Deep Robotic Learning

Bachelor thesis, Research assistant, Master thesis

Research focus: Automation and Robotics, Machine Learning, Service Robotics and Mobile Manipulation
Courses of study: Electrical Engineering, Informatics, Mechanical Engineering, Mechatronics, Related Disciplines

Job Description

The emergence of the new artificial intelligence (AI) technologies opens new opportunities and aspects to approach the problems and challenges in the field of robotics. One example is giving robots the ability to learn and explore instead of being manually programmed by humans. A promising machine learning approach, such as reinforcement learning (RL) addresses the problem of the automatic learning of optimal decisions over time, by constantly improving through trial and error, thus leading to higher autonomy and generalization. Combined with deep learning (DL), robots could additionally gain the ability to better handle unstructured environments and deal with their high-dimensional nature. This is known as deep reinforcement learning (DRL), which essentially characterizes as a general-purpose adaptive algorithm, that enables us to combine abstract and ambiguous sensory input such as camera image directly with robotic control and robot manipulation. We will explore and test different algorithms and learning technics that will enable robots to learn and generalize faster while relying strongly on transfer learning. I am looking for highly motivated students that are open to learn the latest AI concepts and apply them in robotics, both in simulation and in the real world.

Your Responsibilities

  • Paper research for the latest state of art in Deep Robotic Learning
  • Training and learning the robot in simulation, while testing different learning approaches
  • Transfer the trained agents in simulation into real-world use case with real robots

Our Offer

  • An interdisciplinary working environment with partners from science, industry and society
  • A working environment and organization close to business
  • Pleasant and motivating work atmosphere
  • Constructive teamwork with constant and mutual support

Your Profile

  • Basic knowledge of C++ or Python
  • Basic knowledge in machine learning, (optional OpenAI-Gym, TensorFlow, PyTorch)
  • Self-reliant thinking and working
  • Fluent in German or English
  • Motivation and commitment

Application

We are looking forward to receiving your PDF application. Please send your application containing the following documents to Mr Atanas Tanev, tanev@fzi.de:

  • Current transcript of records
  • Curriculum vitae in tabular form

Job Description

  • Start: as soon as possible
  • Supervising institute at KIT: IAR - Institut für Anthropomatik und Robotik | Prof., Rüdiger, Dillmann