Affective human–robot interaction

 

 

 

Human-machine interaction in intelligent sensor network (Minotaurus) is a joint project between the Biomimetics and Intelligent Systems Group and the Center for Machine Vision Research, funded by the European Regional Development fund in 2010-2012.

However, the human-robot interaction has been an active research area at the BISG since early 2000'. Parallelly, the same research area, in particular the affective human-robot interaction, is being studied with the support of the Academy of Finland.

The goal of the currrent project is to create natural communication between the human and the machine. An  intelligent sensor network will be implemented to support the operation of humans and robots.  Practical applications for human-robot interaction will be implemented as well. In addition, the HRI system will be made easily extandable.

 

Demonstrations:

 

 

Intelligent sensor network
•Generalized representation for information
•Sensor information
•Communication if system components
•Information fusion from various sources
•Real-time models of state of the environment
•Visualization of the state
•Sharing services and distributed intelligence

Sensors: Cameras, microphones, motion detectors

 

 

Machine vision and camera network
•Multi-purpose vision system
•Camera network
•Human related vision algorithms:
•Detection of faces
 
 
•Recognition of emotions and expressions
•Detection of gesture movements
•Visual commands to robot (e.g. Pointing objects,...)
 
Human-robot interaction
•Multi-modal communication
•Acceptable  by humans
•Supporting sensor network
•Implementation technologies for the robot
•Speech to text -> text dialog -> text to robot’s speech
 
 
•Other inputs: vision system, reasoning from other inputs

 

Information representation and integration
•How to glue everything to one real-time operating system
•Distributed stand-alone modules
•Key technology is the information representation
•How to represent past, present, future, and previous knowledge
 
Mobile robot applications

 

Contact person: Antti Tikanmäki

 

 

 

 

 

 

Last updated: 17.10.2013