Vai al contenuto

Adaptive robotic manipulation through body perception and artificial intelligence

Development of new reconfigurable surfaces equipped with flexible sensors and controlled by AI-based learning tools for handling fragile and delicate objects

Partners involved

MOZART is a Consortium made up of IT Universitetet i København (computer science research coordinator), Italian Institute of Technology (research on poroelastic materials and sensors), Trinity College Dublin (SSH research), Cabinplant AS (food processing and packaging machinery supplier), UNI – Italian Standardisation Body (standardisation), ICONS Foundation (communication and dissemination), AVEC (B2B and B2I), École Polytechnique Fédérale de Lausanne (research on poroelastic materials, sensors and origami robotics), Brno University of Technology (user interfaces), University of Southern Denmark (SSH and human-robot interaction research).

Benefits

The European project MOZART develops new reconfigurable surfaces equipped with flexible sensors and controlled by learning tools based onartificial intelligence for handling fragile and delicate objects

The Challenge

Food packaging automation has evolved steadily over recent decades.
However, the handling of fragile and fresh food still largely depends on complex manual processes.
Manual work puts workers' health at risk and multiplies the risk of injury, as well as increasing problems of quality control of products, food waste and hygiene.
The European project MOZART was created to study the latest robotic technologies and system solutions to push the boundaries of the food industry towards economic, social and environmental sustainability.

The Solution

 

The European project MOZART, funded by the European Union's Horizon Europe research and innovation programme under grant agreement No 101069536, aims to advance the state of the art in robotic manipulation of soft and heterogeneous objects, currently limited to simple pick-and-place operations.

MOZART takes the first step towards solving this problem by developing a completely new approach based on innovative reconfigurable surfaces equipped with flexible sensors and controlled by AI-based learning tools, called AUTOMATs (Autonomous Manipulation using Morphing Modular Mats).

The Project was created to achieve the following objectives:

  • support the green transition of the food industry by developing soft tactile robotics for handling fresh and fragile food products;
  • extend the state of the art in robotic perception, interaction and manipulation, and address the modularity and composability of solutions;
  • constantly monitor, assess and ensure the sustainability (ecological, economic, social) of the project work;
  • prepare for the development of the AUTOMAT product using feedback from demonstrations and working towards standardisation;
  • disseminate, communicate and efficiently exploit the project results.

 

Resources: available use cases and documented trials

 

CWA 18409:2026

17-08-2026

CWA 18409:2026 “Grasp-free robotic surface manipulation – Vocabulary, characteristics, performance criteria and related test methods” is the result of the work of the MOZART consortium, led by project partner UNI – Italian Standardisation Body. The document serves as a shared reference for manufacturers, integrators and food processors, ensuring greater interoperability and adoption of the AUTOMAT system.

Download ↓

Exploring collaborative in-situ visual programming of industrial robotic workstations in augmented reality on handheld devices.

15-05-2026

In this paper by Brno University of Technology, user interfaces for collaborative robot programming are explored through a handheld augmented reality system that allows physically co-located users to jointly create and edit robot programs. The study highlights the challenges of coordination, workspace awareness and interaction in shared industrial robotic environments.

Download ↓

Surface manipulation using tunable and flexible porous-elastic soft sensors

29-04-2026

A poster by IIT presents a 3D-printed "smart skin" for robotic manipulators, able to sense and guide objects by adjusting its stiffness. It adapts to different tasks and materials, enabling reactive, physics-based manipulation. The poster was selected among the three finalists of Robosoft 2026.

Download ↓

Shape control – The electronic layer

20-02-2026

This info pack focuses on how the MOZART system is powered and coordinated. The electronic layer is the foundation that distributes power, manages communication and orchestrates the machine's sensors and actuators.

Download ↓

Shaping movement – The Origami actuator layer

20-02-2026

This info pack focuses on how actuation is achieved within the MOZART system. The Origami Robotic Actuator (ORA) is the layer in which mechanical modules reconfigure the surface to move the objects placed on it, enabling precise and adaptive manipulation.

Download ↓

Shaping perception: the Smart Skin layer

20-02-2026

This info pack focuses on how the MOZART system interprets its surroundings. The Compliant Poro-Elastic Smart Skin (COPESS ) is the layer between an object and the machine handling it, acquiring detailed tactile information that enables adaptive manipulation.

Download ↓

Discover the MOZART project

Ecosystem partners involved

Interested in adopting this methodology?