By Dan Smigrod, Managing Editor, Humanoid Home News.
Source: The Conversation
Context: Expert analysis of household-task reliability
Published: September 8, 2026 (source article)
Laundry offers a revealing test of household robotics, argues Wai-keung Fung, a senior lecturer at Cardiff Metropolitan University. In a September article for The Conversation, he explains why changing fabric and unpredictable arrangements demand more than repeating a practiced movement.
His argument concerns adaptation: a robot must recognize what is happening and respond when a grasp or a fold does not go as intended. He also explains why a skill learned on one robot cannot simply be assumed to work on a machine with different hands, cameras and balance requirements.
The Open X-Embodiment research project provides relevant background. Its own project page describes a shared dataset with more than one million robot trajectories spanning 22 robot embodiments. The work explores learning across platforms; those dataset figures do not establish that a consumer humanoid can complete a household laundry routine.
Measure the whole job
HHN’s practical question is what the person gets back at the end of a session. A useful test would define the starting pile, the expected result and the amount of help permitted before the robot begins. Recording abandoned garments and interventions would make the result easier to judge than a successful clip alone.
This is an explanation of evaluation challenges, not a new HHN laboratory test or a declaration that every current robot has the same abilities.
Humanoid Home News Takeaway: Household usefulness is easier to assess when the job and the help are both visible. For Humanoid Home News readers, the informative demonstration is one that shows the unfinished work and recovery as well as the successful steps.
Source: The Conversation. Author’s discussion of the article. Open X-Embodiment research.
