Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Smart Teaching - Master Thesis in collaboration with CSEM
Industrial grasping and pick-and-place highly benefit from robotic systems, allowing fast and precise execution of highly repetitive tasks. The common and widely established approach implies each single action of each single robot to be manually preprogrammed within a teaching phase, and simply repe
Industrial object manipulation highly benefits from robotic systems, allowing fast and precise execution of repetitive tasks. The common approach implies each action of each robot to be manually preprogrammed (teaching). While very effective when the conditions remain constant and predictable, each variation (desired and unexpected) requires the system to be manually adjusted. A new teaching phase requires trained engineers and is time consuming. Moreover, current teaching is limited to guiding the robot along reference trajectories, while the knowledge of the human operator goes way beyond simple kinematic information. Instead of hard-coding complex trajectories and policies, it would be much more effective teaching how to interact with the object, how to approach it in order to grasp it correctly, in which way the object should be moved from A to B, etc. An effective method commonly used to teach to another human is showing the new task with few demonstrations. **Can we teach to a robot in a similar way and can we improve the transfer of knowledge and experience from the human to the robotic system?** AI in form of deep reinforcement learning and other machine learning techniques combined with vision-based learning by demonstration offer today powerful tools, well suited to those fields where agility and adaptation are required. **Goal of this project is investigating, designing, and evaluating new AI-based methods to teach tasks (e.g. pick and place, move along trajectories, grasping, object manipulation) to a collaborative-robot (see figure) in an intuitive and agile way in order to improve the knowledge transfer from the human operator to the robotic system.** At CSEM Center Alpnach we have a strong background in Machine Learning and Machine vision and we develop innovative robotic and automation solutions. The Thesis will be conducted at CSEM (Center Alpnach) and co-supervised by Prof. R. Riener (SMS-Lab).
Industrial object manipulation highly benefits from robotic systems, allowing fast and precise execution of repetitive tasks. The common approach implies each action of each robot to be manually preprogrammed (teaching). While very effective when the conditions remain constant and predictable, each variation (desired and unexpected) requires the system to be manually adjusted. A new teaching phase requires trained engineers and is time consuming. Moreover, current teaching is limited to guiding the robot along reference trajectories, while the knowledge of the human operator goes way beyond simple kinematic information. Instead of hard-coding complex trajectories and policies, it would be much more effective teaching how to interact with the object, how to approach it in order to grasp it correctly, in which way the object should be moved from A to B, etc. An effective method commonly used to teach to another human is showing the new task with few demonstrations. **Can we teach to a robot in a similar way and can we improve the transfer of knowledge and experience from the human to the robotic system?** AI in form of deep reinforcement learning and other machine learning techniques combined with vision-based learning by demonstration offer today powerful tools, well suited to those fields where agility and adaptation are required. **Goal of this project is investigating, designing, and evaluating new AI-based methods to teach tasks (e.g. pick and place, move along trajectories, grasping, object manipulation) to a collaborative-robot (see figure) in an intuitive and agile way in order to improve the knowledge transfer from the human operator to the robotic system.** At CSEM Center Alpnach we have a strong background in Machine Learning and Machine vision and we develop innovative robotic and automation solutions. The Thesis will be conducted at CSEM (Center Alpnach) and co-supervised by Prof. R. Riener (SMS-Lab).
▪ State-of-the art and technology maturity (literature review)
▪ Design of method and algorithms
▪ Implementation of learning/control algorithms
▪ Evaluation of system performance (simulation and real world)
▪ State-of-the art and technology maturity (literature review) ▪ Design of method and algorithms ▪ Implementation of learning/control algorithms ▪ Evaluation of system performance (simulation and real world)
**About CSEM**
CSEM, founded in 1984, is a Swiss research and development center (public-private partnership) specializing in microtechnology, nanotechnology, microelectronics, system engineering, photovoltaics and communications technologies. Around 450 highly qualified specialists from various scientific and technical disciplines work for CSEM in Neuchâtel, Zurich, Muttenz, Alpnach, and Landquart. Further information is available at www.csem.ch.
**About CSEM**
CSEM, founded in 1984, is a Swiss research and development center (public-private partnership) specializing in microtechnology, nanotechnology, microelectronics, system engineering, photovoltaics and communications technologies. Around 450 highly qualified specialists from various scientific and technical disciplines work for CSEM in Neuchâtel, Zurich, Muttenz, Alpnach, and Landquart. Further information is available at www.csem.ch.
▪ MSc student in Mechanical Engineering or Robotics, Systems & Control
▪ Good programming skills (Python, c++, c#, MATLAB)
▪ Knowledge of and interest in machine learning
▪ Knowledge of Linux OS is preferred
▪ MSc student in Mechanical Engineering or Robotics, Systems & Control ▪ Good programming skills (Python, c++, c#, MATLAB) ▪ Knowledge of and interest in machine learning ▪ Knowledge of Linux OS is preferred
Dr. Francesco Crivelli (francesco.crivelli@csem.ch), CSEM SA
Prof. Robert Riener (riener@hest.ethz.ch), ETH, SMS-Lab
Dr. Francesco Crivelli (francesco.crivelli@csem.ch), CSEM SA
Prof. Robert Riener (riener@hest.ethz.ch), ETH, SMS-Lab