Results
Discover the outcomes of the HARTU project
HARTU sets out to deploy new technologies in different relevant industrial use-cases. This document has been compiled based on input from each of the consortium members and describes the scenarios of 5 industrial partners.
In line with the user-centred and industrial driven approach followed by the HARTU project, the consortium has implemented an initial set of demonstrators as part of this strategy. A total of 11 prototypes are available, corresponding to the 8 use cases defined by the project. The document includes also the methodology that is used for the User Research, as well as the initial insights on this topic which are the result of the literature review and the various workshops that have been held in the first 10 months of the project with different stakeholders from the 5 Industrial companies offering the validation scenarios. Finally, Legal and Ethical aspects that have to be considered are presented.
This document describes the development of the Human-AI Teaming Interaction Model (HATIM). HATIM is a descriptive model of Human-AI Interaction, developed to support the creation of teams paring Human and Autonomy (AI agents able to operate autonomously) in manufacturing. The model is composed of a selection of variables that, according to the literature, can describe the Human Operator, the Autonomous Agent and the resulting Team. These have been selected and organised in macro blocs of Operator, AI-Agent, Mediator, Environment and Task, to describe the connection between them and how they affect teaming. A particular innovation of the model is a novel approach to the definition and description of the human and AI roles within the team, based on concepts from narrative and storytelling.
The document includes the theoretical basis of the model in the form of a literature review, a description of the model and its components, a guide on the application of the model as part of a User Centred Design effort, and a trial application of the model in the context of the HARTU project.
This document presents the work carried out to define the new skills required of operators in the TO-BE scenarios envisioned within the Industry 5.0 context, such as the use cases addressed in the HARTU project.
The first part introduces the Skill–Task Mapping tool developed to support data collection and subsequent analysis. Based on a literature review and insights gathered during workshops with project partners, a structured list of skill —organised into eight categorie —has been defined.
The second part presents the Skills Transformation Map, which describes the three new human profiles (i.e. Smart Line Operator, Plant Flow Keeper, Tech Solver) expected to operate in future Industry 5.0 scenarios, together with their associated tasks and skills. The Skills Transformation Map has been implemented as an interactive tool equipped with section-specific charts, enabling users to navigate and explore the data according to their interests (e.g., identifying the most relevant skills across operator profiles or analysing the tasks carried out by a specific human role).
Finally, the document introduces the Human-Centric Design and Training Guidance. This set of guidelines and sub-guidelines, organised into five sections (i.e. Procedures, workspace and working practices, Human Performance Transition Factors, Teams & Communication, Skills, roles and responsibilities, Ethics and Liability) and divided into design and training guidelines, is intended to support design teams in the transition toward and development of smart manufacturing work environments. The Guidance ensures that human factors and user experience considerations are integrated throughout the design process and that appropriate training programmes can be planned to prepare operators whose roles will evolve and require new competencies. Together, the Skills Transformation Map and the Human-Centric Design andTraining Guidance provide direction on how to develop acceptable and trustworthy solutions for smart manufacturing, as well as how to customise training programmes to ensure proper operator preparation.
A new release of the real-world scenarios based on the analysis and requirements reported in D1.1 has been deployed. In this release some changes in the setups have been included and all technical results have been integrated and are ready to be validated in the last six months of the project.
This document, Deliverable D1.6 of the HARTU project, presents the final validation results of the AI-enhanced robotic technologies developed throughout the project’s duration. The primary objective has been to evaluate these solutions across five real world industrial scenarios encompassing seven specific use cases in the assembly and logistics sectors, transitioning from research-driven prototypes to validated industrial applications. To ensure a continuous improvement cycle and precise alignment with industrial requirements, the validation methodology was structured through a series of iterative Test Sprints. This agile approach allowed the consortium to progressively test, refine, and validate system capabilities, ensuring that every technological component was evaluated under actual operating conditions before final integration.
The report is structured to provide a comprehensive analysis of each scenario and its corresponding use cases. For every case, the document details the final implementation state, the quantitative and qualitative results achieved, and the specific conclusions derived from the validation. This comprehensive breakdown offers a clear view of both the individual performance of the HARTU tools and their collective effectiveness in diverse and flexible manufacturing environments. Overall validation results have been highly positive, particularly regarding the performance of AI algorithms in complex picking and precision assembly tasks, where the system demonstrated significant reliability. Furthermore, the validation successfully confirmed the effectiveness of the electro-adhesion (EA) gripper for versatile material handling, alongside the Fiber Bragg Grating (FBG) sensing technology, which provided the high-precision data essential for advanced process control. These achievements mark a significant milestone in creating an adaptable robotic platform capable of managing the high variability inherent in modern industry.
Despite these successes, the validation process also identified critical lines for future improvement to ensure full commercial competitiveness. The most prominent technical challenge concerns cycle times, which require further optimization to meet the productivity demands of high cadence industrial environments. Across the scenarios, findings indicate that reducing data processing latencies and further refining coordinated motion control are essential steps to eliminate operational bottlenecks. By addressing these areas for improvement and enhancing overall temporal efficiency, the HARTU system will be positioned as a fully competitive solution ready to transform automation across the European industrial landscape.
The current deliverable aims to report the main results of tasks T2.1 and provides relevant information about Reference Architecture or high-level specification of the HARTU software solutions that will be provided by technical partners (TEK, AIMEN, ITRI and DFKI) to address pilot needs and requirements. Tasks T2.1 covers the overall project approach on Reference Architecture model, reporting the results on technical drawings and software components specifications identifying key technical challenges to be addressed in HARTU, in terms of implementation and architecture definition.
HARTU has developed a set of simulation tools (SimEnv) to help system integrators configure some features of a handling application. This document describes these tools.
This document describes the main concepts of HARTU-APP-MANAGER, the tool that HARTU makes available to end users and system integrators to create and control robotic applications, thus giving rise to two different functionalities: HARTU-APP-CREATOR and HARTU-APP-CONTROL. The core of HARTU-APP-MANAGER is based on the concept of behaviors trees and its implementation using BehaviorTree.CPP 3.8, a C++ library for building BehaviourTrees. The GUI is inspired and based GROOT, an advanced open IDE for creating and debugging BehaviourTrees. HARTU-APP-MANAGER is an open tool that will be made accessible through the GITLAB infrastructure created by HARTU.
The overall topic of the HARTU project is to develop robotic technologies which allow a more flexible robotic automation of manufacturing and logistics tasks. One element towards achieving this goal is the application of data-driven machine learning approaches. In the course of the method development in the project, multiple datasets related to robotic assembly have been created for this purpose. This document describes a dataset of user demonstrations for learning and generalizing robotic assembly skills, a dataset of measurements of a novel type of force sensor using fiber optics, as well as image datasets of the objects being handled in various containers.
This document presents initial the dissemination, exploitation, and communication plan developed to promote the project and raising awareness and visibility about its progress and achievements. This plan outlines the dissemination goals, the methodology behind it, the overall dissemination approach, and the channels used, and identifies the dissemination actions planned for the project duration. It is designed as a practical framework for day-to-day communications activities, which will range from the communication through the project’s social channels and its website, to graphic activities, such as posters and brochures to be proposed at events where HARTU will participate. It includes a section devoted to the exploitation of HARTU’s results, which includes a series of strategic activities and a structured methodology that will help to identify a detailed exploitation plan encompassing the project and the partners’ needs.
This deliverable is an update of the D6.1 Dissemination, Communication and Exploitation plan submitted in June 2023. It provides an overview of the progress and effectiveness of HARTU’s communication strategy at the 18-month mark. It encompasses a comprehensive review of past
events, communication actions implemented, graphical products developed, and an evaluation of the communication success against the Key Performance Indicators (KPI). Additionally, the update identifies deviations and areas for improvement observed as the project has advanced.
HARTU will adhere to FAIR principles by: Making data findable, Making data accessible, Data re-usability. HARTU project will use and generate several datasets. HARTU will make them available to stakeholders unless some restrictions apply for commercial or privacy reasons.
Scientific publications
Read our researchers’ studies
In European Robotic Forum 2024, by TEK, Italy. (15th ERF, Volume 1). Here.
In European Robotic Forum 2024, by AIMEN, Italy. Here
In 2024 IEEE International Conference on Robotics and Automation, by DFKI, May 2024, Japan. Read it here.
In International Conference ADM2024, by DBL, 2024. Here.
In SII2025 by DFKI, January 2025, Germany. Read it here.
Simulation-Based Synthetic Data Generation for Automated Training of Bin-Picking Segmentation Models
In ICARA 2025, by TEK, February 2025, Croatia. Read it here.
In Machine 2025, by TEK. Read it here.
In European Robotic Forum 2024, by AIMEN, Germany. Here
In Proceedings Volume 13639, 29th International Conference on Optical Fiber Sensors (2025), by AIMEN, Portugal. Here.
In Journal of the Mechanics and Physics of Solids, 2025, by POLIBA. Here.
In Tribology International, 2025, by POLIBA. Here.
In Advanced Materials, 2025, by OMNI. Here.
In Extreme Mechanics Letters, 2023, by OMNI-POLIBA. Here.
In Journal of Open Source Software, 2025, by DFKI. Here.
In Intelligent and Fuzzy Systems (INFUS 2025), 2025, by DFKI. Here.
In Frontiers in Robotics and AI, 2025, by DFKI. Here. (The final, formatted version of the article will be published soon)
Promotional materials
Help HARTU spread its work
Read the 1st HARTU press:
Videos
HARTU’s third webinar
Translating research into practice: turning real industrial needs into smart, flexible and human-centered solutions. This third webinar brought together experts from HARTU, SMARTHANDLE, AGILEHAND, and MASTERLY projects to explore how research and innovation are being translated into real-world manufacturing solutions, showcasing concrete use cases.
HARTU’s second project video
Discover how operators, engineers, and robots collaborate within HARTU’s vision of human-centred automation, embracing new skills, evolving roles, and seamless teamwork between humans and intelligent systems in the factories of the future.
HARTU’s second webinar
Shaping ethical and inclusive futures: the human side of innovation in robotics. This second webinar brought together experts from HARTU, SMARTHANDLE, AGILEHAND, and MASTERLY projects to explore the future of Human-AI Teaming and Human-robot interaction in manufacturing. The webinar focuses on the critical role of humans in increasingly automated industrial processes.
HARTU’s 1st project video
Welcome to the HARTU project! In this video, we present an overview of our research, its objectives, and the five industrial case studies we’re focusing on.
HARTU’s first webinar
How to pick (almost) anything: soft grippers revolution. The webinar is focused on the innovative gripper technologies that HARTU intends to implement in its case studies.
