Metraipolis › archive › dc_debd62013872

Bimanual Manipulation Within an 8 GB Budget: Zero-Copy Sensing and Quantized ACT on an Entry-Level Jetson — arXiv:2608.03938

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.03938
License (as archived)
CC0-1.0
Archived
2026-10-31T13:08:25.752Z
SHA-256 of the archived text
027e6650a31e4c40fce96b188660a9687939e7c51a8a93d30da1bbccc473cffe
Stable URL of this page
https://metraipolis.com/documento/dc_debd62013872

How to cite

Bimanual Manipulation Within an 8 GB Budget: Zero-Copy Sensing and Quantized ACT on an Entry-Level Jetson — arXiv:2608.03938 — arXiv (en), https://arxiv.org/abs/2608.03938 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_debd62013872 (Metraipolis archive dc_debd62013872, sha256 027e6650a31e4c40fce96b188660a9687939e7c51a8a93d30da1bbccc473cffe)
BibTeX
@misc{dcdebd62013872,
  title = {Bimanual Manipulation Within an 8 GB Budget: Zero-Copy Sensing and Quantized ACT on an Entry-Level Jetson — arXiv:2608.03938},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.03938},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_debd62013872; sha256 027e6650a31e4c40fce96b188660a9687939e7c51a8a93d30da1bbccc473cffe},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.03938. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Bimanual manipulation policies trained with imitation learning are typically evaluated on workstation or datacenter-class GPUs, leaving the cost of deploying them on embedded hardware largely uncharacterized. We present a bimanual SO-101 system running entirely on an NVIDIA Jetson Orin Nano Super (8 GB), the entry-level tier of NVIDIA's embedded line, using a desktop GPU (RTX 3070) only for offline training, evaluated on pick-and-place of a deformable beanbag. First, we build a GStreamer capture pipeline backed by NVMM buffers that removes redundant host-device copies from three-camera sensing. Contrary to expectation, the conventional path fit the memory budget and dropped no frames; what zero-copy sensing recovers is CPU headroom (peak single-core utilization 98.0% to 77.0%) and worst-case latency (117.31 ms to 101.52 ms). Second, we train ACT and Diffusion Policy on identical demonstrations, each at its own reference budget (100k gradient steps for ACT, 200k for Diffusion Policy). ACT converges to a task-competent policy (19/20 trials) while Diffusion Policy does not converge to a usable one (0/10) even at twice the step count, which we attribute to differing convergence costs rather than an accuracy ceiling. Third, we convert ACT to TensorRT. FP16 reduces mean inference latency from 114.02 ms to 17.93 ms (6.4x) and INT8 to 12.65 ms (9.0x), with task success preserved at all three precisions (19/20, 18/20, 19/20). We report two findings not previously documented for ACT: TensorRT's general-purpose INT8 calibration quantizes the ResNet18 backbone but accepts zero of 145 transformer layers, explaining INT8's negligible size reduction over FP16 (0.9%) despite a further 28% latency gain; and the need for quantization is conditional on ACT's action-chunking configuration, feasible in full precision at n_action_steps = 100 but not at the per-step re-prediction temporal ensembling requires.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-08-05 in cs.RO).
Categorie arXiv: Robotics (cs.RO) ; Systems and Control (eess.SY).
Pagina del lavoro: https://arxiv.org/abs/2608.03938 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-31T12:55:56.569Z dalla pagina https://arxiv.org/catchup/cs.RO/2026-08-05?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).