News

We present an integrated vision system for autonomous forklifts that combines RGB-only 6D pallet pose estimation with gesture- and face-based...

We present a digital twin of a smart production cell that synchronizes objects bidirectionally between a real and a virtual...

REMA-Net is a lightweight dual-stream framework for real-time human action recognition that pairs RGB visual features with semantic embeddings derived...

SAR-SLAM is an RGB-D SLAM framework that keeps tracking reliable when people and objects move through the scene. It combines...

This study investigates efficient speech-to-intent recognition for human–robot interaction in elderly-care environments in German, targeting deployment on resource-constrained platforms such...

We introduce GESTRO, a dataset of 11 dynamic gestures for the intuitive teleoperation of mobile robots, comprising 3,300 samples recorded...

A comprehensive review of gaze estimation methodologies, tracing the evolution from conventional model- and feature-based approaches to modern deep learning...

We introduce a conceptual temporal Shared Mental Model (SMM) that maps changing human-human and human-agent team dynamics using temporal graph...

A comprehensive review of synthetic dataset generation approaches for 3D object detection, covering both traditional 3D modeling and neural image...

An end-to-end pipeline that automates all aspects of 3D dataset generation by leveraging Radiance Fields for high-quality 3D modeling and...

We present a real-time digital twin of the TIAGo service robot built in Unity, featuring authentic kinematic mirroring, real-time sensor...

We conducted a user study in a realistic collaborative human-robot interaction scenario with unscripted tasks to evaluate how well current...

RoSA 3 extends the Robot System Assistant framework with a React-based touchscreen interface and a locally running speech pipeline that...

An efficient and lightweight network for gaze estimation that leverages progressive attention mechanisms to emphasize crucial eye features while maintaining...

We propose a novel HAR framework integrating residual networks, Bi-LSTM, and multi-head attention with a motion-based frame selection strategy. It...

We present IM HERE, a formal effort-based framework for modeling engagement across human–robot, human–human, and robot–robot interaction. It captures relational...

We present a fully automated pipeline that leverages Radiance Field–based universal 3D representations to rapidly generate high-quality models of arbitrary...

With the rapid development of large language models (LLMs), their integration into autonomous systems has become essential. This integration significantly...

We introduce a new approach to predict human engagement in human-robot interactions (HRI), focusing on eye contact and distance information....

We propose an enhanced SLAM framework that combines RGB-D and 2D LiDAR data using late fusion and adaptive resampling for...

We present our recent work on gaze estimation. Our approach is a novel two-branch CNN architecture with a multi-loss approach...