Intelligent robots in everyday environments
There are a growing number of everyday applications for artificial intelligence and robotics such as social robots, self-driving cars, and automated shops. We expect that future society will be ubiquitous with various robots. Towards such a robotized society, we aim to study robots that interact with people and operate in harmony alongside them. Such robots are embodied agents with sensory feedback that have the ability to interact in real-time with their environments. Consequently, we conduct fundamental research in intelligent robotics, human interaction, sensor networks, and artificial intelligence (AI). We are aiming for applications in the service industry, elderly care, health care, collaborative work places, and learning. Below are some example research themes:
Various sensors and IoT (Internet of Things) devices will be ubiquitous in our daily life. We investigate techniques to instrument distributed sensor networks to recognize human social behaviors; e.g. state-of-the-art approaches such as deep neural network (DNN). This enables us to observe peoples’ daily behavior and build models from observed “big data” to provide a comprehensive insight about collective social behavior; e.g. studying one year’s worth of data from pedestrians in a shopping mall. Furthermore, we integrate a sensor network and robots to develop a sophisticated network robot system with “ambient intelligence”.
We made datasets from a shopping mall available, e.g. people trajectories and their interaction with a robot <link>
Robots can serve as a natural interface for computer/IoT systems. Hence, we conduct a series of studies for human-robot interaction. One focus is endowing “human likeness” to a robot (both behaviorally and computationally), which will allow it to interact with people in a natural manner. For instance, we have realized human-like interaction using gaze and gesture.
Our recent challenges aim to investigate methods to allow a robot to understand human intuition, and make it interact with people in a morally acceptable way. Further, we investigate the social acceptance of robotic systems “in the wild” with real-world field studies.
Cyber-physical system (CPS)
We study modeling techniques that reproduce relevant human behaviors in simulation, so that a system can anticipate how its behaviors can potentially influence people. For instance, we have developed a pedestrian model and simulated their behaviors around a social robot.
With this simulation a robot can anticipate the gathering of a curious crowd around itself (see the Figure), or potential bullying from nearby children, so that it can avoid problematic situations before they occur.
Here are listed past and present research projects with the research funding provided by Japanese government and governmental funding agencies.
JST CREST “Human-Robot Interaction for Symbiotic Robots in a Public Space in a City” (2017-2023)
Since artificial intelligence and robotics technologies have recently grown fast, it is expected that diverse human works will be eventually replaced by
artificial intelligence and robots. However, although many works expected to be automated in reality also play the role of bringing safety feelings
to our society, the robots today are yet far from providing such safety feelings. They even fail to be morally considered. This project aims to solve
this problem by providing moral interaction capability to robots, and realize a symbiotic society between people and robots. < link >
JST CREST “Enabling a mobile social robot to adapt to a public space in a city” (2010-2016)
This project is aimed to realize a future city environment in which a social robot roams in a public space and offers useful services, such as information-providing and physical assistance. In order to realize this project, two kinds of research will be conducted. The first one will be aimed at achieving a technique based on a sensor network that will be able to understand how the humans use a public place. The second investigation will be aimed at enabling a robot to roam the environment without hindering the movement of people, while talking kindly to walking people. < link >