DRIFT: Diffusion-based Rule-Inferred For Trajectories

TL;DR

DRIFT is a diffusion-based framework that generates high-fidelity, smooth, and precise robot trajectories by integrating global topological constraints and local obstacle attention, improving trajectory planning in unstructured environments.

Contribution

This paper introduces DRIFT, a novel diffusion-based trajectory generator combining GNN-based scene perception and graph-conditioned temporal attention for improved robot path planning.

Findings

Achieves centimeter-level imitation fidelity (0.041m FDE)

Maintains competitive smoothness (27.19 Jerk)

Balances trajectory smoothness and precision effectively

Abstract

Trajectory generation for mobile robots in unstructured environments faces a critical dilemma: balancing kinematic smoothness for safe execution with terminal precision for fine-grained tasks. Existing generative planners often struggle with this trade-off, yielding either smooth but imprecise paths or geometrically accurate but erratic motions. To address the aforementioned shortcomings, this article proposes DRIFT (Diffusion-based Rule-Inferred for Trajectories), a conditional diffusion framework designed to generate high-fidelity reference trajectories by integrating two complementary inductive biases. First, a Relational Inductive Bias, realized via a GNN-based Structured Scene Perception (SSP) module, encodes global topological constraints to ensure holistic smoothness. Second, a Temporal Attention Bias, implemented through a novel Graph-Conditioned Time-Aware GRU (GTGRU), dynamically attends to sparse obstacles and targets for precise local maneuvering. In the end, quantitative results demonstrate that DRIFT reconciles these conflicting objectives, achieving centimeter-level imitation fidelity (0.041m FDE) and competitive smoothness (27.19 Jerk). This balance yields highly executable reference plans for downstream control.