Evo-Depth: A Lightweight Depth-Enhanced Vision-Language-Action Model

TL;DR

Evo-Depth introduces a lightweight depth-enhanced vision-language-action model that improves robotic manipulation by integrating implicit depth features from RGB images without extra sensors, achieving high performance with low resource usage.

Contribution

It proposes a novel, efficient depth encoding and spatial enhancement framework that enhances VLA models without additional hardware or large models.

Findings

Evo-Depth outperforms existing models on four simulation benchmarks.

It achieves the highest success rate in real-world experiments.

The model is compact, with only 0.9B parameters, and uses less GPU memory.

Abstract

Vision-Language-Action models have emerged as a promising paradigm for robotic manipulation by unifying perception, language grounding, and action generation. However, they often struggle in scenarios requiring precise spatial understanding, as current VLA models primarily rely on 2D visual representations that lack depth information and detailed spatial relationships. While recent approaches incorporate explicit 3D inputs such as depth maps or point clouds to address this issue, they often increase system complexity, require additional sensors, and remain vulnerable to sensing noise and reconstruction errors. Another line of work explores implicit 3D-aware spatial modeling directly from RGB observations without extra sensors, but it often relies on large geometry foundation models, resulting in higher training and deployment costs. To address these challenges, we propose Evo-Depth, a lightweight depth-enhanced VLA framework that enhances spatially grounded manipulation without relying on additional sensing hardware or compromising deployment efficiency. Evo-Depth employs a lightweight Implicit Depth Encoding Module to extract compact depth features from multi-view RGB images. These features are incorporated into vision-language representations through a Spatial Enhancement Module via depth-aware modulation, enabling efficient spatial-semantic enhancement. A Progressive Alignment Training strategy is further introduced to align the resulting depth-enhanced representations with downstream action learning. With only 0.9B parameters, Evo-Depth achieves superior performance across four simulation benchmarks. In real-world experiments, Evo-Depth attains the highest average success rate while also exhibiting the smallest model size, lowest GPU memory usage, and highest inference frequency among compared methods.