Michael Görner

Robots require a range of capabilities for goal-directed interaction with changing environments. These include safe and predictable motion models, perceptual scene understanding, the ability to infer scene affordances, and models of the outcomes of potential actions.

My research contributes to two complementary perspectives on autonomous manipulation:

1. Task-Level Planning for Long-Horizon Manipulation
Specifying and executing manipulation behaviors in changing environments remains a core challenge. I investigate computational structures that enable robots to plan and execute sequences of actions to achieve long-horizon goals in the physical world.

2. Data-Efficient Skill Learning on Real Hardware
Most low-level manipulation capabilities are either hand-designed or trained extensively in simulation, followed by a fragile sim-to-real transfer to the real system. As an alternative, I explore methods for data-efficient exploration and policy learning directly on physical robotic systems. This approach bypasses the need for exact system identification and implicitly compensates for real hardware behavior at the cost of relying on expensive physical rollouts.

By combining structured planning with data-driven learning, my work aims to enable autonomous systems that are both generalizable and grounded in real-world performance.

• University of Osnabrück
  M.Sc. with distinction, Cognitive Science / Robotics
  2012–2015

• University of Potsdam
  B.Sc., Informatics / Knowledge Processing
  2008–2012

• University of Hamburg / Research Associate
  Interdisciplinary Sino-German Collaboration Project
  2016-2025

• Independent Contractor / Robotics Integration
  2018-2019

• Tsinghua University, Beijing
  Research Project Secondee Marie Skłodowska-Curie Fellowship
  2018

• Student Researcher & TA at University of Potsdam and University of Osnabrück

2024

Sensor-agnostic Visuo-Tactile Robot Calibration Exploiting Assembly-Precision Model Geometries

Pluck and Play: Self-supervised Exploration of Chordophones for Robotic Playing

2023

Multi-Stage Book Perception and Bimanual Manipulation for Rearranging Book Shelves

A Multimodal Robotic Blackjack Dealer: Design, Implementation, and Reliability Analysis

2022

Coordinating human-robot collaboration by EEG-based human intention prediction and vigilance control

Efficient and Collision-Free Human-Robot Collaboration Based on Intention and Trajectory Prediction

Reinforcement Learning with Vision-Proprioception Model for Robot Planar Pushing

2021

Mobile Manipulation Hackathon: Moving Into Real-World Applications

2020

Detection and Reconstruction of Transparent Objects with Infrared Projection-based RGB-D Cameras

Crossmodal Pattern Discrimination in Humans and Robots: A Visuo-Tactile Case Study

Self-Adapting Recurrent Models for Object Pushing from Learning in Simulation

Learning Local Planners for Human-aware Navigation in Indoor Environments

2019

MoveIt! Task Constructor for Task-Level Motion Planning

PointNetGPD: Detecting Grasp Configurations from Point Sets

Vision-based Teleoperation of Shadow Dexterous Hand using End-to-End Deep Neural Network

2018

Comparison of Multimodal Heading and Pointing Gestures for Co-Located Mixed Reality Human-Robot Interaction

Simulation of the SynTouch BioTac Sensor