This article was updated on the 3rd September 2019.
Ramona Heim / Shutterstock
Infant robots have been widely used in the study of human development. Such robots will help scientists explore the complexity and uniqueness of human development, and in turn, gain a better understanding of a child’s behavior and mind.
In 2011, a robotic infant first circled the internet with its ability to mimic real-life facial expressions and cognitive abilities.
Meet Affetto, the world’s pioneer infant robot designed by engineers at Osaka University, Japan by Hisashi Ishihara, Yuichiro Yoshikawa and Professor Minoru Asada, a team of researchers at the university aiming to construct a child robot with all of the natural facial expressions and cognitive development characteristics that emerge as a direct effect of interaction with a caregiver.
The following video demonstrates Affetto being put through a facial motion test. I must say, at first glance this footage is rather overwhelming, but the level of precision in facial motion shown by Affetto is simply inspiring.
Making Baby Affetto
Up until recently, Affetto has only been a head imitating the physical characteristics of an infant. Research development as part of the Asada project has now given baby Affetto a biomimetic body to work with. Affetto is a flexible infant robot made up of 22 pneumatic actuators to control its body (3 actuators are placed in the neck region to the robotic body, 7 actuators have been used for each arm, 1 actuator has helped control the chest location to the body, and 4 pneumatic actuators have helped control movement of the waist). The following video demonstrates the movement of a prototype upper body for baby Affetto.
The head structure to Affetto has 12 degrees of freedom, which refers to the number of rotations that define the full orientation of the body structure to Affetto. The idea of using actuators that are operated by hydraulic fluid pressure or pneumatic pressure as opposed to electric motors allows for increased flexibility and enhances direct interaction with the human, which is particularly important when considering physical comfort and social interaction with a human caregiver.
Affetto’s face is covered with a soft urethane elastomer gel, making this robot much safer and comfortable to interact with. At present, baby Affetto is only a head and body; intricate structures such as the hands and feet are in development and the entire project is funded up to 2016 at which point there will be an advanced bio-inspired tactile robot structure with the hope of being put into practice to unravel a deeper understanding of child cognitive development.
Another robotic infant, Baby Noby has lifelike similarities with real infants. In 2010, Japanese scientists designed the robot designed to stimulate the development and behavior of a real infant. It has 600 sensors across its body to be able to touch.
Attempts have been made in the past to build cognitive development robots under the JST Erato Asada Project. One example of this attempt is known as Noby (Baby robot); again, this project inspires the idea of using lifelike robots to better understand cognitive development.
Noby was developed with soft skin and tactile sensors to encapsulate the body of the robot to enhance comfort between the human and the infant robot. Noby offers a promising insight on early cognitive development; however, when studied under simulation, caregivers may not treat and interact with the baby robot as though it is human, which could be a result of the caregiver knowing that this interactive robot is not going to grow and doesn’t possess the mind of a human.
It is the social interaction that is important here, but to get to the point where a baby robot contacts a human caregiver in a social setting involves much more than biomechanics. According to brain development simulation research under the JST ERATO Asada Project, advocation of cognitive development robotics involves:
- Simulating brain development using a lifelike robot body
- Image analysis to understand human-robot interaction in real-time
- Application of a muscle-skeleton system to develop a more realistic child robot
- Carrying out behavioral observations on the child robot to analyze the kinematics of rhythmic movement practiced under controlled conditions
What is interesting about the brain development simulation model is the monitoring of brain waves generated in a child robot that is tuned and compared to the brain activity in humans. This research also highlights the possibility of using a mirror neuron system that will help researchers understand the cognitive development of self.
The concept of mirror neurons is fundamental to how we perceive others and our world around us making the adaptation of this type of system in cognitive development robots a crucial point in such research. The basic idea to a mirror neuron is based on a neuronal cell firing a signal both when the body it inhabits carries out a task and when the body observes another object/human/animal performing the same task. Look at the following video to illustrate this concept.
New Robotic Infants Today
There is always the concern that not everyone will adapt well to an infant robot-like Affetto or Noby, possibly because of the lack of realistic childlike facial expressions and overall appearance and so this will be a very sensitive and challenging hurdle to overcome to increase the chances of exploring human-robot interaction.
Future research on cognitive development robots must consider incorporating additional factors into the robotic system to enhance human-robot interaction, including voice, body temperature, the flexibility of the skin, and the ability of the robot to smell, so that this structure is perceived to be more lifelike and more welcoming to interact with a caregiver.
A greater understanding of developmental psychology and neuroscience will be expected from research into advancing cognitive development robot technology, to the point where it may be possible for a caregiver to develop an emotional attachment to a structure that closely resembles life itself.
Today, there are new developments in robotic infants. Sophia The Robot made the waves on the internet because of her artificial intelligence that makes her converse amazingly with humans. Now, she has a little sister, an infant robot, called Little Sophia. Designed by Hanson Robotics, the robotic infant is commercially-available for kids to play with.
The doll has various abilities – a wide range of facial expressions, facial tracking and recognition, interactive chat with user, storytelling, jokes and games, interacts with children while teaching STEM and AI, AR function for the perfect selfie, and walks.
Moreover, there are infant simulators used today. For instance, RealCare Baby 3, formerly known as Baby Think It Over, is the world’s most advanced infant simulator. It is mainly used as a learning aid to add meaning to teach early childhood, infant health and parenting.