Skip to content
Agientry
← Back to Videos

UFO: Unsupervised RL Framework for Humanoid Robot Balance Recovery

Loading video
Loading video

RoboParty introduces UFO, an unsupervised reinforcement learning framework enabling humanoid robots to quickly recover balance from pushes and pulls without hand-crafted reward functions.

HumanoidReinforcement LearningBalance ControlUnsupervised RLOpen SourceRoboParty
Category: humanoid
Author: @ZappyZappy7
Date: 2026-10-07T00:00:00
Duration: 71.2s
Source: https://x.com/ZappyZappy7/status/2107796695492993241
Reference: https://roboparty.github.io/UFO/

Related Videos

Delta Intelligence unveils Delta-0: a humanoid foundation model for whole-body loco-manipulation across 69 DoFs
4:34

Delta Intelligence unveils Delta-0: a humanoid foundation model for whole-body loco-manipulation across 69 DoFs

@HumanoidRTech · 832 views · 2026-09-28
HumanoidMobile ManipulationWhole-Body Control
IROS 2026 Workshop on Athletic Humanoid Robotics: Dance Demos and an Open-Problems Panel
0:12

IROS 2026 Workshop on Athletic Humanoid Robotics: Dance Demos and an Open-Problems Panel

@ZiyunClaudeWang · 745 views · 2026-09-27
IROS 2026HumanoidAthletic Robotics
TactileStep: Tactile Foot Control for Humanoid Locomotion
0:32

TactileStep: Tactile Foot Control for Humanoid Locomotion

@YimingLi9702 · 470 views · 2026-09-26
TactileStepHumanoidTactile Sensing
Hunter V2: Building an Open-Source Humanoid From the Joints Up
1:38

Hunter V2: Building an Open-Source Humanoid From the Joints Up

@clankrmedia · 931 views · 2026-09-25
HumanoidOpen SourceAssembly
Open-RAIL: China Mobile's Real-Time Execution Layer Between VLA Models and Humanoids
1:49

Open-RAIL: China Mobile's Real-Time Execution Layer Between VLA Models and Humanoids

@CyberRobooo · 737 views · 2026-09-23
VLAHumanoidReal-time control
Light-O1: Whole-Body Intelligence Scaled by Human Action Pretraining
2:23

Light-O1: Whole-Body Intelligence Scaled by Human Action Pretraining

@LightOrigins_ · 722 views · 2026-09-21
Light-O1Light OriginsWhole-Body Control

As an Amazon Associate, we earn from qualifying purchases.