Scott Kuindersma, a research scientist at Boston Dynamics, presents Atlas research at Robotics Today, followed by a panel discussion with host Alberto Rodriguez, guest specialist Sangbae Kim, and audience members.
Kuindersma describes Boston Dynamics’ goal of building robots with human or animal level performance in movement, perception, and decision making. He distinguishes Atlas as a research platform from products such as Spot, Handle, and Pick, and summarizes the hardware features that support Atlas’s strength and agility, including high power batteries, hydraulic systems, and lightweight components.
Atlas’s research goal is to improve dynamic whole body control, with parkour serving as a demanding test of agility, environmental response, and varied movement. Kuindersma outlines a system that creates motion templates offline through trajectory optimization, then uses perception and model predictive control (MPC) to adapt and execute them on the robot.
The optimization process first plans the robot’s momentum and then finds a full body motion that satisfies that plan and task specific constraints. Kuindersma shows examples including jumps onto platforms, hand supported landings, and failed flips caused by toe slip and actuator limits. Online MPC can adjust contact locations, timing, and forces when the robot’s real motion differs from its template.
Depth cameras and plane fitting help Atlas identify nearby surfaces and plan contacts. MPC can retarget stored motions to objects in the robot’s actual surroundings, then blend consecutive jumps so the robot moves through a course without stopping. Adding single footed motions to the behavior library lets Atlas traverse obstacles with jogging steps as well as jumps.
Kuindersma says Atlas is still building a foundation in agility, control, and state estimation; future work includes perception, autonomy, planning, learning, and interaction. In the panel, he discusses using data driven methods to complement model based control, balancing model detail against computational limits, and learning from repeated attempts. He advises students to get hands on robot experience and choose methods based on the problem.
Briefs summarise talks and interviews; timestamps open the video at that moment. Claims in a talk are the speaker's own.
Robert Playter, CEO of Boston Dynamics, and Atlas engineers Zach Jackowski and Scott Kuindersma discuss training and deploying AI-powered humanoid robots at Hyundai’s Georgia factory and Boston Dynamics’ Massachusetts lab.
A Boston Dynamics Prototype Operations team member explains how the company builds and improves Atlas, from precision machining through assembly and testing, alongside its engineers.
Boston Dynamics engineers discuss the design, simulation, and dexterity of Atlas’s new hands.
Boston Dynamics’ robotics team discusses Atlas’s path to industrial work at Hyundai Motor Group’s Robotics Metaplant Application Center, with a host visiting the facility.
The Atlas Director of Robot Behavior at Boston Dynamics talks with NEXT host John Koetsier about Atlas’s design, robot intelligence, factory integration, and why legs can suit industrial work.
Chris Thorn, Atlas hardware team lead at Boston Dynamics, discusses the new Atlas design with industrial designers Aaron Abroff and James Kucera at Boston Dynamics.
