the robot. Instead, the actor plays the role of a user interfacing with the robot and the audience will be in this way able to watch and then involve in a discussion with the characters themselves to investigate aspects and issues of interest to the audience.

This provides freedom in creating the scenarios and scripts for the actors and semi actors (the robots). The scenarios include very advanced functions of the robots that they might not exactly have, such as: smooth space navigation, natural language interaction and reasoning and all other social skills and intelligence that we humans have. Resulting in having significant feedback about the targeted robotic system form the targeted users involved in the experiments.

Starring Actors

In the experiment of this pilot study a Pioneer and Peoplebot were the electronic actors in the scenarios after being modified by the LIREC team at the university of Hertfordshire. Following we list all individual tasks involved in the play: 1- Actors: improvise scenarios that might face the robots 2- Semi-actors: robot machines that interact with the user demands and communication 3- Facilitator: answers questions that the actor cannot 4- Scientific Facilitator: answers scientific questions that the facilitator cannot The scenarios to work on are three types of functions the robots are able to do: 1- Fetch and Carry functions: where the robot is asked to bring different kind of objects 2- Interaction: including either verbal communication of non-verbal communication with the expressive ability of the robot 3- Migration: shifting the robot's artificial mind from the current physical body to another physical body/ platform. This type of research is one of the topics LIREC interested in and working on [5].

The wide range of audience's age and the variation in their backgrounds insured that various ways of thinking and judgment were involved in tackling the different aspects and level of issues related to the scenario of the play. After the scenario is played, the audience is allowed to ask questions and discuss them with the human players involved in the act. One key aspect is that the actor stays in character during the discussion which insures the continuity of the act situation to the extent that it the gives the impression that it is real so that the audience will not feel disconnected from the illusion of the scenario. This allows them to ask questions about the current situation and when posing hypothetical questions, a concrete and solid based discussion will be triggered.

The Progress of the Wheel

After discussing the stage action, we will now review the reaction of the audience and further talk about how this would benefit the robot development wheel. The audience reactions focused on five categories as following: 1- Intelligence aspect: took the most time of the discussion but it was only discussed once 2- Personalization: also had a higher number of duration but it was infrequent as well 3- The memory aspect of the robot had the lowest frequency among other aspects. The memory aspect of the robot did not fascinate the audience. This doesn’t mean that this aspect is not essential in a robot. Contrarily it is very essential to the extent that it is supposed to exist in a robot system. 4- The communication of the robot: how the robot verbally and non-verbally communicates with humans. This discussion is important in terms of showing what the public thinks the communication should and could be and how it might be developed with the most substantial aspects to focus on. Personalization of communication was discussed along with memory. Meaning that the ability of the robot to speak according to user's preference is something significant and preferred. Audience also expressed that natural language communication with the robot is preferred more than certain keywords communication. 5- The physical embodiment and appearance: also was frequently discussed. Again along with personalization where it showed that having a user defined physical appearance of the robot would be helpful and useful. It is noteworthy to discuss this aspect in the future development of robots. We will also tackle personalizing the appearance aspect in another work in LIREC framework of research.

The Wheel In Motion

The substantial results of this novel experiment using theatre in Human-Robot interaction prove its validity and significance in extracting more information about user preferences especially in what is related to communication, as it is the effective aspect in theatre acting. This doesn’t imply that communication is the only possible aspect to discuss but well incorporation of other aspects of robots through communication scenarios is also a promising option to consider of in the future. It all depends on the structuring of scenarios. Further collaboration with theatre scenarists, psychologists, scientists, and artists to work on such scenarios would create rich discussion scenarios to incorporate in the future planning of these experiments.

Read more: Wizard of OZ Robot house

References: [1] A. Green, H. Huttenrauch & K. Severinson-Eklundh, “Applying the Wizard of Oz Framework to Co-Operative Service Discovery and Configuration”, Proceedings of the 13th IEEE International Workshop on Robot and Human Interactive Cmmunication (RO- MAN 2004), Kurashiki, Okayama Japan, (2004), pp. 575-580. [2] A. Green, E. A. Topp & H. Huttenrauch, “Measuring Up as an intelligent Robot – On the Use of High-Fidelity Simulations for Human-Robot Interaction Research”, Proceedings of Performance Metrics for Intelligent Systems Workshop (PerMIS ’06), Gaithersburg, MD, USA, (2006), pp. 247-252. [3] M. L. Walters, K. Dautenhahn, S. N. Woods, K. L. Koay, R. te Boekhorst, D. Lee, “Exploratory Studies on Social Spaces between Humans and a Mechanical-looking Robot”, Connection Science 18(4) , (2006), pp. 429-442. [4] D. S. Syrdal, K. Dautenhahn, M. L. Walters, K. L Koay “Sharing Spaces with Robots in a Home Scenario – Anthropomorphic Attributions and their Effect on Proxemic Expectations and Evaluations in a Live HRI Trial”. Proc. AAAI Fall 2008 Symposium ‘AI in Eldercare: New Solutions to Old Problems’ November 7-9, Washington, DC, USA, accepted (2008) [5] D. S. Syrdal, K. Dautenhahn, M. L. Walters, K. L Koay “Sharing Spaces with Robots in a Home Scenario – Anthropomorphic Attributions and their Effect on Proxemic Expectations and Evaluations in a Live HRI Trial”. Proc. AAAI Fall 2008 Symposium ‘AI in Eldercare: New Solutions to Old Problems’ November 7-9, Washington, DC, USA, accepted (2008)