Workshop on Safe Real-World Robot Autonomy

Overview

Robots are playing an increasingly important role in our lives. When deploying robots in the real world, safety of operation is an essential requirement. In the past, traditional perception, planning, and control techniques have been successfully applied to automate robots in static environments, where the dynamics of the robots are relatively well-characterized. However, with the continuously advancing capabilities of robots comes higher demands for safety. We expect robots to perform complex tasks in unstructured and dynamic environments, which requires robots to act safely despite uncertainties. In recent years, machine learning algorithms, especially reinforcement learning, have been developed to achieve complex tasks. However, safety and robustness considerations for these algorithms have been lacking till recently. As noted in the “Roadmap for US Robotics (2020)”, a key factor to enable next-generation robotics applications is safety, which is ingrained in all aspects of automation including perception, planning, and control. In this workshop, we bring together researchers from these fields to facilitate interdisciplinary discussions and initiate collaboration on the topic of safe autonomy for real-world applications. To this end, our workshop has four components: a tutorial on Safe Robot Autonomy and three moderated discussion sessions each addressing an important theme related to safe robot autonomy: (i) Safety Definitions and Requirements, (ii) Open Challenges and Opportunities for Integrating Theoretic and Data-driven Approaches, and (iii) Evaluation of Safety-Aware and Safety-Assured Algorithms. We envision this workshop to be an effort towards a long-term, interdisciplinary exchange on the development of safe real-world robotic systems.

Discussion Themes

Safety Definitions and Requirements

  • What are the attributes that are important to consider in theory and in experiments to evaluate algorithms for safety-critical applications (e.g., repeatability, generalization, efficiency, transferability from sim2real)?
  • What do we want to see from a theoretic perspective for machine learning (e.g., type of guarantees) and vice versa (e.g., generality of results)?
  • Is it possible to reach a consensus on the definition of safety for real-world applications?

Open Challenges and Opportunities for Integrating Theoretic and Data-Driven Approaches

  • What are some promises and limitations of traditional perception, planning, and control approaches and their learning-based counterparts?
  • Is it possible to guarantee safety a priori? How is the issue of safety addressed differently in theoretic and in data-driven approaches?
  • How do we best leverage and combine the expertise of the machine learning community and the established theoretical results when aiming for safe robot applications?

Evaluation of Safety-Aware and Safety-Assured Algorithms

  • What are some common benchmarks that are used in the robotics community? How do we measure “safety” and establish fair comparisons?
  • What are the practical issues faced in real-world robotics applications, which are lacking in current benchmarks? What factors or aspects should be further emphasized to guide algorithm design for real-world autonomy?
  • How do we move forward to maximize the accessibility of robot platforms to facilitate the transition from simulation environments to physical systems?

Invited Speakers

Naira Hovakimyan, University of Illinois Urbana-Champaign (Confirmed)
Anirudha Majumdar, Princeton (Confirmed)
Byron Boots, University of Washington (Confirmed)
Vincent Vanhoucke, Google (Confirmed)
Ryan Gariepy, Clearpath (Tentatively Confirmed)
Andreas Krause, ETH (Tentatively Confirmed)

Program

Below is a tentative program for the workshop. Times are in UTC.

13:00 – 13:10: Introduction
13:10 – 14:10: Tutorial on Recent Advances in Safe Robot Autonomy
14:10 – 14:20: Coffee break
14:20 – 15:20: Discussion Topic #1 – Safety Definitions and Requirements for Real-World Applications
15:20 – 15:30: Coffee break
15:30 – 16:30: Discussion Topic #2 – Challenges and Opportunities to Leverage Theory and Data for Safety Assurance
16:30 – 16:40: Coffee break
16:40 – 17:40: Discussion Topic #3 – Evaluation of Safety-Aware and Safety-Assured Algorithms
17:40 – 18:00: Concluding Remarks

Organizers

Angela Schoellig, University of Toronto
Melissa Greeff, University of Toronto
SiQi Zhou, University of Toronto
Animesh Garg, University of Toronto
Somil Bansal, University of Southern California

University of Toronto Institute for Aerospace Studies