10 Hits in 3.1 sec

Social-DualCVAE: Multimodal Trajectory Forecasting Based on Social Interactions Pattern Aware and Dual Conditional Variational Auto-Encoder [article]

Jiashi Gao, Xinming Shi, James J.Q. Yu
2022 arXiv   pre-print
The proposed model is evaluated on widely used trajectory benchmarks and outperforms the prior state-of-the-art methods.  ...  Pedestrian trajectory forecasting is a fundamental task in multiple utility areas, such as self-driving, autonomous robots, and surveillance systems.  ...  Social-BiGAT (Kosaraju et al. 2019) , following Bicycle-GAN, realized multi-modal prediction by constructing a bijection between the outputted trajectories and the latent noise vector, at the expense  ... 
arXiv:2202.03954v1 fatcat:ackcjynrcrcwnkoqf73reflh3m

Scene Gated Social Graph: Pedestrian Trajectory Prediction Based on Dynamic Social Graphs and Scene Constraints [article]

Hao Xue, Du Q.Huynh, Mark Reynolds
2020 arXiv   pre-print
Pedestrian trajectory prediction is valuable for understanding human motion behaviors and it is challenging because of the social influence from other pedestrians, the scene constraints and the multimodal  ...  In the proposed SGSG, dynamic graphs are used to describe the social relationship among pedestrians.  ...  Based on Bicycle-GAN [60] , Social-BiGAT [22] develops a bijection between the output trajectories and the latent space input to the trajectory generator.  ... 
arXiv:2010.05507v1 fatcat:4pdljljosva5dl64cih7wb4m24

Conditional Generative Adversarial Networks for Speed Control in Trajectory Simulation [article]

Sahib Julka, Vishal Sowrirajan, Joerg Schloetterer, Michael Granitzer
2021 arXiv   pre-print
CSG is comparable to state-of-the-art GAN methods in terms of the benchmark distance metrics, while being simple and useful for simulation and data augmentation for different contexts such as fast or slow  ...  We present Conditional Speed GAN (CSG), that allows controlled generation of diverse and socially acceptable trajectories, based on user controlled speed.  ...  LSTM networks integrated with spatial attention pooling SGAN [10]: GAN network with max pooling approach to predict future human trajectories Graph-LSTM [48] : Graph convolution LSTM network using dynamic  ... 
arXiv:2103.11471v1 fatcat:o2o5uddd4ndwjazdqva2ykvusa

CoMoGCN: Coherent Motion Aware Trajectory Prediction with Graph Representation [article]

Yuying Chen, Congcong Liu, Bertram Shi, Ming Liu
2020 arXiv   pre-print
First, we cluster pedestrian trajectories into groups according to motion coherence. Then, we use graph convolutional networks to aggregate crowd information efficiently.  ...  Forecasting human trajectories is critical for tasks such as robot crowd navigation and autonomous driving.  ...  Social-BiGAT [12] : A generative model using Bicycle-GAN for multimodal prediction and GAT for crowd interaction modeling. Table 1 , we compare our models with various baselines.  ... 
arXiv:2005.00754v2 fatcat:icwihhwjt5epxbmjorukkba5ju

A Spatial-Temporal Attentive Network with Spatial Continuity for Trajectory Prediction [article]

Beihao Xia, Conghao Wang, Qinmu Peng, Xinge You, Dacheng Tao
2021 arXiv   pre-print
First, spatial-temporal attention mechanism is presented to explore the most useful and important information.  ...  To solve these limitations, we propose a novel model named spatial-temporal attentive network with spatial continuity (STAN-SC).  ...  Specially, we use Social GAN-P as our one baseline. -Social-BiGAT [11] : A multimodal model based on graph attention network which also considers the impact of scene when predicting.  ... 
arXiv:2003.06107v3 fatcat:7zyov5xnwvgxbiekbb7vxmmd3a

LTN: Long-Term Network for Long-Term Motion Prediction [article]

YingQiao Wang
2020 arXiv   pre-print
We first generate a set of proposed trajectories with our proposed distribution using a Conditional Variational Autoencoder (CVAE), and then classify them with binary labels, and output the trajectories  ...  Our Long-Term Network integrates both the regression and classification approaches.  ...  it's corresponding social interaction, and at the GRU decoder a full distribution of the predicted trajectory is produced. • Social-BiGAT [9] (S-BiGAT): This is an LSTM-GAN with Graph Attention Network  ... 
arXiv:2010.07931v1 fatcat:qz7hyasbenenhczi63izdajrwe

Robust Trajectory Forecasting for Multiple Intelligent Agents in Dynamic Scene [article]

Yanliang Zhu, Dongchun Ren, Mingyu Fan, Deheng Qian, Xin Li, Huaxia Xia
2020 arXiv   pre-print
Trajectory forecasting, or trajectory prediction, of multiple interacting agents in dynamic scenes, is an important problem for many applications, such as robotic systems and autonomous driving.  ...  In this paper, we present a novel method for the robust trajectory forecasting of multiple intelligent agents in dynamic scenes.  ...  The social-BiGAT method [Kosaraju et al., 2019] proposes a graph attention network to encode the interactions among humans in a scene and a recurrent encoder-decoder architecture to predict the trajectory  ... 
arXiv:2005.13133v1 fatcat:p5np5vi7kzd2toyholi3jeeyye

A Novel Graph based Trajectory Predictor with Pseudo Oracle [article]

Biao Yang, Guocheng Yan, Pin Wang, Chingyao Chan, Xiaofeng Liu, and Yang Chen
2020 arXiv   pre-print
Pedestrians'social interactions are captured by the proposed GA2T (Graph Attention social Attention neTwork) module.  ...  In this work, we propose GTPPO (Graph-based Trajectory Predictor with Pseudo Oracle), which is a generative model-based trajectory predictor.  ...  Social-BiGAT [32] : An improved version of SGAN by using the bicycle structure to train the generator.  ... 
arXiv:2002.00391v1 fatcat:22tghx7nffcfhd7vhhc57jeoae

Predicting Vehicles Trajectories in Urban Scenarios with Transformer Networks and Augmented Information [article]

A. Quintanar, D. Fernández-Llorca, I. Parra, R. Izquierdo, M. A. Sotelo
2021 arXiv   pre-print
More recently, simpler structures have also been introduced for predicting pedestrian trajectories, based on Transformer Networks, and using positional information.  ...  Our model exploits these simple structures by adding augmented data (position and heading), and adapting their use to the problem of vehicle trajectory prediction in urban scenarios in prediction horizons  ...  Another interesting approach to model spatial interactions for trajectory forecast is through Graph Convolutional (GNN) or Graph Attention (GAT) Networks.  ... 
arXiv:2106.00559v2 fatcat:qofkflvrtnhgzigs2xcrqkmxiu

A Review of Deep Learning-Based Methods for Pedestrian Trajectory Prediction

Bogdan Ilie Sighencea, Rareș Ion Stanciu, Cătălin Daniel Căleanu
2021 Sensors  
an overview of the available datasets, performance metrics used in the evaluation process, and practical applications.  ...  The current paper reviews the most recent deep learning-based solutions for the problem of pedestrian trajectory prediction along with employed sensors and afferent processing methodologies, and it performs  ...  /0.58 FDE: 0.78/1.18 Social-BiGAT [98] Graph-based generative adversarial network in the form of graph attention network (GAT) that learns reliable feature representations that encode the social interactions  ... 
doi:10.3390/s21227543 pmid:34833619 pmcid:PMC8619260 fatcat:u7wapci74jdlpcbpn2uu5ljomy