Discover the STAR Booklet, STAR technologies and MARKET by STAR

29/12/23

Artificial intelligence (AI) systems in the manufacturing sector are increasingly replacing human tasks improving the automation of production. These systems need to be safe, trusted and secure, even when operating in dynamic, unstructured and unpred...

Share this news
Stay up to date!

Follow us on Twitter

Follow us on Linkedin

Subscribe to our monthly Newsletter!

Artificial intelligence (AI) systems in the manufacturing sector are increasingly replacing human tasks improving the automation of production. These systems need to be safe, trusted and secure, even when operating in dynamic, unstructured and unpredictable environments to be able to react to different situations and security threats. Ensuring safety and reliability of these systems is a key prerequisite for deploying them at scale and for fully leveraging the benefits of AI in manufacturing.

The Horizon 2020 STAR project has just made available the Booklet integrating a series of articles and blog posts prepared by the project partners. This article collection sheds light in many different aspects of trusted artificial intelligence for industrial use cases featuring also work and achievements of the STAR project towards safe, secure and ethical AI systems.

The articles in the Booklet are divided into the six sections, namely:

  • Part I: H2020 STAR Project Overview
  • Part II: Human-AI Systems Collaboration and Human-Centred Manufacturing
  • Part III: Industrial Systems Safety
  • Part IV: Explainable and Trustworthy AI
  • Part V: AI Cybersecurity
  • Part VI: AI Ethics and Regulations

The STAR Booklet has just become available digitally! You can download it here.

STAR technologies for trusted AI solutions

Artificial intelligence (AI) systems in the manufacturing sector are increasingly replacing human tasks improving the automation of production. These systems need to be safe, trusted and secure, even when operating in dynamic, unstructured and unpredictable environments to be able to react to different situations and security threats. Ensuring the safety and reliability of these systems is a key prerequisite for deploying them at scale and for fully leveraging the benefits of AI in manufacturing.

The Horizon 2020 STAR project is developing a number of technologies for trusted AI solutions that address different domains such as Cyber Security, Human-Robot collaboration, and Safety. These assets, resulting from continuous research, development, and validation, are crucial enablers of security and safety in production lines.

STAR components support authentication procedures, querying, browsing, accessing, and modifying data, orchestrating data flow, and can be leveraged to find holistic solutions that can increase the overall trustworthiness of several production systems.

  • Distributed Ledger Services for Data Reliability (DLSDR): A trusted decentralised solution for industrial data provenance and traceability covering tracking and tracing of raw data, AI models/algorithms and AI analytics.
  • Runtime Monitoring System (RMS): RMS provides a real time service which collects security related data from monitored IoT system components or applications. RMS enables appropriate filtering and data transformation mechanisms for reporting irregular measurements, that might be related to an attack/abuse case, and are used to drive the STAR Security Policy Manager.
  • AI Cyber-Defense Strategies (ACDS)AI Cyber defense tool for the protection of manufacturing AI data pipelines against poisoning and evasion attacks.
  • Risk Assessment and Mitigation Engine (RAME): Risk assessment and Mitigation Engine for the management of the lifecycle security incidents and risk indicators in manufacturing environments.
  • Security Policies Manager (SPM) – Security Policies Repository (SPR): The Security Policy Manager is a multipurpose tool for defining a system of rules for automatically identifying cyber threats. The rule definition is specific to the application domain and the type of data available (e.g., GPS, CPU consumption, logs). Additionally, the Security Policy Manager is agnostic to the data source, allowing the definition of policies for both hardware and software components.
  • XAI Models and LibraryA set of techniques that help develop more explainable models while at the same time preserving their high-performing learning functionalities in real-world manufacturing environments and applications.
  • Simulated Reality (SR): Synthetic data generation and intelligent oversampling methods. This can be used to improve the performance of machine learning models, when little data exists or when skewed distributions are found.
  • Active Learning (AL)Algorithms that enable finding most informative data samples from unlabelled data, which allow to increase the learning of machine learning models while minimising the labelling effort.
  • Production Processes Knowledge Base (PPKB)Prototype knowledge-graph encoding information regarding users’ perception of anomaly and heat maps showing either potential defects or where a machine learning algorithm focuses (pays attention to in the image) to determine whether a defect exists.
  • Multimodal Workers’ Training Platform: Web service that combines Natural Language Processing and Workers Training Platforms to offer a solution that allows operators to learn more about their occupation, detect knowledge gaps and get training recommendations. All offered through chatbots and multimodal interfaces.
  • Feedback ModulePrototype for feedback service implemented for a demand forecasting and logistics planning proof of concept.
  • AMR Safety: Solution for automated visual analysis and robots deployed in next generation work floor, using a computer vision module detecting empty areas merged with a dynamic robot path planning engine for secure robot displacements.
  • Human-Centred Digital TwinThe Human Digital Twin Core Infrastructure is an extensible and flexible IIoT – industrial internet of things – based platform supporting the creation of customised data representations of production systems and their entities, including humans.
  • Fatigue Monitoring SystemThe Fatigue Monitoring System uses artificial intelligence (AI) models relying on machine learning to estimate the exertion level of subjects based on static data (e.g. age, weight, etc.) and dynamic data (e.g. HR, EDA, skin temperature).
  • Workers Activity Recognition: The Workers Activity Recognition Module recognises worker’s activities by using time-series sensor data from wearable sensors including accelerometer, gyroscope, magnetometer, and capacitive sensors to optimise the interaction between humans and mobile robots and prevent collisions.

For more information about the project, visit the STAR website: www.star-ai.eu 

MARKET by STAR: the insights on the current trends, success stories and future developments related to Safe and Trusted Human-Centric Artificial Intelligence in Manufacturing

With the growth and improvement of production automation in manufacturing, Artificial Intelligence (AI) systems must be safe, trusted, and secure, even when operating in dynamic, unstructured and unpredictable environments. In this regard, one of the STAR project goals is to research and make available new technologies to enable standard-based, secure, safe, reliable and trusted human-centric AI systems in manufacturing environments.

MARKET by STAR is the entry point where all the interested visitors can consult the information representing the results of the project at one place along with the broad range of additional resources and services related to trustworthy Artificial Intelligence in manufacturing.

In the MARKET by STAR, visitors can find the following content:

  • Assets – Tools/components developed in the scope of the project. You will find a list of all the assets and a page with details for each of them.
  • Success Stories – Use cases provided by the three STAR pilots with a description and characteristics.
  • Services – Content/platforms developed directly in the scope of the project, including:

     o   STAR Courses – the four courses developed in STAR.

     o   Workshops – the workshops organised during the execution of the STAR project.

     o   STAR Book – the STAR open access book.

     o   Worker’s Training Platform – The platform oriented to skill assessment and enhancement developed in STAR.

     o   AI Trustworthiness Framework – A self-assessment form for the trustworthiness of AI systems. It consists of a set of questions that relate to different aspects AI systems.

  • External Resources – Online content/sources not developed in the scope of the project but directly related and of high relevance to the project scope and goals:

     o   External Courses – List of courses from external sources, linked with the topics of STAR, such as AI, Cybersecurity, Privacy, IoT, etc.

     o   Relevant Communities – List of relevant communities to the STAR stakeholders, such as EFFRA, AI4EU, DFA and IoT-Catalogue.com

Visit MARKET by STAR: https://www.market.star-ai.eu/