Sakhanda Wire
NVDA $153.63 -1.61% MSFT $506.84 -2.15% GOOGL $280.27 -1.22% META $619.77 -1.06% AMZN $256.65 -0.43%
← Back to the news

Reward AI Introduces OM-1 Robot Policy Trained Only on Human Demonstrations

Reward AI has introduced OM-1, or Omnibody Model 1, a general-purpose manipulation policy trained exclusively on human demonstrations recorded with a sensorized 7-DoF wearable glove. The company says the training used neither teleoperation data nor data collected by robots. OM-1 is designed to control industrial arms and humanoids at human speed, while its underlying approach is intended to work across different robot bodies.

The glove combines tactile and proximity sensing, cameras, and hand-pose tracking. Reward AI reports that electromagnetic tracking reduced mean overshoot error at 67 cm/s to 9.5 mm, compared with 24.9 mm for visual-inertial tracking, a 60% reduction. Each sensor stream is processed at its native sampling rate, while a separate reinforcement-learning control layer handles dynamics, disturbances, and delays in simulation.

According to Reward AI, OM-1 can learn a new task from fewer than 30 minutes of human data. The company says published demonstrations cover arms, legged humanoids, and wheeled mobile manipulators, with clips shown at normal speed. However, no model weights, code, dataset, API, success rates, benchmark comparisons, or research paper have been released, so the system is not currently available for independent deployment.

Written by Sakhanda Wire AI from reporting by MarkTechPost. Facts, quotes and full context are in the original article below.

Originally published by MarkTechPost on

Read the original on MarkTechPost ↗

The reporting and the image belong to MarkTechPost and are credited and linked above; the text on this page is Sakhanda Wire's own summary.

← Back to the news

More stories

All the latest news