In many applications mobile intelligent robots cannot rely on foreknowledge about their environment for orientation. Being able to locate in undefined surroundings and the precise mapping of the environment is therefore crucial for mobile robots. This project tackles the development of a mobile robot system that is able to autonomously navigate in unknown outdoor environments to cooperate with human interaction partners. Interactions are proactively initiated by the robot by analyzing potential communication partners for interaction willingness.

The robot’s capabilities comprise

• Mapping, localization and navigation in unknown environments
• Motion estimation and tracking of dynamic objects
• Detection and identification of interaction partners
• Detection of interaction willingness by analyzing body and head pose, as well as facial expressions

Simulatenous localization and mapping

• Mapping, localization and navigation in unknown environments
• Map and estimated position has to be robust to minimize drift over time

Visualization of the Tiago robot building a 2d grid map (laser based) and a 3d point cloud map(camera based) of its environment (a lab).

Collision Avoidance

• Dynamic environments are natural conditions in real world scenarios
• Therefore it's crucial to detect possible collision threats and act appropriately

Publications

• Marc-André Fiedler, Michal Rapczynski, Ayoub Al-Hamadi
  Fusion-Based Approach for Respiratory Rate Recognition From Facial Video Images
  IEEE Access PP(99):1-1, July 2020
• Thorsten Hempel, Ayoub Al-Hamadi
  SLAM-Based Multistate Tracking System for Mobile Human-Robot Interaction
  Image Analysis and Recognition, 368-376, Springer International Publishing, 2020
  
• Thorsten Hempel, Ayoub Al-Hamadi
  Pixel-Wise Motion Segmentation for SLAM in Dynamic Environments
  IEEE Access 08:164521 - 164528, September 2020, DOI: 10.1109/ACCESS.2020.3022506

Contact:

Thorsten Hempel, Marc-André Fiedler, Ayoub Al-Hamadi