M.Sc. Annika Peters

Annika Peters studied computer science (Intelligent Systems) in Bielefeld, Germany. She received her Master's degree from Bielefeld University in November 2008. In her thesis she worked on action structuring and the analysis of multi-modal input (Statistical analysis of characteristics of Infant Directed Interaction with respect to action structure).

In January 2009, she joined the research group for Applied Informatics and the 'Cognitive Interaction Technology - Center of Excellence' (CITEC) at the University of Bielefeld, where she is doing research on non-verbal interaction between robots and humans, especially spatial body/machine movements in Human-Robot Interaction within the Home Tour scenario. The current research platform is BIRON.

BIRON passing and 
spatially prompting 
a person in a hallway

She is member of the CITEC Graduate School and is currently working towards her PhD. She is associated with the CITEC's research project Spatial Movement Concepts in Human Robot shared Environments.  Her supervisors are Marc Hanheide and Franz Kummert.

Current project description:

Spatial Movement Concepts in robot-shared environments (non-verbal spatial human-robot communication)(Citec Column A and C)
When humans and robots share space there is more to the interaction between them than verbal communication. For example spatial prompts can be non-verbal communicative cues. A motion with the entire machine or body (translational, rotational) can be a prompt and therefore is a means of communication.

Having scenarios in mind in which a future robot should help in shops, homes, and offices, a robot needs to detect, react to non-verbal human cues. Based on situational proxemics and spatial concepts, the goal is to make human-robot interaction smooth, not only with users but also with bystanders, by using non-verbal human cues.

Envision a narrow passage (e.g. a doorway, hallway, or a small room) in which two persons have to manage the space around them, e.g., they want to walk in opposite directions. Usually in such situations, humans seem to be well capable of understanding their mutual intentions and negotiate non-verbally how they both achieve their goals, e.g., make room for each other in order to pass the narrow space.

It is immediately evident that bodily communication such as spatial prompting (defined by Anders Green 2009) and proxemic behaviors (E.T. Hall 1969 and others) are a natural means of communication in human-human interaction, their presence and role in human-robot interaction (HRI) has so far mostly been investigated regarding verbal prompts performed by a non-mobile robot towards a human user.

Through a range of conducted observational studies, we found that non-verbal spatial prompting occurs intuitively when humans are prompting robots and that spatial prompting of the robot is perceived as such. Another outcome of the studies is that the detection of human spatial prompts is possible with the cameras and the laser installed on the robot platform BIRON, so that the sensor readings can be anchored with the new robot behavior. Hence, future robots benefit from perceiving and understanding but also from using these prompts to achieve a more intuitive communication about spatial actions in human-robot interaction.

Research is so far based on two robot platforms, BIRON(Bielefeld/Germany) and Snoopy (Lund/Sweden). Both are technical looking robots on wheels. Due to their hight (ca. 130cm) and possibility to move around, detect humans via camera and laser range finder, they are especially used for robot-human interaction studies involving movements with the entire robot.

*BIRON (=BIlefeld Robot companiON): mix of GuiaBot and PatrolBot by MobileRobots, 180° laser coverage, camera
*Snoopy (LTH, Sweden): Pioneer P3-DX by MobileRobots, 360° laser coverage

Close cooperation with Elin Anna Topp at Lund University (LTH), Robotics and Semantic Systems Group, exist.

Very brief description:
My goal is to use non-verbal cues (full body motion, bodily communication (proxemic)) as means
of feedback from the robot towards the human and as general means of communication.

*How should a robot behave/react/signal in situations in which speech
is not the first means for communication?
*Could a robot spatially prompt a human?

In addition I am examining the human part in human-robot interaction.
*Is a human prompting a robot? -- Is a robot able to detect spatial prompts?

Please contact me if you have questions or if you like more information. Contact me if you like to write a Bachelor - or Master Thesis in this field.

Keywords of latest research