Digital Apollo | David A. Mindell | Talks at Google

TL;DR

Observations from autonomy and robotics in extreme environments can provide valuable insights for terrestrial and everyday environments where autonomy and robotics are becoming more prevalent.

Transcript

OWEN: OK. Welcome to Authors@Google. Thanks everybody for coming. This is the second talk I've arranged, and so I'm really happy to have you all here. For the past 15 years, David Mindell has been combining engineering and historical research to investigate the evolution of humans' relationships to machines. His engineering work has been primarily ... Read More

Key Insights

• ♻️ Autonomy and robotics in extreme environments involve a dynamic interplay between human, remote, and autonomous systems.
• 🤑 Full autonomy is often an easier technical problem than autonomy with rich human involvement, as the latter requires more sophisticated technology and design considerations.
• ⌛ Autonomy is not a fixed concept and should be understood as something bounded by time, bandwidth, and human context.
• 🪡 For autonomous systems to be trusted, humans need simplified mental models and an understanding of the system's decision-making processes.
• 🤯 Autonomy must be designed with human limitations and fatigue in mind, actively engaging the cognitive state of the operator.

Questions & Answers

Q: Why is it important to understand the interactions between human, remote, and autonomous systems?

Understanding these interactions helps us develop more effective and safe autonomous systems, as well as redefine traditional roles and identities (e.g., pilot, explorer, scientist).

Q: How does autonomy in extreme environments differ from autonomy in everyday environments?

Autonomy in extreme environments is often bounded by human interventions and is influenced by factors like time delays and limited bandwidth. In everyday environments, autonomy may have different levels and complexities, requiring a rethinking of human-machine relationships.

Q: How do extreme environments inform our understanding of autonomy and robotics?

Extreme environments provide insights into the complexities and challenges of autonomy, such as the need for human approvals, safety considerations, and the potential for system failures. These insights can be applied to improve autonomy in everyday environments.

Q: What are the key considerations when designing autonomous systems for real-world applications?

Designers must consider issues of safety, verification, validation, liability, and human trust. They need to ensure that systems can handle failures, have clear human-automation interfaces, and meet regulatory requirements.

Summary & Key Takeaways

• The speaker, David Mindell, has combined engineering and historical research to study human-machine relationships in extreme environments, such as deep ocean exploration and space flight.

• He discusses the evolution of human, remote, and autonomous systems in these environments, highlighting the ways in which they interact and affect each other.

• Mindell emphasizes that autonomy is not a fixed concept, but rather something that is bounded by human interventions and context.