Pint of Robotics: From Tracking to Planning: Context Understanding for Autonomous Systems, Jan-Nico Zaech, ETH Zürich

Speaker:  Jan-Nico Zaech, ETH Zürich

Title: From Tracking to Planning: Context Understanding for Autonomous Systems

When: Wednesday 19th May, 1700-1800

Where: Zoom (free but registration required)

Abstract: Autonomous systems are currently moving from isolated and controlled settings into more open environments that include interactions with other robots as well as humans. While this comes with a range of opportunities, it also poses important research challenges. The new environments change dynamically and require robust methods to detect and track other agents. Together with an understanding of intention, smooth and safe interactions in the shared space need to be ensured. Finally, the way a robot can solve a task is influenced by the environment and should be adapted while the robot proceeds towards achieving its goal. The talk covers three projects on different aspects of this pipeline, online tracking of agents in 3D, action prediction for road-users, and trajectory planning to solve an imaging task.

Short Bio: Jan-Nico Zaech is a Ph.D. candidate at the Computer Vision Lab at ETH Zürich (Switzerland). His research mainly focuses on computer vision for autonomous systems and vehicles. He received his M.Sc. degree at the Friedrich-Alexander-Universität Erlangen-Nürnberg (Germany). As visiting scholar at the Johns Hopkins University (USA) he worked on intelligent X-ray imaging systems and contributed to the Square Kilometre Array signal-processing pipeline at Callaghan Innovation (New Zealand).