Home Research



The focus of our research lies on robots acting in human environments. We develop techniques that allow robots to adapt their behavior to the environment and to the surrounding people, thereby exploiting semantic information about objects and information about the activities of users.

In particular, we have introduced several novel methods for environment modeling as well as for planning navigation and manipulation actions for wheeled and biped robots. Among them are techniques for 3D environment perception and exploration, footstep planning, constrained manipulation planning, human-aware navigation, and imitation of human whole-body motions. Currently, we especially focus on motion planning and navigation through cluttered and dynamic scenes as well as on generating foresighted robot behavior by predicting human motions over a longer time horizon.

The lab has been involved in several projects with international partners. Our work is currently supported by the Research Unit Anticipating Human Behavior and the Cluster of Excellence PhenoRob--Robotics and Phenotyping for Sustainable Crop Production funded by the German Research Foundation. Within Phenorob, we investigate methods for coverage of individual plants with a robot's sensors to learn a 4D model (3D+time) to monitor the plant growth and further factors. Here, we build upon methods we developed for the exploration of indoor scenes that include the articulation of objects. Furthermore, we will contribute perception and motion planning techniques to the new H2020-FETOPEN project RePAIR -- Reconstructing the Past: Artificial Intelligence and Robotics Meet Cultural Heritage, which will start in 2021.

Previously, Maren Bennewitz was PI in the projects SQUIRREL, ROVINA, and First-MM funded by the EU as well as in the Cluster of Excellence BrainLinks-BrainTools, in the Transregional Research Center SFB/TR8 Spatial Cognition, and in the research training group on Embedded Microsystems, all funded by the German Research Foundation.