Spectral Temporal Graph Neural Network for Trajectory Prediction

Abstract

An effective understanding of the contextual environment and accurate motion forecasting of surrounding agents is crucial for the development of autonomous vehicles and social mobile robots. This task is challenging since the behavior of an autonomous agent is not only affected by its own intention, but also by the static environment and surrounding dynamically interacting agents. Previous works focused on utilizing the spatial and temporal information in time domain while not sufficiently taking advantage of the cues in frequency domain. To this end, we propose a Spectral Temporal Graph Neural Network (SpecTGNN), which can capture inter-agent correlations and temporal dependency simultaneously in frequency domain in addition to time domain. SpecTGNN operates on both an agent graph with dynamic state information and an environment graph with the features extracted from context images in two streams. The model integrates graph Fourier transform, spectral graph convolution and temporal gated convolution to encode history information and forecast future trajectories. Moreover, we incorporate a multi-head spatio-temporal attention mechanism to mitigate the effect of error propagation in a long time horizon. We demonstrate the performance of SpecTGNN on two public trajectory prediction benchmark datasets, which achieves stateof-the-art performance in terms of prediction accuracy.

A Review on Intention-aware and Interaction-aware Trajectory Prediction for Autonomous Vehicles

“…Alternatively, Mo, Xing and Lv [105] used a 3 × 3 × m grid, in which the embedded features of the target vehicle and surrounding vehicles are stored, respectively, in the center and edges cells of the grid. In addition, convolutional neural networks have also been used to extracts geometrical features from the surroundings [23], [51], [106], [107].…”

A Review on Intention-aware and Interaction-aware Trajectory Prediction for Autonomous Vehicles

“…Alternatively, Mo, Xing and Lv [105] used a 3 × 3 × m grid, in which the embedded features of the target vehicle and surrounding vehicles are stored, respectively, in the center and edges cells of the grid. In addition, convolutional neural networks have also been used to extracts geometrical features from the surroundings [23], [51], [106], [107].…”

“…Alternatively, Cao et al [107] proposed a framework that applies spectral convolutional operations in the space and time, called Spectral Temporal Graph Neural Network (ST-GNN). They used Inverse Graph Fourier Transforms (IGFT) to combine both results, and also a Multi-Head Attention Mechanism (MHA) to reduce the propagation error for long time horizons.…”

“…In addition to spatial interaction, some primary studies also used graphs to represent road geometry [90], [107], [124], [136], [137]. Quehl et al [124] modeled the road network using digraph with transition probabilities that express the tendencies of a target vehicle to choose the next waypoint.…”

“…Attention mechanism is another deep learning technique that is often used with graph-based trajectory prediction, which differentiate the importance of the adjacency vertices in the target vertex embedding [51], [106], [107], [127], [130], [131], [133], [134], [136], [137]. Ye et al [134] proposed a Graph Self-Attention Network (GSAN) with dynamic weighted digraph to model the interaction of traffic participants.…”