JARVIS project
QUARK: Bringing Intelligence and Flexibility to Quality Inspection
The QUARK (QUality Assurance with Robotic Knowledge) project marks a significant step forward in the evolution of smart manufacturing.
Funded under the JARVIS Open Call, QUARK aims to transform traditional quality inspection into a flexible, intelligent, and human-centered process — combining artificial intelligence, collaborative robotics, and augmented reality (AR) to enhance both precision and usability.
The challenge
In many manufacturing environments, particularly in the plastics industry, quality control is still rigid and time-consuming.
Each time a new product or defect type is introduced, engineers must manually reprogram inspection routines and recalibrate vision systems — a process that can take hours or even days.
This lack of flexibility leads to downtime, increased costs, and reduced responsiveness to market demands.
The QUARK approach
QUARK addresses these challenges by developing a robotic inspection system powered by AI and guided by human interaction.
The system is built upon an existing UR10 collaborative robot and multi-camera vision station already operating at GERMANPLAST, which serves as the real industrial testbed for validation.
The project introduces two key JARVIS tools:
-
AR Interfaces, enabling operators to visualize inspection feedback and configure tasks intuitively without programming skills.
-
Object Recognition & Pose Estimation, allowing the robot to detect and adapt to different product geometries in real time.
On top of these, the system integrates AI models enhanced with synthetic data, improving defect detection accuracy and robustness — even for rare or complex defects.
Human-centered by design
What makes QUARK unique is its human-centric philosophy.
Rather than replacing operators, the system empowers them.
Through AR-based interfaces, operators can teach, guide, and refine the robot’s behavior, while maintaining full oversight of the inspection process.
This approach not only boosts efficiency but also increases trust, engagement, and job satisfaction.
Expected impact
By the end of the pilot, QUARK aims to:
-
Reduce setup and reconfiguration times by at least 30%.
-
Increase defect detection accuracy by 20%.
-
Achieve operator acceptance above 60%.
-
Reach TRL 7 validation in a real industrial environment.
These results will demonstrate how advanced human–robot collaboration can deliver tangible benefits to SMEs seeking flexible, high-quality production systems.
The team behind QUARK
The project is implemented by:
-
IDEA Soc. Coop. – technology provider and system integrator, responsible for AI development, software integration, and project coordination.
-
GERMANPLAST S.r.l. – industrial end-user, providing the production environment, process knowledge, and operator feedback for usability validation.
Together, they embody the synergy between innovation and industry that the JARVIS programme is designed to foster.
Looking ahead
Over the coming months, the QUARK team will focus on integrating JARVIS tools, expanding the AI dataset with synthetic images, and deploying the automated palletization module.
The final goal: a fully validated, AI-driven quality inspection system ready for real-world manufacturing.
Stay tuned for updates as we continue to advance toward a more adaptive, intelligent, and human-centered manufacturing future.
This pilot is developed as part of @JARVIS Project, funded under the Horizon Europe programme. JARVIS provides mentoring and tools complementing our Human-Robot Collaborative Solution, and accelerates our path towards human-centric, sustainable, and future-proof industrial innovation.
🔗 Learn more: https://jarvis-project.eu/
hashtagJARVISproject hashtagHorizonEurope hashtagHRI hashtagAI hashtagRobotics hashtagInnovation hashtagDigitalTransformation hashtagHumanCentricAI
