The Human-Robot Interaction (HRI) is an evolving field that requires multiple trials and the application of appropriate assessment methods (Bartneck et al., 2008). In addition, with the increasing number of robots being applied in a wide variety of sectors (metalworking, welfare, medical, etc. – continue with examples) it is increasingly necessary to define appropriate methods for the evaluation of HRI. Currently, many of the methods used are borrowed from other fields, such as psychology and social sciences (Kidd & Breazeal, 2005) but, although HRI has elements in common with the area of human-computer and human-human interaction, the wide diffusion of robotic platforms makes it necessary to specialize in evaluation methods for the design of their interaction with humans.

Research on assistive, social, interactive or purely functional robots has shown how human beings tend to create links with robots according to the same mechanisms that regulate the relationships between man and man. This implies the involvement of affective, emotional, communicative and social factors that cannot be quantified in a similar way to the analysis of elements such as usability, effectiveness or efficiency.

One of the first models that analyses the probability of acceptance of a technology but also the influences underlying this acceptance is the TAM (later expanded in TAM 2) developed by Davis (1989) and followed by UTAUT (later expanded in UTAUT 2), developed by Venkatesh et al., (2003) following the study of eight theoretical models based on the application of intention and use as key dependent variables. They are: (1) Theory of Reasoned Action – TRA; (2) Motivation Model; (3) Theory of Planned Behaviour – TPB; (4) TAM; (5) a combination of TAM and TPB; (6) Model of Perceived Control Utilization (PC); (7) Innovation Diffusion Theory; (8) Social Cognition Theory. Regarding the concepts of non-verbal human-robot communication, Kanda et al., (2008) have developed a model to analyze the variability of human friendly behaviors based on the aesthetic appearance of robots. There are, then, factors related to the evaluation of HRI in therapeutic and care settings such as the robot’s behavioral adaptation and flexibility. Moreover, the validity of these methods depends, in many cases, on the truthfulness of the information collected from users, for example, answers given during the administration of questionnaires, etc.). As a consequence, long-term evaluation techniques of the same HRI methods have been developed: Bethel & Murphy (2009) conducted psycho-physiological studies, identifying the correlations between the different signals acquired, in order to obtain reliable and precise results.

With regard to HRI evaluation methodologies with elderly users, Chen & Chan (2014) developed STAM, expanding TAM with specific factors related to people over 65 and, specifically, to gerontechnologies. Heerink et al., (2009) developed the Almere TAM (based on social and functional acceptance) as they did not consider the UTAUT adequately tested for this type of users.

The emotional aspect and the value of empathy in HRI were addressed by Leite et al., (2013) who developed a specific model to assess how much the empathic behavior of the robot positively affects its perception by users. Cramer et al., (2010) also addressed this issue, measuring on quantitative scales the variables of perceived ability, trust (dependability, credibility) and closeness. Moshkina & Arkin (2005) applied ethological and emotional models (i.e. TAME – Traits, Attitude, Moods and Emotions) to evaluate and design long-term human-robot interaction typologies. The correspondence between the user’s personality and that of the robot is a determining factor for acceptance: Tapus & Mataric (2008) proposed an evaluation model based on the learning of the robot and its adaptation to the changing needs, needs and expectations of the user.