DiffuTraj: A Stochastic Vessel Trajectory Prediction Approach via Guided Diffusion Process

TL;DR

This paper presents a novel framework to conceptualize the trajectory prediction task as a guided reverse process of motion pattern uncertainty diffusion, in which it progressively remove uncertainty from maritime regions to delineate the intended trajectory.

Abstract

Maritime vessel maneuvers, characterized by their inherent complexity and indeterminacy, requires vessel trajectory prediction system capable of modeling the multi-modality nature of future motion states. Conventional stochastic trajectory prediction methods utilize latent variables to represent the multi-modality of vessel motion, however, tends to overlook the complexity and dynamics inherent in maritime behavior. In contrast, we explicitly simulate the transition of vessel motion from uncertainty towards a state of certainty, effectively handling future indeterminacy in dynamic scenes. In this paper, we present a novel framework (\textit{DiffuTraj}) to conceptualize the trajectory prediction task as a guided reverse process of motion pattern uncertainty diffusion, in which we progressively remove uncertainty from maritime regions to delineate the intended trajectory. Specifically, we encode the previous states of the target vessel, vessel-vessel interactions, and the environment context as guiding factors for trajectory generation. Subsequently, we devise a transformer-based conditional denoiser to capture spatio-temporal dependencies, enabling the generation of trajectories better aligned for particular maritime environment. Comprehensive experiments on vessel trajectory prediction benchmarks demonstrate the superiority of our method.

Figures (9)

• Figure 1: Illustration of the gradual transition in vessel motion from a state of uncertainty to certainty throughout the reverse diffusion process.
• Figure 2: The framework of our DiffuTraj model. $\otimes$ denotes multiplication. $FC$ denotes fully-connected layer. The comprehensive details of the Encoder ($\psi$) and the Transformer-based Conditional Denoiser ($\theta$) are elaborated in section \ref{['sec4.2']} and \ref{['sec4.3']}, respectively.
• Figure 3: Ablation studies on decoder backbone network designs under the 2-hour prediction horizon on the Danish Straits dataset. Empty indicates trajectory generation without using the decoder backbone network.
• Figure 4: Prediction performance under different sampling steps $\gamma$ on the Danish Straits testing set over 4 hours. The trend of ADE is represented by blue lines, while the trend of FDE is represented by green lines.
• Figure 5: Visualization of denoised trajectories at each reverse diffusion step $k$. Our model progressively removes uncertainty and derives deterministic future trajectories through the learned reverse diffusion process. The trajectories start from a Gaussian noise distribution ($k$ = 100) and undergo a non-Markov denoising process to arrive at the predicted trajectories ($k$ = 0), successfully simulating the process from diversity to determinism.