Mosaic: An Extensible Framework for Composing Rule-Based and Learned Motion Planners

TL;DR

Mosaic is an extensible framework that combines rule-based and learned motion planners for autonomous driving, enhancing safety, transparency, and performance without retraining.

Contribution

It introduces a structured arbitration graph approach that decouples trajectory verification from generation, enabling effective integration of diverse planners.

Findings

Achieves state-of-the-art performance on nuPlan benchmark with 95.48 CLS-NR.

Reduces at-fault collisions by 30% compared to individual planners.

Outperforms constituent planners on interPlan benchmark by 23.3%.

Abstract

Safe and explainable motion planning remains a central challenge in autonomous driving. While rule-based planners offer predictable and explainable behavior, they often fail to grasp the complexity and uncertainty of real-world traffic. Conversely, learned planners exhibit strong adaptability but suffer from reduced transparency and occasional safety violations. We introduce Mosaic, an extensible framework for structured decision-making that integrates both paradigms through arbitration graphs. By decoupling trajectory verification and scoring from the generation of trajectories by individual planners, every decision becomes transparent and traceable. Trajectory verification at a higher level introduces redundancy between the planners, limiting emergency braking to the rare case where all planners fail to produce a valid trajectory. Through unified scoring and optimal trajectory selection, rule-based and learned planners with complementary strengths and weaknesses can be combined to yield the best of both worlds. In experimental evaluation on nuPlan, Mosaic achieves 95.48 CLS-NR and 93.98 CLS-R on the Val14 closed-loop benchmark, setting a new state of the art, while reducing at-fault collisions by 30% compared to either planner in isolation. On the interPlan benchmark, focused on highly interactive and difficult scenarios, Mosaic scores 54.30 CLS-R, outperforming its best constituent planner by 23.3% - all without retraining or requiring additional data. The code is available at github.com/KIT-MRT/mosaic.