Robot morphology and form
When designing robots for human-robot interaction (HRI), it is important to consider both the robot’s function and appearance. The debate over whether function determines form or vice versa does not apply in HRI as the two are interconnected. HRI designers can choose from several types of robots, including androids, humanoids, zoomorphic robots, and minimalist robots. Some designers also create “robjects” that are based on objects rather than living creatures, such as social trashcans or robotic piggy banks. Making appropriate decisions about the robot’s form, function, level of autonomy, and interaction modalities based on the users and context is a crucial aspect of HRI design.
Affordances
The idea of “affordances” is crucial in designing HRI. This concept originated in ecological psychology and refers to the inherent relationship between an organism and its environment. It was further developed by Don Norman to describe the perceivable relationships between an organism and its environment that enable certain actions. A designer needs to incorporate user expectations and cultural perceptions while making the product’s affordances explicit. Design affordances are important in developing common ground between robots and humans, allowing people to understand robot capabilities and limitations appropriately and adapt their interactions accordingly. A robot’s appearance and interaction modalities are important affordances that people tend to assume are commensurate with its capabilities. For example, if a robot looks like a human or has eyes, people expect it to act like a human or see. The affordances used by designers are particularly important for signalling appropriate ways of engaging with robots because they are novel interaction partners.
Read more about the concept of affordances:
https://dl.acm.org/doi/10.1145/1409040.1409044
http://www.worldcat.org/oclc/896794768
Design patterns
The task of HRI design involves not only creating a robotic platform but also enabling certain interactions between humans and robots in various social contexts. This means that HRI designers need to consider not only the characteristics of individual robots, such as appearance, sensing abilities, or actuation, but also what Peter Kahn calls “design patterns” in HRI, which are inspired by Christopher Alexander’s idea of design patterns in architecture. These patterns describe a problem that occurs repeatedly in our environment and provide a core solution that can be used multiple times. In HRI, designers should develop abstract patterns that can be combined and integrated to describe interactions with the social and physical world. For example, a didactic communication pattern, where the robot assumes the role of a teacher, could be combined with a motion pattern to create a robotic tour guide. These patterns can be developed through an iterative design process based on observation of human interactions, prior empirical knowledge, and designers’ experiences with HRI. The focus of design should be on the relationship between humans and robots.
Read more about Christopher Alexander’s design patterns in architecture:
https://www.worldcat.org/title/961298119
Read more about Peter Kahn design patterns:
https://dl.acm.org/doi/10.1145/1349822.1349836
Design principles in HRI
The process of designing human-robot interactions (HRI) involves considering not only the physical characteristics of robots but also the design patterns that facilitate appropriate interactions between humans and robots in various social contexts. Rather than focusing solely on the physical appearance and capabilities of a robot, designers should take into account how they fit into and convey specific human-robot interaction (HRI) design patterns. They should also design the robot’s features and capabilities as affordances that accurately convey its intended interaction capabilities and purpose. To this end, researchers in HRI have proposed the following principles that Bartneck et al. summarizes in their book that should be considered when developing appropriate robot forms, patterns, and affordances in HRI design.
Matching the form and function of the design: For example, a humanoid robot would be expected to behave like a human, whereas a cleaning robot could have a less anthropomorphic design. Design can also prompt people to associate specific social norms and cultural stereotypes with robots. For instance, people might expect a female robot to be more knowledgeable about dating or a robot made in China to be more knowledgeable about tourist destinations in that country.
Underpromise and overdeliver: The concept of “Underpromise and overdeliver” is important in designing robots, as raising people’s expectations of a robot’s intelligence or companion-like nature can lead to disappointment if those expectations are not met. To avoid negative evaluations of robots, it is better to decrease people’s expectations, possibly by not calling the design a “robot”.
Interaction expands function: People tend to fill in the blanks in their interactions with robots, so designing them in an open-ended way can be useful, particularly for robots with limited capabilities. This approach has been successful with the seal-like robot Paro, which is interpreted as a pet-like character despite its limited abilities.
Do not mix metaphors: The design of a robot should take into account all of its components, such as its abilities, actions, and interaction possibilities, in a cohesive manner. If a robot is made to look like a human, it may be unsettling if only some parts of its body are covered in skin. Similarly, if a robot is made to look like an animal, it would be unusual for it to speak like an adult or to teach math. This is associated with the concept of the Uncanny Valley, which suggests that when abilities, behaviour, and appearance are not appropriately matched, people tend to view the robot negatively.
To further read about design principles in HRI you can explore this book:
https://www.cambridge.org/core/books/humanrobot-interaction/2C042DEB4D0ECFFA5485857314E885BC
References
Bartneck, C. et al. (2020) Human-Robot Interaction: An Introduction. Cambridge: Cambridge University Press. Available at: https://doi.org/10.1017/9781108676649.