Periodic Reporting for period 1 - AI-PRISM (AI Powered human-centred Robot Interactions for Smart Manufacturing)
Période du rapport: 2022-10-01 au 2024-03-31
AI-PRISM will provide industrial users with human-centred artificial intelligence (AI)-based solutions to create a more efficient, resilient, digital, sustainable and high-quality European manufacturing industry.
To do so, we will develop an integrated and scalable environment with solutions adapted to dynamic and unpredictable manufacturing scenarios that require tasks that are difficult to automate and where speed and versatility are essential to meet users' needs. Furthermore, the solutions will be specific to semi-automated and collaborative manufacturing in flexible production processes and will not require specific robotic programming skills.
Our solutions ecosystem will have four main pillars.
1. A human-centred collaborative robotic platform oriented to ease hard-to-automate manufacturing tasks.
2. A human-robot cooperative environment powered by trustworthy AI.
3. Social human-agent-robots teams collaboration — AI-based safety monitoring and robot control mechanisms to detect and avoid unsafe situations and ensure social and physical safety.
4. An open-access network portal to offer compliant infrastructure.
To evaluate our solutions' performance, transferability, scalability and large-scale deployment, we will perform demonstrations in real operating environments. Specifically, in four user pilots involving key manufacturing sectors — furniture, food/beverage, built-in appliances and electronics —, types of robots and industrial processes that are difficult to automate, plus a generic demonstration facility.
In addition to seeking quantitative improvements in the manufacturing sector, AI-PRISM aims to use technological innovation to support a paradigm shift in which AI, robotics and Social Sciences and Humanities (SSH) are integrated into the manufacturing domain for the improvement of flexible production processes, becoming a viable and widespread alternative for European factories.
During the next three years, 25 partners from 12 countries will join forces to make AI-PRISM a reality. From educational institutions to research and technology organisations, robot manufacturers, industries and use case providers; our interdisciplinary consortium brings together all the actors of the human-robot collaboration value chain and involves key experts in SSH, standardisation, exploitation, and dissemination.
To do so, we will develop an integrated and scalable environment with solutions adapted to dynamic and unpredictable manufacturing scenarios that require tasks that are difficult to automate and where speed and versatility are essential to meet users' needs. Furthermore, the solutions will be specific to semi-automated and collaborative manufacturing in flexible production processes and will not require specific robotic programming skills.
Our solutions ecosystem will have four main pillars.
1. A human-centred collaborative robotic platform oriented to ease hard-to-automate manufacturing tasks.
2. A human-robot cooperative environment powered by trustworthy AI.
3. Social human-agent-robots teams collaboration — AI-based safety monitoring and robot control mechanisms to detect and avoid unsafe situations and ensure social and physical safety.
4. An open-access network portal to offer compliant infrastructure.
To evaluate our solutions' performance, transferability, scalability and large-scale deployment, we will perform demonstrations in real operating environments. Specifically, in four user pilots involving key manufacturing sectors — furniture, food/beverage, built-in appliances and electronics —, types of robots and industrial processes that are difficult to automate, plus a generic demonstration facility.
In addition to seeking quantitative improvements in the manufacturing sector, AI-PRISM aims to use technological innovation to support a paradigm shift in which AI, robotics and Social Sciences and Humanities (SSH) are integrated into the manufacturing domain for the improvement of flexible production processes, becoming a viable and widespread alternative for European factories.
During the next three years, 25 partners from 12 countries will join forces to make AI-PRISM a reality. From educational institutions to research and technology organisations, robot manufacturers, industries and use case providers; our interdisciplinary consortium brings together all the actors of the human-robot collaboration value chain and involves key experts in SSH, standardisation, exploitation, and dissemination.
The work performed and the main achievements so far are described below organised around the main objectives of the project.
Technical and Methodological Framework Development (WP1, WP2): The activities focused on establishing a shared project vision and a robust technical and methodological framework to guide all development activities. A detailed analysis of use case scenarios and requirements was conducted, with an emphasis on human-centered objectives and needs. This analysis formed the basis for the development of a reference architecture that integrates the different system components.
Human Centered Collaborative Robotic Platform (WP3):The final objective is to develop a ROS-based collaborative robotic platform able to digitalize the collaboration environment, including human interactions, and enabling real-time communication networks and services. In this period, we have successfully developed a ROS-based framework that serves as the runtime environment for all ROS-based components of the platform. Additionally, we have integrated a comprehensive Industrial Internet of Things (IIoT) Platform to facilitate information exchange between devices, and a Data Platform that provides data storage and streaming services, along with Software Defined Networking (SDN) functions to manage real-time communications and ensure Quality of Service (QoS). The development also included components to sense and calibrate the collaboration environment and to digitalize it effectively.
AI-PRISM AI Enhancing Tools (WP4) :The objective is to develop a set of AI-based tools to enhance the reasoning, perception, and coordination capabilities of collaborative robots, facilitating human collaboration in complex environments. To this aim, we have developed a Continuous Integration/Continuous Delivery (CI/CD) framework that allows for the deployment of AI-PRISM components as Kubernetes applications in a distributed cluster architecture, enhancing modularity and flexibility to adapt to various use case scenarios. Additionally, during this period, initial releases of AI-enhancing components have been made available for deployment at pilot sites. These components include perception and reasoning modules such as zero-shot 6D pose estimation, programming by demonstration modules like kinaesthetic learning, and human-machine interfaces (HMI) such as voice control.
Main achievements are:
Social sciences to guide the design and implementation of AI-PRISM solutions (WP5) :The objective is to establish a solid scientific foundation based on social sciences to guide the design and implementation of AI-PRISM solutions and to demonstrate their impact qualitatively and quantitatively. A human-centered methodology was developed to steer the design and implementation of collaborative robotic solutions, focusing on direct interactions and feedback from end users.
During this period, multiple workshops were conducted at various pilot sites, involving experiments and interviews with end users. These workshops aimed to identify situations that cause mental or physical discomfort. The insights gained from these workshops were used to derive primary human-centered objectives and refine the requirements and specifications, ensuring they effectively enhance working conditions.
System Demonstration and Validation (WP6): The objective is to integrate, deploy, and validate the technical approach and results of the AI-PRISM project across four large-scale industrial demonstration sites: Furniture (Andreu World), Food/Beverage (Athenian Brewery), Built-in Appliances (Silverline), and Electronics (Vigo System), along with a generic demonstrator by the technology provider (Keba). Main achievements include the identification of the different technical components (hardware and software) involved in each of the pilots, and the establishment of monitoring mechanisms to control the commissioning of equipment and the development of the required modules. Currently, all hardware equipment and software modules have been defined and acquired by each pilot. The integration process is underway, with initial deployment of the modules occurring at RTO partners' laboratories for testing and pre-validation.
Network of Open Access Pilots (WP7): The main objective is to develop a network of open access pilots to engage academic and industrial actors, leveraging AI-PRISM simulation services for demonstration, training, certification, and development activities. The achievements in this period include the definition and identification of the tools and components of the Open-Access Platform, the refinement of the eatures and definitions for each tool, and the establishment of the tools and repositories to enable the development of the software tools. Additionally, we have defined and implemented a pilot use case (TAU) that will be used as a reference for forthcoming pilot sites.
Technical and Methodological Framework Development (WP1, WP2): The activities focused on establishing a shared project vision and a robust technical and methodological framework to guide all development activities. A detailed analysis of use case scenarios and requirements was conducted, with an emphasis on human-centered objectives and needs. This analysis formed the basis for the development of a reference architecture that integrates the different system components.
Human Centered Collaborative Robotic Platform (WP3):The final objective is to develop a ROS-based collaborative robotic platform able to digitalize the collaboration environment, including human interactions, and enabling real-time communication networks and services. In this period, we have successfully developed a ROS-based framework that serves as the runtime environment for all ROS-based components of the platform. Additionally, we have integrated a comprehensive Industrial Internet of Things (IIoT) Platform to facilitate information exchange between devices, and a Data Platform that provides data storage and streaming services, along with Software Defined Networking (SDN) functions to manage real-time communications and ensure Quality of Service (QoS). The development also included components to sense and calibrate the collaboration environment and to digitalize it effectively.
AI-PRISM AI Enhancing Tools (WP4) :The objective is to develop a set of AI-based tools to enhance the reasoning, perception, and coordination capabilities of collaborative robots, facilitating human collaboration in complex environments. To this aim, we have developed a Continuous Integration/Continuous Delivery (CI/CD) framework that allows for the deployment of AI-PRISM components as Kubernetes applications in a distributed cluster architecture, enhancing modularity and flexibility to adapt to various use case scenarios. Additionally, during this period, initial releases of AI-enhancing components have been made available for deployment at pilot sites. These components include perception and reasoning modules such as zero-shot 6D pose estimation, programming by demonstration modules like kinaesthetic learning, and human-machine interfaces (HMI) such as voice control.
Main achievements are:
Social sciences to guide the design and implementation of AI-PRISM solutions (WP5) :The objective is to establish a solid scientific foundation based on social sciences to guide the design and implementation of AI-PRISM solutions and to demonstrate their impact qualitatively and quantitatively. A human-centered methodology was developed to steer the design and implementation of collaborative robotic solutions, focusing on direct interactions and feedback from end users.
During this period, multiple workshops were conducted at various pilot sites, involving experiments and interviews with end users. These workshops aimed to identify situations that cause mental or physical discomfort. The insights gained from these workshops were used to derive primary human-centered objectives and refine the requirements and specifications, ensuring they effectively enhance working conditions.
System Demonstration and Validation (WP6): The objective is to integrate, deploy, and validate the technical approach and results of the AI-PRISM project across four large-scale industrial demonstration sites: Furniture (Andreu World), Food/Beverage (Athenian Brewery), Built-in Appliances (Silverline), and Electronics (Vigo System), along with a generic demonstrator by the technology provider (Keba). Main achievements include the identification of the different technical components (hardware and software) involved in each of the pilots, and the establishment of monitoring mechanisms to control the commissioning of equipment and the development of the required modules. Currently, all hardware equipment and software modules have been defined and acquired by each pilot. The integration process is underway, with initial deployment of the modules occurring at RTO partners' laboratories for testing and pre-validation.
Network of Open Access Pilots (WP7): The main objective is to develop a network of open access pilots to engage academic and industrial actors, leveraging AI-PRISM simulation services for demonstration, training, certification, and development activities. The achievements in this period include the definition and identification of the tools and components of the Open-Access Platform, the refinement of the eatures and definitions for each tool, and the establishment of the tools and repositories to enable the development of the software tools. Additionally, we have defined and implemented a pilot use case (TAU) that will be used as a reference for forthcoming pilot sites.