Thesis Project

Fatemeh Mohammadi Amin

"My research interest is focused on Computer Vision, and Deep Learning applied in Human-Robot Interaction. I am currently working on Safe Human-Robot Collaboration in a real industry application, and I aim to have a step change in human safety during collaboration."

Homepage

Thesis: Real-time dynamic modelling of Human-Robot Collaboration environments and human-aware motion planning

Supervisors:
Prof. Dr. Davide Scaramuzza (UZH)
Prof. Dr. Hans Wernher van de Venn (ZHAW)

Keywords: Industry 4.0, Safe Human-Robot Collaboration, safe real-time motion planning, collision avoidance

Abstract:Today, modern manufacturing systems require increased levels of automation for fast and low-cost production, but at the same time high levels of flexibility and adaptability to dynamic production environments. In order to reach this goal, the Industry 4.0 paradigm aims to establish a major step-change in conventional production methods by enabling Human-Robot Collaboration in a shared working environment with volitional physical contact. Collaborative robots (Cobots) is a major trend in industry 4.0 where the operator directly accesses and interacts with the robot. Safety during interaction in unstructured and dynamic environments is a requirement for complex robotic systems. In my thesis, safe human-robot interaction and safe real-time motion planning for Cobots in dynamic industrial environment will be considered. Ensuring human safety as the main part of this project can be solved by human activity recognition. Human activity can be detected by considering the change in the spatial configuration of a human skeleton within an industrial working environment. Recently we are working on environment monitoring and, in consequence, human action recognition. This thesis's main goal is to design a reliable framework for real-time safe human-robot collaboration, using a novel 3D human skeleton extraction that is independent of the environment. Safe real-time motion planning for robot manipulation will be the final part of this thesis.

Planned Duration: Spring 2020 - Fall 2024