Robot Videos: Multimodal Robot, Spot's Strong Arm, More

Video Friday this is your weekly selection of amazing robotics videos collected by your friends on IEEE spectrum robotics. We're also posting a weekly calendar of upcoming robotics events for the next few months. Please send us your events to turn on.

IN 2025: October 19-25, 2025, HANGZHOU, CHINA

Enjoy today's videos!

Caltech Center for Autonomous Systems and Technologies (CAST) and the Institute for Technological Innovation in Abu DhabiThe UAE recently held a demonstration of X1, a multi-robot system developed as part of a three-year collaboration between the two institutes. During the demonstration, the M4 multimodal robot developed by CAST is launched in drone mode from the back of a humanoid robot. It lands and goes into driving mode and then back again as needed. The demonstration highlighted the progress that is possible when engineers from multiple institutions are at the forefront autonomous systems and technologies do collaborate.

[ Caltech Center for Autonomous Systems and Technologies ]

Point robot performs dynamic manipulations of the entire body using a combination reinforcement learning and sampling-based control. The behavior shown in the video is completely autonomous, including dynamic selection of contacts on the arm, leg and torso, as well as coordination between manipulation and movement processes. The tire weighs 15 kg (33 lb), making its mass and inertial energy significant compared to the weight of the robot. External motion capture The system was used for ease of use, with an external computer connected via Wi-Fi performing intensive computational operations.

Spot's hand stronger than I thought. Also, the hand and foot collaboration is pretty wild.

[ Robotics and AI Institute ]

Figure 03 represents unprecedented progress in adoption humanoid robots from experimental prototypes to deployable, scalable products. By combining advanced perception and tactile intelligence with home-safe design and mass production readiness, Fig has created a platform that can learn, adapt and perform in both home and commercial environments. Figure 03, created for Helix, at home and around the world, lays the foundation for true general purpose robotics that can change the way people live and work.

The baby and dog in these clips make me very, very nervous.

[ Figure ]

Researchers have invented a new super-agile robot that can skillfully change shape thanks to amorphous characteristics similar to the popular robot. Marvel anti-hero Venom. The researchers used a special material called electromorphic gel (e-MG), which allows the robot to demonstrate shape change features that allow them to bend, stretch, and move in ways that were previously difficult or impossible through manipulation electric fields from ultra-light electrodes.

[ University of Bristol ]

This is of course very preliminary, but I like the idea four-legged robots physically helping each other overcome such obstacles.

[ Robot Perception and Learning Lab ]

Have we reached the peak of the dynamic humanoid yet?

[ Unitree ]

Dynamic manipulation, such as robotic object tossing, has recently gained attention as a new paradigm for speeding up logistics operations. However, the focus was on where the object landed, regardless of its final orientation. In this work, we present a method that allows a robot to accurately “throw and flip” objects into a desired landing pose (position and orientation).

[ LASA ]

I don't really care about “industry-oriented” quadrupeds. I'm very excited about four-legged animals that can be ridden.

[ MagicLab ]

I'm not at the point yet to trust any humanoid regarding priceless ancient relics. Any humanoid, not just a robot.

[ LimX ]

This CMU RI workshop is taught by Matt Mason, CMU Professor Emeritus, entitled “Journey Through Manipulation.”

In the talk, I will return to my career in manipulation research, focusing on projects that can provide useful lessons for others. We'll start with my beginnings in WITH The AI Lab and my master's thesis, which is still my most cited work, then continues with my arrival at CMU, a discussion with Allen Newell, the exercise of creating a coherent research program, and how that led to a second and third childhood. The conversation will conclude with a discussion of lessons learned.

[ Carnegie Mellon University Robotics Institute ]