RAD-LAD: Rule and Language Grounded Autonomous Driving in Real-Time

TL;DR

This paper introduces LAD, a fast real-time language-action planner, and RAD, a rule-based planner, demonstrating that combining both enhances autonomous driving with reliable and adaptive decision-making.

Contribution

The paper presents LAD and RAD, two complementary planning systems, and shows their integration improves real-time autonomous driving performance and explainability.

Findings

LAD achieves ~3x lower latency than prior models.

RAD attains state-of-the-art performance among rule-based planners.

Hybrid planning combines strengths of rules and language for better driving decisions.

Abstract

We present LAD, a real-time language--action planner with an interruptible architecture that produces a motion plan in a single forward pass (~20 Hz) or generates textual reasoning alongside a motion plan (~10 Hz). LAD is fast enough for real-time closed-loop deployment, achieving ~3x lower latency than prior driving language models while setting a new learning-based state of the art on nuPlan Test14-Hard and InterPlan. We also introduce RAD, a rule-based planner designed to address structural limitations of PDM-Closed. RAD achieves state-of-the-art performance among rule-based planners on nuPlan Test14-Hard and InterPlan. Finally, we show that combining RAD and LAD enables hybrid planning that captures the strengths of both approaches. This hybrid system demonstrates that rules and learning provide complementary capabilities: rules support reliable maneuvering, while language enables adaptive and explainable decision-making.