|
Yufei Zhu I am Yufei Zhu, a doctoral student at the Robot Navigation & Perception Lab, Örebro University in Sweden, under the supervision of Martin Magnusson, with co-supervision by Andrey Rudenko, Tomasz P. Kucner, and Achim J. Lilienthal. I am working on the EU horizon 2020 project DARKO and affiliated with Wallenberg AI, Autonomous Systems and Software Program (WASP). My research is part of DARKO EU project, which aims to develop agile production robots that are efficient, safe and able to operate in work environments shared with humans. A significant aspect of this environment is the need for robots to be aware of human presence and intentions, allowing for smooth and intuitive interactions. To achieve this, the focus is on learning probabilistic representations of human motion patterns, i.e. maps of dynamics (MoDs). By encoding spatial or spatio-temporal patterns of human motion as environmental features, the learned MoDs can be exploited to infer goal locations, constraints, and preferences implicitly, which is crucial for multimodal long-term human motion prediction. Efforts are also directed toward enhancing the efficiency and flexibility of learning these patterns, including online updates with new observations. Before starting my doctoral studies, I worked as a software engineer at Klarna AB and Ericsson AB in Stockholm, Sweden. Github / Google Scholar / LinkedIn / Email |
|
News |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Research |
|
Fast Online Learning of CLiFF-maps in Changing Environments
Yufei Zhu, Andrey Rudenko, Luigi Palmieri, Lukas Heuer, Achim J. Lilienthal, Martin Magnusson ICRA 2025 Present an online update method of the CLiFF-map (an advanced map of dynamics type that models motion patterns as velocity and orientation mixtures) to actively detect and adapt to human flow changes. |
|
Trajectory Prediction for Heterogeneous Agents: A Performance Analysis on Small and Imbalanced Datasets
Tiago Rodrigues de Almeida*, Yufei Zhu*, Andrey Rudenko, Tomasz P. Kucner, Johannes A. Stork, Martin Magnusson, Achim J Lilienthal IEEE Robotics and Automation Letters (RAL) Present a set of conditional pattern-based and efficient deep learning-based baselines, and evaluate their performance on robotics and outdoors datasets (THÖR-MAGNI and Stanford Drone Dataset). |
|
LaCE-LHMP: Airflow Modelling-Inspired Long-Term Human Motion Prediction By Enhancing Laminar Characteristics in Human Flow
Yufei Zhu, Han Fan, Andrey Rudenko, Martin Magnusson, Erik Schaffernicht, Achim J Lilienthal ICRA 2024 Present the Laminar Component Enhanced (LaCE) LHMP approach inspired by data-driven airflow modelling. LaCE-LHMP extracts laminar patterns in human dynamics and uses them for motion prediction, mitigating the impact of anomalous data in an unsupervised manner. |
|
CLiFF-LHMP: Using Spatial Dynamics Patterns for Long-Term Human Motion Prediction
Yufei Zhu, Andrey Rudenko, Tomasz P Kucner, Luigi Palmieri, Kai O Arras, Achim J Lilienthal, Martin Magnusson IROS 2023 Present CLiFF-LHMP, which uses CLiFF-map, a specific MoD trained with human motion data recorded in the same environment to generate multi-modal trajectory predictions over extended periods of time. |
|
A Data-Efficient Approach for Long-Term Human Motion Prediction Using Maps of Dynamics
Yufei Zhu, Andrey Rudenko, Tomasz P Kucner, Achim J Lilienthal, Martin Magnusson 5th LHMP workshop ICRA23 Explore the data-efficient characteristics of using Maps of Dynamics (MoD) for long-term human motion prediction. |
|
THÖR-MAGNI: A large-scale indoor motion capture recording of human movement and robot interaction
Tim Schreiter, Tiago Rodrigues de Almeida, Yufei Zhu, Eduardo Gutierrez Maestro, Lucas Morillo-Mendez, Andrey Rudenko, Luigi Palmieri, Tomasz P Kucner, Martin Magnusson, Achim J Lilienthal The International Journal of Robotics Research 2024 Present a new large dataset of indoor human and robot navigation and interaction, called THÖR-MAGNI, that is designed to facilitate research on social human navigation: for example, modeling and predicting human motion, analyzing goal-oriented interactions between humans and robots, and investigating visual attention in a social interaction context. |
Project |
|
Dynamic Agile Production Robots that Learn and Optimise Knowledge and Operations
DARKO EU Project Örebro University, TUM, Bosch, University of Pisa, EPFL, University of Lincoln, ACT Operations Research DARKO is innovating agile production robots for efficient and safe intralogistics in warehouses. |
|
Object Picking and Constrained Placement by Visual Reasoning
Yufei Zhu, Shih-Min Yang, Rishi Hazra, Kamran Hosseini, Karol Wojtulewicz We designed a robotic system for precise object manipulation, integrating a perception module, visual reasoning module, and an in-hand perception and control module. |
|
From Website Source |