July 19, 2023 - AI-PRISM

The development and implementation of robotic processes requires experts. Untrained users need a more natural way of coordinating human and robotic tasks to speed up and have the precision for task execution. One of the challenges on discrete manufacturing control is that the robotic system is variable.

Flexibility and Speeding Up processes

We will place modules to work with minimal training. This will include natural ways to teach a new process, to adapt existing ones, and “debug” their execution by non-programmers.

Expected results:

Decrease setup time by 12%
Decrease time for training by 20%
Decrease mental load operation and training
Decrease number of training cycles
Increase the number of communication modes
Improve usability of human-robot collaboration through multi modes communication

Many challenges have already been solved by industrial robots in this field, but there is an opportunity for cobots. Processes performed at Silverline final control workstation consist of 12 steps performed manually. Some of them are causing mental and health issues in workers.

Intelligent and Adaptable Control Station

The current tasks all performed by humans will be conducted either by robot, human or both.

Expected results:

Increase resource efficiency by 15%
Reduction of total production time by 20%
Reduction of production costs by 10%
Reduction of cognitive load required from the operator by 10%
Reduction of unwanted personnel movement by 5%

Reduction by 30% of workers in heavy task and high-risks areas
Reduction by 50% of OPEX dos custom/dynamic tasks
Reduction by 30% of time to execute custom/dynamic tasks

Painting and sanding processes are performed manually by skilled operators. Some of the pieces of furniture require very elaborate painting which is difficult to automate.

Collaborative environment
Cobots will learn different tasks by human demonstration such as picking, loading, unloading, painting, sanding.
Humans can focus on tasks that provide a higher added value for the company.

Expected results:

Increase by 20% painting process productivity
Decrease by 50% quality defects

Brewing production line is highly automated but some tasks are difficult due to their complexity. Humans end up performing monotonous and strenuous tasks, repetitive and heavy duties, which expose them to repetitive strain injury or musculoskeletal disorders risks.

Learning from human demonstrations
The tasks involving cobots will include extra labeling, operating clarks/forklifts, returned bottles classification, packing and palletizing custom orders, heavy tasks in the filtration preparation process and garbage collection and disclosure in geofenced areas based on safety protocols.

Humans will first perform the task once or several times to teach the robot and the next steps will be automatic.
Cobots will interact with the surrounding space by sensing the environment to ensure human safety, while being seamless connected with the other systems of the factory. They will execute coordinated and sequenced dynamically defined logistics tasks and customized orders.

Expected results:

Reduction by 30% of workers in heavy task and high-risks areas
Reduction by 50% of OPEX dos custom/dynamic tasks
Reduction by 30% of time to execute custom/dynamic tasks

At VIGO facilities approximately 5000 per year devices are being produced. The delicate nature of the components and variable production volume makes the automatization of the production with classic solutions very difficult and uneconomical.

AI Algorithms
Adapting AI Algorithms to the differences in components. The operator will teach the robot the correct positioning of the elements to be glued together.

Expected results:

Reduction of time of the operation from 5 to 1 minute
Less waste from 10% to 5%

AI Tools to enhance perception, reasoning and control capabilities in cobots

The industry requests more advanced and flexible solutions for using cobots in real production scenarios.

Building a distributed and calibrated network of sensors.

Novel models for the dynamic pose estimation of multiple people with respect to robot statuses.

Simplified deep graph partitioning method for detecting space-time patterns.

Techniques to classify and predict the future intention of operators based on accumulated behavioural data and the evolving context of processes and the environment.

Results:

Enhancing the global perception and reducing uncertainties in person tracking and modelling.

Positioning of AI-PRISM

The demand for cobots is constantly growing in industrial environments due to their versatility and low prices. However, implementing scenarios where robots work and learn from human operators autonomously while synchronizing their operations in a safe industrial environment with personnel is still in the conceptual phase.

A novel real-time platform for human-robot interaction, providing different skills for a smart manufacturing context.

The platform will be based on ROS (Robot Operating System) – highly recognized open-source framework for robotic software development that provides stable packages and tools as well as tutorials for developing different robotic applications.

Results:

Usage of ROS will decrease the risk of miscommunication between software components, increasing reliability and maturity.

Implemented ROS solutions will be a backbone of AI-enhanced modules.

Robot programming without highly skilled personnel – programming by demonstration

Nowadays, one of the higher trends is to incorporate methodologies to learn and perform every aspect of the task at hand by exploiting a massive number of observations.

Learning from Demonstration (LfD) aims at the generalization of a motion model based on movement cobots learned from observation of humans performing a task.

Probabilistic algorithms to generalize robot motions that can react online, enabling the generation and online adaptation of new policies in non-demonstrated situations.

Encoding a 6D rigid body trajectory information with a representation invariant to variation in contextual information such as reference frame, scale, and velocity.

Results:

Less intrusive interfaces to enable expert operators to perform demonstrations without disturbing their craftsmanship.

Craftsmanship can be further refined by exploiting the robot capabilities leveraging supervised learning techniques.

Increase the current manipulation capabilities, flexibility, and reliability to automate industrial tasks.

Our methodologies can generalize trajectories to non-demonstrated situation using information from a single demonstration.

Social sciences and humanities applied to collaborative robotics

As robotics, automation and digitisation become increasingly more pervasive in society. Most published studies are from an engineering approach and still do not include the social sciences or humanities.

Integrating social sciences and humanities throughout, from the first activities to identify requirements for the design of new systems to the final stage of evaluation and validation of those systems.

Involving an elaborate set of reliable novel and traditional human analysis techniques.

Physiological and sensor monitoring measures triangulated with rich descriptions and feedback from real industrial workers.

Results:

We will fully explore workers needs and preferences in relation to different collaborative systems and sociocultural settings.

Teknopar

Turkiye | Private Company

TEKNOPAR provides customers with a wide range of services and end-to-end solutions where integrated applications of OT and IT have been supplied.

Ikerlan S. Coop

Spain | Technology Centre

It is a leading centre in technology transfer and in the contribution of competitive value to industry. It offers integral solutions combining different technological domains in two main areas: Electronics, Information and Communication Technologies (EICT), and Energy and Mechatronics

NTTD Everis Italia Spa

Italy | Private Company – Affiliated Entity

In Italy, NTT Group has 5,000 professionals committed to building a unique and open community of people, guided by shared values. They intelligently anticipate the future in a unique, creative and generous collaborative environment.

NTTD Everis Spain SL

Spain | Private Company

NTT DATA – a part of NTT Group – is a trusted global innovator of IT and business services. They help clients transform through consulting, industry solutions, business process services, IT modernization and managed services.

The Spanish Association for Standardization, UNE

Spain | Non-profit Association

The Spanish Association for Standardization, UNE, is legally designated as the National Standardization Body of Spain. Through the development of technical standards, they help improve the quality, competitiveness and innovation of companies, products and services.

A&G Technology Co

South of Korea | Private Company

As an Expert engineering company, A&G promise to provide the best support/service based on the needs of customers.

Electronics and Telecommunications Research Institute, ETRI

South of Korea | Research Institute

Since its foundation in 1976, ETRI, a global IT research institute, has been making an immense effort to provide Korea with remarkable growth in the field of IT industry. It delivers Korea as one of the top IT nations in the World, by continuously developing the world’s first and best technologies.

Profactor GMBH

Austria | Research Institution

PROFACTOR researches for the competitiveness of the European Industry. Their technological developments strive for efficiency and sustainability, and work to benefit society. They are pioneers of thinking production.

Cranfield University

United Kingdom | Education Institution

Cranfield’s world-class expertise, large-scale facilities and unrivalled industry partnerships are creating leaders in technology and management globally. Their expertise is in a deep understanding of technology and management and how these work together to benefit the world.

Universitat Politécnica de Valencia

Spain | Education Institution

It is one of the 5 Spanish universities included in Reuter´s European Elite, which gathers the 100 higher education institutions that contribute most to science, and technological R&D.

Tampere University of Applied Sciences Fast-Lab

Finland | Education Institution

FAST-Lab works for the seamless knowledge integration of humans and machines/robots creating smart environments by capitalising the advances on Information and Communication Technologies.

Technological Institute of Informatics

Spain | Research Institute

Since 1994, ITI has developed R&D applied to the needs and problems of companies, looking for technological solutions that respond to social and economic challenges, improve industrial competitiveness, and promote a more intelligent and sustainable society.

Fondazione Bruno Kessler

Italy | Research Institute

It is the top Research Institute in Italy, ranked at the 1st place for scientific excellence within three different subject areas and for the economic and social impact according to the latest quality of research ANVUR evaluation.

Katholieke Universiteit Leuven

Belgium | Education Institution

KU Leuven is dedicated to education and research in nearly all fields. Its fifteen faculties offer classes and degree-granting academic programmes, whilst research activities are organised by departments and research groups.

Łukasiewicz Research Network – Industrial Research Institute for Automation and Measurements PIAP

Poland | Research Institute

Institute that provides Polish entities and foreign partners with solutions using the latest technologies in robotics, automation, measurement techniques and 3D printing, as well as space technologies for over 50 years.

Robotnik Automation SLL

Spain | Private Company

They design, manufacture and market mobile robots and mobile manipulators. Their company was founded in 2002 and is currently a reference in mobile robotics worldwide and in Europe.

Athenian Brewery SA

Greece | Private Company

It is the largest beer production and distribution company in Greece, with an active presence in the country for almost 60 years. Along with its business development, the company has invested in the modernization of the production process and more.

Vigo Photonics

Poland | Private Company

VIGO Photonics is a manufacturer of semiconducting materials and instruments for photonic and microelectronic, specialized in MWIR and LWIR detectors and modules. They have a complete front – end and back- end production line from epitaxy of II-IV and III-V groups, through detector chips, lasers and their assembly and integration with electronics.

Silverline

Turkiye | Private Company

The main field of activities of Silverline are the production of classical & chimney hoods, gas hobs and built-in ovens for the house appliances. Their company consists of several manufacturing firms on total basement of 103.000 m² with apprx. 1800 employees in Merzifon (Turkiye).

NTT Data Romania SA

Romania | Private Company

NTT Data Romania uses information technology to create new paradigms and values, which help contribute to a more affluent and harmonious society.

KEBA AG

Austria | Private Company

It is an automation expert represented by its own subsidiaries all over the world, from the USA to Europe and Asia. Close networking with its customers and a high level of understanding for their industries and challenges have always been as important to KEBA as offering correspondingly industry-optimized solutions.

Transilvania Digital Innovation Hub

Romania | Private Company

It is an important actor fostering innovation and digital transformation services in the Nord-West region of Romania. Now, Transilvania DIH represents a strong consortium of 13 partners: 3 Universities, 1 City Hall, 1 R&D institution and 8 clusters and is part of the European DIHs network.

COMAU S.p.A

Italy | Private Company

As a member of Stellantis, it is a worldwide leader in delivering advanced industrial automation products and system. Its portfolio includes technology and systems for electric, hybrid and traditional vehicle manufacturing, industrial robots, collaborative and wearable robotics, autonomous logistics, dedicated machining centers and interconnected digital services and products able to transmit, elaborate and analyze machine and process data.

Teknopar Industrial Automation

Turkiye | Private Company

TEKNOPAR provides customers with a wide range of services and end-to-end solutions where integrated applications of OT and IT have been supplied.

AUSTRALO Alpha Lab MT

Estonia | Private company

AUSTRALO is a marketing agency specializing in growth hacking for research and innovation. They work closely with partners in the high-tech industry, entrepreneurship hubs, academia, digital transformation, Green Deal and healthcare, advocating for a trustworthy, fair and sustainable data-driven economy.

Wings ICT Solutions Information & Communication Technologies IKE

Greece | Private Company

It is a company that delivers innovative, end-to-end digital solutions (software and hardware) and transformations for vertical sectors, namely; Environment / Nature, Utilities and Infrastructures, Production and Manufacturing and Service sectors, based on its extensive expertise in IoT, wireless infrastructures, AI and AR/VR.

Andreu World SA

Spain | Private Company

International company, leader in good design and development of seating tables with high added value, enabling people to enjoy the spaces they inhabit in a smart and healthy way.

Gender equity

Gender equity continues to be a challenge in AI and Technology. Sociocultural data will inform gender-based considerations and additional inquiries will be made if needed throughout the project to ensure gender equality is maintained. By having an awareness that we are all different, we will try to ensure that when we design, test and validate our AI algorithms, unnoticed machine learning biases and unintentional information delivery bias are minimised. During validation, we will aim to achieve balanced groups of male and female participants for completion of questionnaires, demos and feedback.

Within the framework of the project, we will strive to raise awareness about the need for greater equality in the AI sector and to see more women in it!

Collaborative environment

Painting and sanding processes are performed manually by skilled operators. Some of the pieces of furniture require very elaborate
painting which is difficult to automate.

Cobots will learn different tasks by human demonstration such as picking, loading, unloading, painting, sanding.

Humans can focus on tasks that provide a higher added value for the company.

Expected results

Increase by 20% painting process productivity
Decrease by 50% quality defects

Collaborative robotics in smart manufacturing

In the future, most operations will be executed by autonomous robots that require sensors or feedback from environmental models to

navigate through space avoiding obstacles.

New AI techniques such as Deep Learning (DL) or the evolution of RL towards DRL (Deep Reinforcement Learning) to make these sy-
stems more autonomous with a degree of understanding of the world.

Intrinsic aspects of robot design, giving rise to cobots and AI-based central controllers that will avoid any dangerous situation.

Result

Safety and security in manufacturing between people and robots, avoiding collision and limiting impact forces to acceptable levels.

Day: July 19, 2023

#AIPRISM25 : Research & development, social sciences and humanities, ethics – Cranfield University

#AIPRISM25 : Research & development and technology provider – KU Leuven

SUBSCRIBE TO OUR NEWSLETTER

SUBSCRIBE TO OUR NEWSLETTER

Expected results:

Expected results:

Expected results:

Expected results:

Expected results:

Results:

Results:

Results:

Results:

Expected results

In the future, most operations will be executed by autonomous robots that require sensors or feedback from environmental models to

Result