AI is undergoing a paradigm shift with the rise of models (e.g., BERT, DALL-E, GPT-3) that are trained on broad data at scale and are adaptable to a wide range of downstream tasks. We call these models foundation models to underscore their critically central yet incomplete character. This report provides a thorough account of the opportunities and risks of foundation models, ranging from their capabilities (e.g., language, vision, robotics, reasoning, human interaction) and technical principles (e.g., model architectures, training procedures, data, systems, security, evaluation, theory) to their applications (e.g., law, healthcare, education) and societal impact (e.g., inequity, misuse, economic and environmental impact, legal and ethical considerations). Though foundation models are based on standard deep learning and transfer learning, their scale results in new emergent capabilities, and their effectiveness across so many tasks incentivizes homogenization. Homogenization provides powerful leverage but demands caution, as the defects of the foundation model are inherited by all the adapted models downstream. Despite the impending widespread deployment of foundation models, we currently lack a clear understanding of how they work, when they fail, and what they are even capable of due to their emergent properties. To tackle these questions, we believe much of the critical research on foundation models will require deep interdisciplinary collaboration commensurate with their fundamentally sociotechnical nature.

This paper discusses an object-oriented event-driven simulation as a digital twin of a flexible assembly cell coordinated with a robot to perform assembly tasks alongside human. The digital twin extends the use of virtual simulation models developed in the design phase of a production system to operations for real-time control, dynamic skill-based tasks allocation between human and robot, sequencing of tasks and developing robot program accordingly. The methodology combines lean methods of manual assembly in human-robot collaboration paving path towards flexible human-robot work teams. The study is validated with an industrial case study involving dexterous assembly tasks.

Over the past years the concept of human centred automation has received a lot of attention to achieve hybrid automation. A form of hybrid automation is taking benefit from the synergistic effect of human-robot collaboration. When used in assembly, the requirement of flexibility, adaptability and safety makes the design and redesign of human-robot collaborative (HRC) systems a complex and prone to error process. The use of time-based continuous simulations can offer safe virtual space for testing and validation thus easing the design of complex HRC systems. However, conventional simulations don't allow to experience the future production system as an end-user in an immersive environment. This paper explores the technological development in VR for design of human-centered production systems and develop a unified framework to combine human-robot simulation with VR. The simulation as an event-driven simulation helped in estimating the humanrobot cycle times, developing process plan, layout optimization and robot control program. The same simulation is used in VR to interact with the production equipment and particularly with the robot. Additionally, AWS Sumerian environment is used to create a virtual robot to assist VR user in the design process.

No abstract