Search papers, labs, and topics across Lattice.
This paper introduces the Adaptive Slow-System Control Gate (ASSCG), which optimizes the invocation of slow systems in autonomous driving planning by making resource-aware decisions on whether to refresh, reuse, or suppress slow guidance. By employing an RWKV backbone and a combination of supervised fine-tuning and GRPO-style reinforcement fine-tuning, ASSCG significantly enhances the performance of fast-slow architectures. The results show a notable improvement in evaluation scores and a substantial reduction in inference latency, demonstrating the efficacy of ASSCG in real-world driving scenarios.
ASSCG cuts inference latency by 60% while boosting performance scores in autonomous driving systems, redefining how LLMs can be efficiently integrated into fast-slow planning architectures.
Large language models (LLMs) can improve autonomous driving planning but are costly to query online, and existing fast-slow planners often rely on hand-designed triggering rules that either over-call the slow system or call it at the wrong times. We formulate slow-system invocation as a resource-aware sequential decision problem and propose the Adaptive Slow-System Control Gate (ASSCG), which makes frame-level Query/Cache/Drop decisions to refresh, reuse, or suppress slow guidance. ASSCG uses an RWKV backbone for efficient long-horizon gating and is trained with supervised fine-tuning followed by GRPO-style compute-aware reinforcement fine-tuning. We apply ASSCG to two different fast-slow architectures: (i) AsyncDriver on nuPlan Hard20 closed-loop evaluation, where ASSCG improves score to 67.28 (+2.28) while reducing average end-to-end inference latency by 60%; and (ii) a RecogDrive-based dual system that we build by replacing its original VLM-2B module with a lightweight ViT-based fast planner and adding an LLM slow planner, evaluated on NAVSIM, where ASSCG achieves 91.4 PDMS (+0.6) and increases average speed by 25%. The project page, including video visualizations and additional results, is available at https://williamxuanyu.github.io/asscg/.
Tsinghua AI