DTDNet: Dynamic Target Driven Network for pedestrian trajectory prediction.

Predicting the trajectories of pedestrians is an important and difficult task for many applications, such as robot navigation and autonomous driving. Most of the existing methods believe that an accurate prediction of the pedestrian intention can improve the prediction quality. These works tend to predict a fixed destination coordinate as the agent intention and predict the future trajectory accordingly. However, in the process of moving, the intention of a pedestrian could be a definite location or a general direction and area, and may change dynamically with the changes of surrounding. Thus, regarding the agent intention as a fixed 2-d coordinate is insufficient to improve the future trajectory prediction. To address this problem, we propose Dynamic Target Driven Network for pedestrian trajectory prediction (DTDNet), which employs a multi-precision pedestrian intention analysis module to capture this dynamic. To ensure that this extracted feature contains comprehensive intention information, we design three sub-tasks: predicting coarse-precision endpoint coordinate, predicting fine-precision endpoint coordinate and scoring scene sub-regions. In addition, we propose a original multi-precision trajectory data extraction method to achieve multi-resolution representation of future intention and make it easier to extract local scene information. We compare our model with previous methods on two publicly available datasets (ETH-UCY and Stanford Drone Dataset). The experimental results show that our DTDNet achieves better trajectory prediction performance, and conducts better pedestrian intention feature representation.

MDST-DGCN: A Multilevel Dynamic Spatiotemporal Directed Graph Convolutional Network for Pedestrian Trajectory Prediction.

Pedestrian trajectory prediction is an essential but challenging task. Social interactions between pedestrians have an immense impact on trajectories. A better way to model social interactions generally achieves a more accurate trajectory prediction. To comprehensively model the interactions between pedestrians, we propose a multilevel dynamic spatiotemporal digraph convolutional network (MDST-DGCN). It consists of three parts: a motion encoder to capture the pedestrians' specific motion features, a multilevel dynamic spatiotemporal directed graph encoder (MDST-DGEN) to capture the social interaction features of multiple levels and adaptively fuse them, and a motion decoder to produce the future trajectories. Experimental results on public datasets demonstrate that our model achieves state-of-the-art results in both long-term and short-term predictions for both high-density and low-density crowds.

Multicomponent Spatial-Temporal Graph Attention Convolution Networks for Traffic Prediction with Spatially Sparse Data.

Predicting traffic data on traffic networks is essential to transportation management. It is a challenging task due to the complicated spatial-temporal dependency. The latest studies mainly focus on capturing temporal and spatial dependencies with spatially dense traffic data. However, when traffic data become spatially sparse, existing methods cannot capture sufficient spatial correlation information and thus fail to learn the temporal periodicity sufficiently. To address these issues, we propose a novel deep learning framework, Multi-component Spatial-Temporal Graph Attention Convolutional Networks (MSTGACN), for traffic prediction, and we successfully apply it to predicting traffic flow and speed with spatially sparse data. MSTGACN mainly consists of three independent components to model three types of periodic information. Each component in MSTGACN combines dilated causal convolution, graph convolution layer, and the weight-shared graph attention layer. Experimental results on three real-world traffic datasets, METR-LA, PeMS-BAY, and PeMSD7-sparse, demonstrate the superior performance of our method in the case of spatially sparse data.

A Deep Learning-Based Framework for Intersectional Traffic Simulation and Editing.

Most of existing traffic simulation methods have been focused on simulating vehicles on freeways or city-scale urban networks. However, relatively little research has been done to simulate intersectional traffic to date despite its broad potential applications. In this paper, we propose a novel deep learning-based framework to simulate and edit intersectional traffic. Specifically, based on an in-house collected intersectional traffic dataset, we employ the combination of convolution network (CNN) and recurrent network (RNN) to learn the patterns of vehicle trajectories in intersectional traffic. Besides simulating novel intersectional traffic, our method can be used to edit existing intersectional traffic. Through many experiments as well as comparative user studies, we demonstrate that the results by our method are visually indistinguishable from ground truth, and our method can outperform existing methods.