I am interested in robust spatial perception for interaction and collaboration.
I am particularly interested in questions of perception in the context of interaction with ( semi-)autonomous systems, which explicitly concerns both directions of perception: the perception of the environment by the system, as well as the perception of the system by humans and how these depend on each other. The central theme of my research group is the mutual understanding of human and robot, which is what makes true collaboration possible.
I understand robust spatial perception to be a continuum of methods addressing the combination of simultaneous localization and mapping (SLAM) with semantic concepts, dynamic aspects (e.g. human motion), as well as object recognition, while taking into account the presence of outliers and false positives on all levels. Robust methods are ideal for safety critical and commercial devices as their goal is not only to be resilient to variations in input data but also to non-optimal parameter settings or user errors due to non-expert users. In the grand vision of intelligent autonomous helpers for everyday tasks operating with humans in a multitude of environments, these are critical capabilities.
Similar considerations apply to applications in other fields: In service robotics in consumer or nursing settings, where robots should act within the same workspace as humans, robustness to user error as well as robust perception of humans, their actions, and related objects is especially important. In safety, security and rescue robotics (SSRR), robustness to clutter and significant changes in appearance as well as drift and wheel slip on uneven ground is required. Autonomous driving applications, especially in GPS-denied settings, require robust and safe localization and perception of surrounding objects under the influence of lighting and weather changes as well as perceptual aliasing.
Conversely, actions taken by the system may not always be interpretable to a human. Therefore, the system must facilitate the perception by the human through technical means, for example by augmented or virtual reality. We develop interactive visualizations and close the loop to planning and control methods.
In my current role as group manager and principal scientist for Human Robot Collaboration at the OFFIS Institute for Information Technology, I apply these concepts in robotics applications in close proximity or direct cooperation with humans. These applications occur in industrial and nursing/consumer settings.
Ph.D. Degree in Robotics under Prof. Dr. Andreas Birk. Dissertation Topic: “Generalized Simultaneous Localization and Mapping (SLAM) on Graphs with Multimodal Probabilities and Hyperedges”. Passed defense in May 2013 with special distinction. Thesis published March 2014.
Two year M.Sc. Program in Artificial Intelligence, specializing in Multimodal Intelligent Systems. Graduated with Honors in December, 2006.
Three year B.Sc. Program in Electrical Engineering and Computer Science.
I am the founding manager of the research group on Smart Human Robot Collaboration, which aims to enable close collaboration between people and (semi-)autonomous/automated systems.
Working on human motion tracking and related sensor fusion in the context of collaborative robotics in industry.
Working on full-body human motion tracking and behavior recognition in an industrial scenario.
Co-wrote proposal, and now responsible for Work Package on Underwater Perception and Modeling as Work Package Leader. Tasks include 3D Localization and Mapping, as well as object modeling using multi-view stereo and sonar systems.
Responsible for coordinating activities of the group and working on robust 2D and 3D SLAM using underwater monocular or stereo vision in dynamic scenes (divers and fish), as well as object detection and recognition (diver body parts) using visual and sonar sensors.
Responsible for coordinating activities of the group and contributing in the areas of 2D and 3D underwater multi-robot robust SLAM with sonar and visual sensors.
The SpaceBot Cup competition was funded by the German Aerospace Center (Deutsches Zentrum für Luft- und Raumfahrt, DLR) and the German Federal Ministry of Economics and Development. Responsible for coordinating the team of 10 PhD students and another PostDoc, as well as the 3D SLAM subcomponent. Most successful team at the competition.
Working on 3D perception and modeling.
Responsible for underwater visual 2D multi-robot SLAM, some coordination of group activities, as well as organizing several summer schools within the scope of the project.
Safety in Production, Lecture in the International Summer School on Human Computer Interaction in Safety Critical Environments, OFFIS Institute for Information Technology, Oldenburg, July 2019
Robuste Umgebungswahrnehmung in der Robotik, Talk at the Department for Electrical Engineering and Computer Science, University of Siegen, Germany, September 2018
Robust Spatial Perception for Underwater Robotics, Talk in the Brown Bag series at the Robotics Innovation Center, DFKI, Bremen, Germany, December 2016
Towards Robust Perception for Mobile Robots, Talk at the Institute for Robotics and Process Control, Technical University Braunschweig, Germany, November 2016
Robuste Lokalisierung und Kartenerstellung: Methoden gegen Ausreißer und Mehrdeutigkeit, Talk at the Institute for Computer Engineering, Universität zu Lübeck, Germany, June 2016
Robust Perception for Underwater Robotics Applications, Talk at the Computer Science Institute, Universität Bonn, Germany, January 2016
Robust Perception in Underwater Robotics, Talk in the seminar series of the Institute of Sensors, Signals and Systems at Heriot-Watt University, Edinburgh, UK, October 2015.
Underwater Human-Robot Collaboration, Talk reporting on the ongoing CADDY project at the IROS2015 Workshop on Bioinspired Underwater Robotics, Hamburg, Germany, October 2015.
Ambiguities in Simultaneous Localization and Mapping, Talk at the Computer Science Institute, Freie Universität Berlin, Germany, April 2014.
Anticipation in Robotics, Invited guest lecture in the Master module “Anticipation in Cognitive Literary and Cultural Studies” at Carl von Ossietzky Universität Oldenburg, Germany, October 2012.
Online SLAM with Visual and Sonar Data, Invited lecture with Heiko Bülow for more than 100 university students at the 1st MORPH summer school, co-located with SAUC-E (Student Autonomous Underwater Vehicle Challenge, Europe) 2012 in La Spezia, Italy, July 2012.
Image Mosaicking for Seabed Mapping, invited talk at the German Federal Maritime and Hydrographic Agency (Bundesamt für Seeschifffahrt und Hydrographie, BSH), Hamburg, Germany, February 2012.
A Spectral Image Registration Method and its Use in Underwater Maximum-Likelihood SLAM, invited lecture with Heiko Bülow in the symposium series at the Institute for Systems and Robotics (ISR), Instituto Superior Técnico (IST), Lisbon, Portugal, November 2011.
Data Analysis and Simulation, hands-on session at the 2nd Co3AUVs Summer School, Bremen, Germany, July 2011.
Multi-Robot Pose Graph SLAM: Theory and Practice, lecture at the 2nd Co3AUVs Summer School, Bremen, Germany, July 2011.
Introduction to SLAM: Particle Filters and the Pose Graph, lecture at the 2nd Co3AUVs Summer School, Bremen Germany, July 2011.
Simultaneous Localization and Mapping, lecture and hands-on session at the 1st Co3AUVs Summer School, Bremen, Germany, June 2010.
Virtual Robot Experiments in a High Fidelity Simulation, Talk at the Forum on “Mobile Robots” at Hannover Fair, Hall 22, co-located with RoboCup German Open, Hannover, Germany, April 2009.
Quality Measures for Mapping: From Test Environments to Analysis Tools, Talk at the Workshop on Good Experimental Methodology & Benchmarks in Cognitive Robotics at CogSys 2008, Karlsruhe, Germany, April 2008.
This combined lecture for Bachelor and Master students focused on the mathematical and methodological underpinnings of robotics, amongst others on spatial transformations and kinematics, sensors, actuators, intro to sensor data processing, and integration, especially with ROS. I contributed several practical sessions applying the theoretical concepts in small problems in ROS using simulation.
Software Engineering lab work for Bachelor students in their 4th semester. Emphasis on social and team-oriented skills related to programming. In groups, students implemented, documented, and tested a robotic system with ROS to compete in a maze task both in simulation and on real robots. I contributed introductory lectures, ROS and hardware support, and guidance for the students.
Advanced projects for Master students. Beyond the scope of a regular lab, students learned to do research in current topics in robotics in a scientifically principled manner and present it appropriately.
Individual Master student projects related to Prof. Birk’s lecture “Robotics”.
This combined lecture for Bachelor and Master students focused on the mathematical and methodological underpinnings of robotics, amongst others on spatial transformations and kinematics, sensors, registration, localization and mapping.
Individual student projects related to the lecture “Autonomous Systems”. I supervised several projects.
Advanced projects for Master students and integrated PhD students. Beyond the “Autonomous Systems Lab”, students were able to broaden the scope of their projects, practice the scientific method, and work more intensively towards first publications. I supervised several projects.
This seminar for Master students and advanced Bachelor students focused on current topics in robotics, especially regarding autonomous micro aerial vehicles, for use in the RoboCup Rescue competition. After the class, several students integrated and applied their work in the Jacobs Robotics team during the RoboCup German Open. I supervised most projects and the subsequent trip to RoboCup.
This project-based seminar for Master students focused on current topics in mobile robotics and medical image processing. I supervised a number of projects within this course. I supervised multiple projects.
Conference Best Paper Finalist at ICRA 2014 for the paper Representing and solving local and global ambiguities as multimodal and hyperedge constraints in a generalized graph SLAM framework, 5 finalists out of 1000 accepted submissions.
Scholar of the Studienstiftung des Deutschen Volkes, PhD Scholarship. Largest and most prestigious organization sponsoring the academically gifted in Germany.