One of the major challenges in developing social robots is making them compatible with the expectations and requirements of the human mind, according to Maja Matariæ, Director of Center for Robotics and Embedded Systems of the University of Southern California.
Matariæ stated that human interaction with embodied machines is different when compared to interactions with a cell phone, computer or other intelligent devices. It is important to know these differences for developing a social robot.
Matariæ has designed social robots for several therapeutic applications. According to him, a social robot can be considered as efficiently designed only when it is able to communicate both verbally and physically through body language and facial expressions. Embedding the correct personality is also equally important. When the personality of the social robot was matched with that of users, people enjoyed their rehabilitation exercises and carried out for longer period, Matariæ added.
Another solution is to match the appearance of a social robot according to the perception of people on its abilities. Ayse Saygin, a faculty member at the Kavli Institute of Brain and Mind and an Assistant Professor at the University of California San Diego, and her colleagues studied the ‘action perception system’ in the human brain when it is adjusted with respect to human motion or appearance. What they had discovered was humanlike look is equally important to humanlike actions.
A social robot must also have the capability to learn socially. Andrea Thomaz from the Georgia Institute of Technology developed a robot capable of learning from humans the manner a person does, such as through speech, demonstration, observation and social interaction.