Implicit Mobile Human-Robot Communication for Spatial Action Coordination with Action-Specific Semantic Environment Modelling

2026–2028
Project Duration
BMBF
Federal Funding
6
Publications
4
Citations

Home

Vision

EDISON envisions a new paradigm of industrial production in which humans and intelligent machines collaborate seamlessly through context-aware photonic intelligence. By integrating high-resolution optical sensing, explainable artificial intelligence (XAI), and adaptive robotics, the project aims to create flexible, intuitive, and responsive production environments.

Manufacturing systems will move beyond rigid programming toward dynamic adaptation driven by human demonstration and contextual understanding. Operators can guide and adjust processes in real time, while cognitive robotic assistants interpret, learn, and execute tasks safely and transparently.

This enables efficient and economically viable small-batch production and supports mass customization without compromising productivity or quality. EDISON aims to set a new standard for human-centered, trustworthy, and sustainable automation.

Intelligent assistance systems enhance human capabilities, reduce operational complexity, and strengthen resilience in rapidly changing industrial environments. By combining photonic intelligence with cognitive robotics, the project contributes to shaping the next generation of Industry 5.0 manufacturing.

The project is carried out by a strong consortium consisting of the Neuro-Information Technology group at Otto-von-Guericke-Universität Magdeburg alongside industrial partners: MartinMechanic GmbH, LUCAS Instruments GmbH, HEMA Formenbau + Kunststoffverarbeitung GmbH, and ZBS e.V., with Pilz GmbH & Co. KG as an associated partner.

Transforming Production with Photonic Intelligence and AI

EDISON’s mission is to fundamentally modernize manufacturing processes by integrating advanced photonic sensor technologies, artificial intelligence (AI), and adaptive robotic assistance systems. The project focuses particularly on small-batch production, enabling it to become more economical, flexible, and safe-meeting the growing demands for efficiency, customization, and innovation in modern industry 5.0.

At the heart of EDISON lies the seamless fusion of high-resolution optical 3D sensing, explainable AI, and intuitive human-robot interaction. Modern photonic sensors capture human actions and environmental conditions in real time. Intelligent algorithms interpret these contextual signals and translate them into precise robotic actions. This approach enhances existing production lines and enables rapid adaptation to new tasks and changing requirements.

Human-Centered, Adaptive, and Trustworthy

Aligned with the guiding principles of Industry 5.0-human centricity, resilience, and sustainability-EDISON develops context-sensitive assistance systems designed to work naturally and safely alongside human operators. Innovative methods such as one-shot learning and imitation learning allow robots to learn new tasks directly from human demonstrations, even with minimal training data. This creates the foundation for self-learning, adaptive production systems.

A key contribution of the Neuro-Information Technology group at Otto-von-Guericke-Universität Magdeburg is the development of a no-code robotics platform. This platform enables new workflows to be derived directly from human demonstrations and executed transparently. Photonic 3D scene data are combined with explainable AI models developed in compliance with the European Union’s AI regulatory framework, ensuring trustworthy and transparent human-robot collaboration.

From the outset, EDISON integrates certification, regulatory compliance, and functional safety into the development process, ensuring that research outcomes can be translated efficiently into real-world industrial applications.

Building the Next Generation of Cognitive Robotics

As part of the RoboSensation research cluster at Otto-von-Guericke-Universität Magdeburg, EDISON represents a significant step toward the next generation of cognitive robotics in industrial production-intelligent, flexible, safe, and explainable.

By combining cutting-edge photonic intelligence with adaptive robotics, EDISON contributes to strengthening sustainable and resilient automation “Made in Germany” and shaping the future of human-centered manufacturing.

News

image image
Annual Project Meeting in Ilmenau
The EDISON project partners met at ZBS e.V. in Ilmenau for their annual project meeting.
image image
Per-Dot Calibration Approach for Single-Shot 3D Sensors Using Structured Light
A per-dot disparity-to-depth calibration for single-shot structured light 3D sensors, combined with an improved chessboard target detection method.
image image
See, Spawn, Synchronize: A Digital Twin Pipeline for Smart Production Cells
A digital twin that synchronizes objects bidirectionally between a real and a virtual production cell using only a single RGB-D...
image image
REMA-Net: Real-Time Human Action Recognition via Caption-Guided Dual-Stream Attention
A dual-stream framework fusing RGB features with caption embeddings for real-time human action recognition on embedded robotic platforms.
image image
SAR-SLAM: Semantic-Aware Recognition for Dynamic SLAM in Robotic Applications
An RGB-D SLAM framework that combines semantic segmentation and geometric motion verification to stay robust in scenes with moving people...
image image
GESTRO: Skeleton-Based Dynamic Gesture Recognition for Robot Teleoperation
A new dataset of 11 dynamic gestures in 3,300 samples plus an attention-based LSTM model for real-time teleoperation of mobile...
image image
Deep Learning-Based Gaze Estimation: A Review
A systematic review of deep learning-based gaze estimation, from traditional methods to appearance-based techniques, and its role in human-robot interaction....
image image
Kick-Off Meeting in Magdeburg
All project partners of EDISON met in Magdeburg for the official project kick-off.
image image
EDISON Project Launch
EDISON officially launched on January 1, 2026, with 2.4 million EUR funding by BMFTR.