Drive As You Like: Strategy-Level Motion Planning Based on A Multi-Head Diffusion Model

TL;DR

This paper introduces a diffusion-based multi-head trajectory planner for autonomous driving that adapts to human preferences and dynamic instructions, achieving state-of-the-art performance and diverse, multi-modal behaviors.

Contribution

It presents a novel multi-head diffusion model with fine-tuning via Group Relative Policy Optimization and LLM-guided strategy selection for flexible, instruction-aware motion planning.

Findings

Achieves SOTA performance on nuPlan benchmark.

Generates diverse, multi-modal trajectories.

Retains strong planning capabilities after fine-tuning.

Abstract

Recent advances in motion planning for autonomous driving have led to models capable of generating high-quality trajectories. However, most existing planners tend to fix their policy after supervised training, leading to consistent but rigid driving behaviors. This limits their ability to reflect human preferences or adapt to dynamic, instruction-driven demands. In this work, we propose a diffusion-based multi-head trajectory planner(M-diffusion planner). During the early training stage, all output heads share weights to learn to generate high-quality trajectories. Leveraging the probabilistic nature of diffusion models, we then apply Group Relative Policy Optimization (GRPO) to fine-tune the pre-trained model for diverse policy-specific behaviors. At inference time, we incorporate a large language model (LLM) to guide strategy selection, enabling dynamic, instruction-aware planning without switching models. Closed-loop simulation demonstrates that our post-trained planner retains strong planning capability while achieving state-of-the-art (SOTA) performance on the nuPlan val14 benchmark. Open-loop results further show that the generated trajectories exhibit clear diversity, effectively satisfying multi-modal driving behavior requirements. The code and related experiments will be released upon acceptance of the paper.