Towards Empathic Touch by Relational Agents - Research Paper Example

Towards Empathic Touch by Relational Agents Author(s) Abstract— Humanity has reached a stage where human-computer interface has reached the threshold of magical possibilities. While designing relational agents, incorporating the sensation of touch as a communication agent, has been a recent development. The innovative experiments so far been conducted in this area are reflected and represented in the three researches discussed in this study. The comparison among these studies has unquestionably proved the positive impact of healing touch administered by a computational device, particularly for hospital patients. But the conclusion is that the technology needs to be more fine tuned. Index Terms—relational agents, haptic, touch, empathic touch. ——————————  —————————— 1 Introduction Emotional healing and even physical healing, to an extent, have a subtle connection with human touch. That is why when we try to console someone in her grief or pain, we unintentionally touch her. Even in clinical practice, touch has been acknowledged to have a role in improving health outcomes [1]. In the age of modern technology adapting to the needs of modern medicine, relational agents can be considered as the best interface between technology and humans. Relational agents have been defined as computer agents used in “health care applications” and “designed to establish long term, social-emotional relationships with people” [2]. This study is an attempt to understand the importance of empathic touch by relational agents by comparing the inferences of three research papers on the same and related topics. 2 THREE PAPERS ON TOUCH AND RELATIONAL AGENTS 2.1 “Touchbots” The first study under scrutiny was carried out by Bickmore, et al., which analyzed the use of a relational agent which is designed in such a way that it physically touches its users “in synchrony with speech and other nonverbal communicative behavior” [2]. Functioning of this agent, named, “touchbot”, as a device that hospital patients can hold on to, while they are in hospital beds, is the topic of examination in this paper [2]. This device is able to sense communication signals like squeezing, stroking or pressing hands and respond to these signals in an empathetic way [2]. A speech-based dialogue system is also part of this agent which consoles, advices and give information to the user [2]. The inferences of a primary study, which examines the capability of the “agent to communicate affect using only the touch actuator” and another study which uses “touch”, “animated face” and “synthetic speech” all together are compared by these researchers [2]. How each of these communication elements rate as affects, as felt by the users are then examined [2]. Along with these two experiments, a third experiment has also been conducted by these authors, in which, “the ability of empathic touch to establish a sense of closeness with a relational agent” is tested [2]. 2.2 “Touchbots” and Individual Variations The second research that is being scrutinized here, is the one carried out by Timothy Bickmore and Rukmal Fernando [3]. This study has focused on the user responses towards “touchbot.” On technology side, the “touchbot” used in this study was less developed than that discussed in the first paper. Only one touch pattern (2 slow inflations of 200 ms duration at an interval of 750 ms) was designed into the system while conducting this study. A “single, brief counseling dialogue about cancer” was the only speech part used [3]. And this too was not used in combination with touch. It was used as a comparison modality [3]. A “16-Likert item self report scale” was used for measurement of responses [3]. And touch receptivity was also measured by using another “10-Likert item composite self report scale.” [3]. 2.3 Touch-sensitive Apparel The third research paper considered in this enquiry is by Cati Vaucelle and Yasmine Abbas [4]. “Touch-sensitive” is the name given by the authors to the apparel that they have designed which is “comforting and alerting apparel with a feedback system” [4]. It is a “matrix made of clothing elements that allows diffusion of tactile information through heat sensors, mechanically-driven textual sensation and liquid diffusion” [4]. This clothing also diffuses aromatic essential oils to the skin surface of the wearer [4]. 3 Methods of study 3.1 “Touchbots” In the first study, the “touchbot” touches users on their hands [2]. For making the touch to be “comfortable and organic”, the haptic output device has a glove on it “with an air bladder sewn into the palm” [2]. The design of the device helps it to synchronize its dialogues with the mode of touch [2]. Around 27 unique touch gestures comprised the emotional range incorporated into the device [2]. Intensity and duration of touch was also manipulated in this [2]. 3.2 “Touchbots”- Individual Variations In the second paper, an important aspect was that apart from user responses, the biological reactions of the body of the users were also assessed. The researchers have revealed that “user heart rate and galvanic skin conductivity were recorded continuously at 256 Hz, using finger-clip sensors.” [3]. Twenty one subjects participated in this study. 3.3 Touch-sensitive Apparel The third research discussed above draws its ideological base from the proven fact that massage is a soothing experience for all human beings and also that massage involves a user and a practitioner. There are also simple machines used for massage like mini massage rollers and this prompted the researchres to believe that a non-human system like an apparel can also serve the purpose equally well. The previously designed hug-over-a-distance jackets and huggable sculptures had also inspired this more innovative research [4]. Four low fidelity prototypes have been deviced by the researchers in modeling touch-sensitive apparel. The low fidelity prototype-1 has “metallic wires sawn into the fabric” and it shrinks mechanically as if in a hug [4]. The low fidelity prototype-2 has textural silicon buttons which “act as pressure points to the person who receives the massage” [4]. The low fidelity prototype-3 has “manually inflatable venyl pockets” and low fidelity prototype-4 has the essential iol diffusing function as its major feature [4]. The “growing miniaturization of computational components” is the most important development that enabled this innovation [4]. There is also an embedded microphone or headphone inside the clothing which helps thecustomer to control the whole mechanism [4]. 4 Critical Analysis Natural languages used by way of text and speech as well as facial expressions have been the major communication channels used by computer agents and especially relational agents, so far [2]. There has been not much effort to incorporate the possibilities of physical touch in this area. The studies conducted using dyads and haptic devices had failed to establish touch as a mediator in electronic communication [2]. And no research has been so far done with conservational touch in this field [2]. It is in this context that the effort to create a touching agent gains significance. This attempt can surely be termed as a breakthrough in the progress of clinical-technological interface. But to say with total conviction that this innovation is technically sound, there need to be a lot of follow up research and also technological upgradation and fine tuning. The study by Salminen et al. has relevance in connection with “touchbot” as it had used a “fingertip Stimulator” and tried to find out specific “patterns of motion” associated with certain emotions that comes with touch [6]. The study had been partially successful in finding out such patterns. This result can be used in further fine tuning of the robotic palm by which it can do acts like patting, caressing, stroking etc. rather than simple squeezes delivered now [3]. 5 RESULTS Twelve persons participated in the first study discussed in the first paper, involving touch alone. The participants were found to attach same emotional labels to 5.8% of items which were played upon them more than one time. But this study failed to make out connections between emotions and touch gesture. But the participants were able to relatively rank each touch gesture, roughly in connection with a particular emotion. In the second study, in the first paper, which used multi-modalities including touch, facial display and speech, again 12 participants were there. In the results, the facial display scored above speech and far above touch. The facial display paired with speech had a pretty good result while facial display paired with touch diminished the result. But a combination of the three gave significant positive results. The third study in the first paper is on the effect of empathic touch on user-agent relationship. This study showed that addition of touch to user interface like a computer agent enhances the relationship, though partially. One major finding of the second paper was that there is a gender differentiation in responses to touch by relational agents. Females showed more positive responses to touch than males according to this study. But a majority of respondents had termed the touch interaction with the “touchbot” as “weird” and “awkward”. There were mixed reactions about whether the touch was natural. And most of the respondents said they felt that the hand was not part of the agent. The major achievements of the third study are the development of “a flexible structure, a mechanism of diffusion, and a feedback system for alerting and comforting the user through haptic means” [4]. This study also proved the possibility of designing self care systems through this kind of equipments. 6 MEANING AND SIGNIFICANCE The results of the first paper show that there need to be more subtle variations of touch incorporated into the “touchbot” system. The temperature variations that occur in a human palm in connection with each emotion or similar minute biological aspects have to be used in further re-designing the “touchbot.” Also, it has to taken into consideration that not a single sensation like touch or speech evokes a response as strong as a combination of sensations, as used in the second study in the first paper. The results of the second paper indicate that not everyone is equally receptive to touch. There can be individual variations to a sensation like touch which is very intimate in its nature. There will be persons who like being touched and those who do not. When touch is administered by a person who is not intimate to the person being touched, the sensation can be disturbing or awkward as described by the respondents in the second research paper. But there needs to be further study to find out that if the touch is not too intruding when used alone as in this case, and if it is better to use touch clubbed with other sensations so that more desirable results could be achieved. 7 ethical implications The IEEE code of ethics makes it mandatory “to accept responsibility in making decisions consistent with the safety, health and welfare of the public, and to disclose promptly factors that might endanger the public or the environment” in its code-1 [7]. It also demands a mandatory obligation from whoever follows it “to improve the understanding of technology, its appropriate application, and potential consequences” in its code-5 [7]. These two codes are followed in all the studies discussed above. By giving proper information on the new technologies discussed without hiding their limitations, the first code is adhered to. The technologies discussed are evidently for the welfare of the public, as is instructed in the first code. The second code is also followed by the researchers as they has proposed to do further research on the topic and thereby improve and understand the technologies adopted in a better way. 8 FUTURE In future research, as Bickmore, et al., have suggested, the use of, “Shape Memory Alloys” and “robotic hands with much higher degrees of dexterity” could play a big role [3]. The conversational touch used by humans itself remains a less explored area as yet. Hence, studies in that direction also could supplement the dream of realizing a soothing human-computer interface. 9 conclusion The role of touch in human-computer interaction is going to be a topic of increased interest in the future as well. Invention of new technologies will add momentum to the enquiries in this area. The miniaturization of computing elements and new developments in robotic technology is going to play a key role while designing this kind of relational agents. Research on conversational touch and emotional responses to touch will be another area which will have to be clubbed with the technological research processes. Thus the relational agents-research can be viewed as the bridge that deletes the gap between human intellect and emotions. Acknowledgments The author wishes to thank all the friends and well wishers who have assisted in this research. This work was supported by --------- References [1] W. Reynolds, The measurement and development of empathy in nursing, Aldershot: Ashgate Publishing, 2000. [2] T.W. Bickmore, R. Fernando, L. Ring, and D. Schulman, “Empathic touch by relational agents”, IEEE Transactions on Affective Computing, Issue:1 (1949-3045), 2010. http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5539766, accessed 29/09/2010. [3] T.W.Bickmore and R.Fernando, “Towards empathetic touch my relational agents”, Proc. of 8th Int. Conf. on Autonomous Agents and Multiagent Systems, (AAMAS 2009) Decker, Sichman, Sierra and Castelfranchi (eds.) May 10-15, 2009, International Foundation for Autonomous Agents and Multiagent Systems. (www.ifaamas.org). [4] C.Vaucelle and Y.Abbas, “Touch-sensitive apparel”, CHI 2007, April 28-May 23, 2007, San Jos, USA. [5] S. Brave, C. Nass, and K. Hutchinson, “Computers that care: Investigating the effects of orientation of emotion exhibited by an embodied computer agent”, Int J Human-Computer Studies, Vol.62, pp. 161-178, 2005. [6] K. Salminen, V.Surakka, L.Lylykangas, J.Raisamo, R. Saarinen, R.Raisamo, J. Rantala, and G.Evreinov, “Emotional and behavioral responses to haptic stimulation”, In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, (CHI’08) 2008, pp.1555-1562. [7] IEEE, “IEEE code of ethics”, Approved by the IEEE Board of Directors February 2006. http://www.ieee.org/membership_services/membership/ethics_code.html accessed 28/09/2010. Read More