Demo: Vision-Language Model-Guided Online Calibration of an Electromagnetic Digital Twin
Demo: Vision-Language Model-Guided Online Calibration of an Electromagnetic Digital Twin
This research introduces a vision-language model (VLM)-guided framework for online calibration of an electromagnetic digital twin, using a Unitree G1 robot and NVIDIA Sionna. The framework employs material classification and waypoint planning to enhance calibration accuracy and reduce travel costs, achieving a normalized mean absolute conductivity error of $1.74 imes10^{-4}$ within 20 meters of travel.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
travel distance
20
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
material classification maps visible materials through ITU-R P.2040 to conductivity priors for Sionna's gradient descent on accumulated received signal strength (RSS) measurements; waypoint planning selects the next measurement location online using residual RSS calibration error and image coverage. In a real indoor scenario, the framework achieves a normalized mean absolute conductivity error of $1.74\times10^{-4}$
Open source S3
within 20 m of travel; random initialization never converges, while random waypoints require over twice the travel.
Open source S4
Source：arXiv Robotics — research abstracts · arxiv.org