Yuezhe Zhang

Hey! This is Yuezhe Zhang 张越哲. I am a PhD candidate at PEARL Lab, TU Darmstadt, under the supervision of Prof. Georgia Chalvatzaki.

I obtained my matster degree with honours in Robotics from TU Delft. I worked on my thesis at Autonomous Multi-Robots Lab and AIRLab Delft, under the supervision of Dr. Javier Alonso-Mora. Previously, I obtained my bachelor degree with honours in Mechatronics Engineering from Xuteli School, Beijing Institute of Technology.

My research interests lie in the development of new integrated task and motion planning algorithms that combine fundamental concepts from optimization techniques with machine learning techniques, aiming at obtaining human-level adaptability and intelligence in real robot systems.

news

Mar 18, 2024	I am joining PEARL Lab, TU Darmstadt as a PhD candidate! See here.
Aug 30, 2023	I graduated from TU Delft with honours!

selected publications

Multi-Modal MPPI and Active Inference for Reactive Task and Motion Planning

Y. Zhang, C. Pezzato, E. Trevisan, and 3 more authors

arXiv preprint, Dec 2023

Abs HTML Website

Task and Motion Planning (TAMP) has made strides in complex manipulation tasks, yet the execution robustness of the planned solutions remains overlooked. In this work, we propose a method for reactive TAMP to cope with runtime uncertainties and disturbances. Particularly, we combine an Active Inference planner (AIP) for adaptive high-level action selection, and a novel Multi-Modal Model Predictive Path Integral controller (M3P2I) for low-level control. The AIP generates alternative symbolic plans, each linked to a cost function and samples for M3P2I. M3P2I employs a physics simulator for diverse trajectory rollouts, deriving optimal control by weighing the different samples according to their cost. This idea enables blending different robot skills for fluid and reactive plan execution, accommodating plan adjustments at both the high and low levels to cope for instance with dynamic obstacles or disturbances that invalidate the current plan. We have tested our approach both in simulations and real-world scenarios.