# LiteVLA-Edge: Quantized On-Device Multimodal Control for Embedded Robotics

Source: [LiteVLA-Edge: Quantized On-Device Multimodal Control for Embedded Robotics](<https://www.sota2.com/research/paper/69a90de12d603e6de3e405dd>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-06T05:00:05.533Z

Authors: Justin Williams, Kishor Datta Gupta, Roy George, Mrinmoy Sarkar

Published: 2026-03-03T03:20:52.000Z

## Abstract

Vision-Language-Action (VLA) models provide a unified framework for perception, language conditioning, and action generation, but many existing systems remain difficult to deploy in embedded robotic settings because of their computational requirements and inference latency. In this paper, we present LiteVLA-Edge, a deployment-oriented VLA pipeline for fully on-device inference on Jetson Orin-class hardware. Our approach combines supervised image-to-action fine-tuning in FP32 with post-training 4-bit GGUF quantization and GPU-accelerated inference through the \\texttt{llama.cpp} runtime. Under our deployment configuration, LiteVLA-Edge achieves a mean end-to-end latency of 150.5\\,ms (approximately 6.6\\,Hz) while operating entirely offline within a ROS~2-integrated perception--reasoning--action pipeline. Rather than introducing a new policy objective, our contribution is a practical systems path for executing compact multimodal control models locally on embedded hardware while preserving modular interfaces between perception, reasoning, and actuation. These results establish timing feasibility for reactive language-conditioned control and provide a reproducible baseline for future task-level evaluation of on-device VLAs in robotics.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Closed-loop robotic control on General VLA/VLM Systems Comparison](<https://www.sota2.com/research/sota/closed-loop-robotic-control-on-general-vla-vlm-systems-comparison.md>): Closed-loop robotic control; General VLA/VLM Systems Comparison; Closed-Loop Frequency (Hz): 6.6.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2603.03380>)

[PDF](<https://arxiv.org/pdf/2603.03380.pdf>)
