Social robots

NR conducts research about the use of social robots for educational purposes. Our objectives are to enhance learning experiences for specific cohorts and to create better levels of understanding between groups.

NR currently has several projects where social robots are a central component of our research.

We have studied how social robots can be used to improve Norwegian vocabulary to children with minority languages in daycare centers. We are currently studying how social robots can aid in social and emotional communication for children with autism spectrum disorder (ASD).

We are also working on how to improve a robot’s communicative skills by fine-tuning its dialogue manager. A dialogue manager decides what the system will do or say at a given time, such as respond to a user or make a movement.

Robot-assisted language learning with children

Norwegian language proficiency increases the likelihood of a person completing schooling and upper secondary education. In addition, improved language skills can contribute to integration and reduce social inequalities.

In 2020, we evaluated the outcome of robot-assisted language learning among four to six-year-olds at two kindergartens in Grorud, a city district in Oslo. The trial, which lasted for six weeks, showed that children with various minority language backgrounds learned 12% more words than the control groups without robots.  

NR developed the software that controls the robot and a mobile app for language learning. The app was created to enable learning at home, i.e., without a robot, and in kindergarten where the robot and app were connected.

The image is a black and white illustration of a robot facing three small children under a sign that says 'Day-care center'- To the right there is a polka dot skirt. The image features text wherein the robot asks the children, "What do you see?" and the children respond to the robot's prompts: "Hehe, look at that!," "A skirt," "A shirt, maybe?". The robot is also shown offering correctio when the children provide an incorrect answer, "Bat is wrong."
Caption: A potential scenario of human-robot interaction at a day-care center

NR can design comprehensive solutions and:

  • map pedagogical and functional requirements and means for interaction between children and robots;
  • develop relevant content that is rooted in existing methodologies for language teaching;
  • implement prototypes;
  • evaluate and test prototypes with suitable methodologies, and assess learning experiences and effects;
  • examine various ethical challenges;
  • conduct in-depth research of deep neural networks that control the way robots use language.

Current projects: 

  • ROSa: 
    Robot-supported education for children between six and twelve-years-old with autism spectrum disorder. The project supports one-on-one learning and covers language, social skills and communication.
  • Graphdial: Human-robot interaction using voice communication.

Completed projects: 

  • ROS: Robot-supported language education in kindergartens. Project period: 2018-2021.
  • Sinoe!: Systematic Norwegian language education for children with ASD transitioning from kindergarten to primary school using social robots. Project period: 2017-2018.
  • A prestudy of social robots in kindergarten examining the potential application of social robots in language teaching. Project period: 2017.