James Kuffner of Toyota Research Institute reflects on 30 years in robotics and the challenges of building safe, intelligent machines, speaking at an IEEE Robotics and Automation Society event.
Kuffner explains that factory robots work in structured settings, while robots in the wider world must handle unknown objects and obstacles. They need online perception, motion planning, and task selection instead of relying on fixed models and sequences.
Humanoid motion planning is difficult because robots have many degrees of freedom, balance needs, and real-time constraints. Kuffner and Steve LaValle developed RRT-Connect, a bidirectional search method, and used it for robot arms; cloud simulations also helped identify stable grasps under perturbations and noise.
Kuffner describes work by Mike Stilman and Satoshi Kagami that combined navigation planning with manipulation. Robots could move obstacles to create free space, then plan a route through it.
Kuffner recounts joining Google’s early self-driving car project and learning that strict rule following could prevent vehicles from merging in real traffic. He later helped establish Toyota Research Institute and says roughly 90% of automated-driving software supports the onboard system through tools for machine learning, simulation, deployment, and data management. He argues that simulation can accelerate testing of rare edge cases, alongside hardware validation.
Kuffner says drivers behave differently by region, so autonomous vehicles should learn local traffic patterns and drive predictably. He distinguishes planning and reasoning from machine learning, which supplies experience from data. He concludes that safety depends on reliable hardware and software, rigorous testing, and human validation of a robot’s highest-level logic.
Briefs summarise talks and interviews; timestamps open the video at that moment. Claims in a talk are the speaker's own.
James Kuffner of Toyota and Woven Planet discusses the future of intelligent machines at the IROS 2022 keynote, with audience questions afterward.
A Toyota Research Institute robotics team presenter describes how Toyota aims to train adaptable robots in factories alongside team members.
Akio Toyoda, president of Toyota Motor Corporation, and Gill Pratt, CEO of the new Toyota Research Institute, outline plans for AI research at a Toyota press conference in Japan.
