Towards Safe Autonomy: Emerging Requirements, Definitions, and Methods

Description

Safety and trust are among the greatest challenges in deploying Robots in a real-world use case, working alongside humans. There is a great opportunity to overcome complexity and innovate to achieve safety and trust. Roboticists have been developing various methodologies to meet this challenge. Methods ranging from safe reinforcement learning, certificate learning, certifiable algorithms, generalizability, run-time verification, and domain adaptation are finding new ways to address the question of robot safety. While this is expanding the scope of what is feasible, leaders in the industry are investing in new safety engineering tools and certification requirements. 

The workshop brings together researchers from diverse backgrounds and showcases how the problem of safety and trust is being tackled from different perspectives. The goal is to see how research in different domains is evolving to solve the overarching puzzle of safe autonomy. The goal is also to network researchers working on different aspects of safe autonomy so that better ways of defining and tackling the challenging problems emerge. 

RSS 2023 Safe Autonomy Workshop: https://sites.google.com/view/rss2023-safe-autonomy 

Join the safe-autonomy mailing list: https://www.safe-autonomy.org/  

Workshop Awards: Link.