Human Centric Robotics
developing robots that understand the neuroscience and control underlying human behaviours
Machines and especially robots that interact with humans need to conform to not only aesthetics and human physiology but also to the human behaviours. This is a big challenge for robots because, relative to human physiology and biomechanics, the understanding of human behaviours is still lacking.
Humans have complex interaction dynamics. Human interaction behaviours are determined, not only by their instantaneous sensory observations but also by predictions developed by behavioral models they possess of their interacting agent. These models are continuously adapted with daily experiences. Furthermore, both the models and their adaptations vary with an individual’s cultural background, age, emotion and disease.
To interact efficiently with a human, a robot needs to understand how the sensory observation and models affect human behaviour, and in the very least be able to behave in a way that the human can understand and quickly adapt to. Clearly, achieving this goal requires improvement not only in robot control but also in our understanding of human interaction neuroscience and psychology.
My research thus includes active, simultaneous research (and collaborations) in robot control and behavioral neuroscience with the aim of providing interacting robots with a comprehensive understanding (in the form of mathematical models) of the human behaviour which they can utilize to efficiently and safely interact with humans. Under the general heading of human centric robotics, my research broadly investigates three fields.
Human motor Neuroscience
Investigating the neuroscience behind human interactions with their environment and fellow humans to understand:
· how humans utilize various visual and haptic signals to control their interactions with their environment and other humans.
· how these interactions affect subsequent human behavior.
· the computational processes underlying these behaviours.
Cognitive factors affecting human interactions
Investigating the cognitive factors that affect human interactions. To understand
· the implicit processes governing perception and action by humans.
· the explicit processes modulating human behavior.
· the computational processes underlying these processes
Developing new robot behaviors
Utilizing the models of human behaviors for:
· developing automatic robot behaviors during human-interactions.
· developing new human– machine interfaces with a better understanding of humans.
· developing robots/machines that humans can adapt faster to, and feel ‘comfortable’ with.
· developing bio-mimetic control and learning by robots.
Current/ Recent Funding
· Japan Society for the Promotion of Science (Kakenhi ‘Challenge’)
· Japan Science and Technology Agency (JST)
· Japan Society for the Promotion of Science (Kakenhi Kiban B)
· EU FP7 Koroibot
· The Leverhulme Trust fund, UK
Other works have designed machines to conform to the human physiology and aesthetics. My work aims to conform machine design and behavior to human behavior.
Dr. Gowrishankar Ganesh,
Senior Researcher (CR1),
CNRS-AIST Joint Robotics Laboratory,
Tsukuba, Ibaraki 305-8568 JAPAN
gans_gs AT hotmail . com