Applied Informatics Group

Universität BielefeldAIWEB

Overview of Research Topics

Group Description

For our research on social and cognitive robots, we draw from the following areas.

  • Speech-Based Dialog

    We see speech-based dialog as a key aspect for robots to engage in a successful interaction with the human users. Besides investigating dialog models, we also focus on adapting speech recognition techniques to the situational context.

    Users interacting with the NAO robot.

  • Visual Scene Interpretation

    Visual scene interpretation is an important ability for intelligent systems. We realize this ability as a complete loop from extracting bottom-up information from visual data, fusing these information with top-down knowledge, and improving bottom-up processing by feeding back interpretations as relevance or context information.

  • Systems Engineering

    Mastering the engineering challenges inherent to the creation of integrated robot systems with their underlying software and system architectures is a particular goal of the engineering research conducted by the Applied Informatics group.

  • Perception of Humans

    Multi-modal perception of a robot's environment is crucial for interacting naturally with human partners in non-laboratory settings. We apply multi-modal perception of activities on a global and local human-centered level.

  • Human Social Interaction & HRI

     

    To enable robot systems to engage in interaction with the human user, we need to understand human multimodal communication and sequential structures in authentic interaction. Inspired by HHI, we develop (adapted) interactional models for robots, which are implemented in autonomous systems and evaluated in HRI studies focusing on the users’ interpretation and the technical performance. 

  • Interaction Design

    Interaction Design is broudly about shaping robots for people´s use in order to improve its perceived and actual usability. To increase the human-robot interaction, we thereby focus mainly on two aspects: the robot´s aesthetic form and its social behaviors.

  • Evaluation of HRI

    One of the aspects we are particularly proud of here at the Applied Informatics group is that we carry out many of our evaluations on autonomous systems in real-world environments, such as regular apartments, in competitions (e.g., RoboCup@Home), and in realistic, autonomous settings in our scenario-specific labs. The often-seen Wizard-of-Oz approach is, for us, also important, but primarily as a stepping stone towards autonomy. Moreover, we also perform a variety of studies on Human-Human-Interaction, to gather input for system design and requirements. The following describes our portfolio of methods and approaches to conduct such evaluations.

  • Learning by Interacting

    Interaction with a tutor plays an important role in the development of children. In order to learn in an interactive scenario, robots require flexible methods, since it is not possible to predetermine all tasks a robot would be involved in.

For members