FIVE-VLA: 再帰的行動メモリによる高速で効果的な自動運転
FIVE-VLA: Fast and EffectIVE Autonomous Driving with Recurrent Action Memory
高解像度画像をわずか98トークンに圧縮する効率的な視覚エンコーダと、過去の行動トークンを条件に用いる再帰的行動メモリ(RAM)を組み合わせ、641Mパラメータで従来のVLAより高速・高精度に軌道予測を行う自動運転モデル。
詳しい要約
1. どんなもの？
2. 先行研究と比べてどこがすごい？
3. 技術・手法の肝は？
4. どうやって有効だと検証した？
5. 議論はある？
6. 次に読むべき論文は？
※ AIが要旨から生成した要約です。正確性は原文をご確認ください。
著者: Kemal Oksuz, Alexandru Buburuzan, Yuhan Yao, Puneet K. Dokania
分類: cs.CV, cs.RO
原文アブストラクト
State-of-the-art vision-language-action models (VLA) for autonomous driving face critical limitations: excessive parameter counts, inefficient high-resolution image processing, and lack of temporal memory. We introduce Fast and EffectIVE VLA (FIVE-VLA) to address these through two key contributions. First, we employ an efficient vision encoder that processes high-resolution ($448 \times 896$) images while generating only 98 tokens, over $5\times$ fewer than existing approaches, and bypass text generation entirely for single-pass trajectory prediction. Second, we propose Recurrent Action Memory (RAM), a lightweight module that conditions action prediction on previous action tokens, providing temporal context critical for manoeuvres such as overtaking and emergency braking. With only 641M parameters, FIVE-VLA completes $\sim$10% more routes without traffic rule infractions than the previous state-of-the-art VLA on the challenging Bench2Drive closed-loop driving benchmark. Non-reactive open-loop simulation on the large-scale real-world NVIDIA Physical AI AV dataset shows 10.2% and 7.7% lower collision-violation rates than SimLingo in single- and four-view settings, respectively. Additionally, FIVE-VLA runs at $\sim$30 fps on an A100 and $\sim$4 fps on a T4 GPU (proxy to an edge device), representing an 8-30$\times$ speedup over previous methods.