M³oCap: Motion Capture with Millimeter-Wave Tags

Abstract

This paper introduces M³oCap, a millimeter-wave (mmWave) tag-based motion capture system that delivers accurate 6 degrees of freedom motion tracking. M³oCap utilizes a single commercial mmWave radar and custom-designed mmWave backscatter tags (2.5 cm × 3 cm) to localize and track the motion of tagged objects. Our system features novel algorithms that effectively isolate weak backscattered signals while accurately recovering phase changes induced by motion, allowing high-rate tracking of tag movements.

Experiments show that M³oCap achieves performance close to commercial motion capture systems – it provides 1000 measurements per second with a median range tracking accuracy of 250 µm, median localization accuracies of 3.34 mm, 3.95 mm, and 4.20 mm in x, y, z dimensions, and median orientation tracking accuracies of 1.7°, 2.3°, and 2.0° across pitch, roll, and yaw. We further demonstrate the system's fine-grained movement tracking capabilities through various motion capture tasks.

System Overview

M3oCap tag next to a coin: a 25 mm by 30 mm board with a 4x4 antenna array, diodes, and an oscillator

M³oCap pairs a commercial 60 GHz mmWave radar with our custom retro-reflective backscatter tags. Each 2.5 cm × 3 cm tag uses a 4×4 Van Atta array to reflect radar signals back toward the radar, while low-power diodes modulate the reflection. A tag uses under 283 µW and costs about $39.

Signal processing pipeline: spectrum, band-pass filtering, IFFT, and multiplication of the two shifted tag signals

Our signal processing pipeline (above) has three key components:

Single-chirp signal isolation: tag modulation shifts the tag signal into frequency bands free of background reflections (a), isolating it within a single chirp, even when people or objects move nearby.
High-rate phase measurement: band-pass filters extract the two frequency-shifted copies of the tag signal (b, c), and multiplying them cancels modulation noise (d), recovering the precise motion-induced phase at 1000 fps.
Multi-tag 6-DoF tracking: phase-based matching identifies each tag, and three or more tags on an object give its 3D orientation in addition to its position.

Motion Capture Results

Tossed box tracking and robotic arm stacking with reconstructed 3D trajectories

Three human poses captured with tags on a jacket and the reconstructed skeletons

Reconstructed trajectories of a box tossed in the air (median 3D error 9.69 mm) and cubes stacked by a robotic arm (8.12 mm), compared to ground truth, and human poses captured from tags placed on clothing.

Code and Dataset

We release code and example data to facilitate future research in this direction. Instructions can be found in the readme of our github repo.

BibTeX

@inproceedings{m3ocap,
author = {Yang, Xin and Liu, Freddy Yifei and Li, Yunshuang and Gong, Yao and Jayaraman, Dinesh and Abari, Omid and Zhao, Mingmin},
title = {Motion Capture with Millimeter-Wave Tags},
year = {2026},
doi = {10.1145/3774906.3800488},
booktitle = {Proceedings of the 2026 ACM/IEEE International Conference on Embedded Artificial Intelligence and Sensing Systems},
pages = {319–333},
location = {Saint Malo, France},
series = {SenSys '26}
}

Acknowledgments

This work was supported by the NSF (CNS-2516559 and CNS-2516558).

This project page template is adapted from Nerfies.