ST‐SIGMA: Spatio‐temporal semantics and interaction graph aggregation for multi‐agent perception and trajectory forecasting
Abstract Scene perception and trajectory forecasting are two fundamental challenges that are crucial to a safe and reliable autonomous driving (AD) system. However, most proposed methods aim at addressing one of the two challenges mentioned above with a single model. To tackle this dilemma, this pap...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Wiley
2022-12-01
|
Series: | CAAI Transactions on Intelligence Technology |
Subjects: | |
Online Access: | https://doi.org/10.1049/cit2.12145 |
_version_ | 1811189134457307136 |
---|---|
author | Yang Fang Bei Luo Ting Zhao Dong He Bingbing Jiang Qilie Liu |
author_facet | Yang Fang Bei Luo Ting Zhao Dong He Bingbing Jiang Qilie Liu |
author_sort | Yang Fang |
collection | DOAJ |
description | Abstract Scene perception and trajectory forecasting are two fundamental challenges that are crucial to a safe and reliable autonomous driving (AD) system. However, most proposed methods aim at addressing one of the two challenges mentioned above with a single model. To tackle this dilemma, this paper proposes spatio‐temporal semantics and interaction graph aggregation for multi‐agent perception and trajectory forecasting (ST‐SIGMA), an efficient end‐to‐end method to jointly and accurately perceive the AD environment and forecast the trajectories of the surrounding traffic agents within a unified framework. ST‐SIGMA adopts a trident encoder–decoder architecture to learn scene semantics and agent interaction information on bird’s‐eye view (BEV) maps simultaneously. Specifically, an iterative aggregation network is first employed as the scene semantic encoder (SSE) to learn diverse scene information. To preserve dynamic interactions of traffic agents, ST‐SIGMA further exploits a spatio‐temporal graph network as the graph interaction encoder. Meanwhile, a simple yet efficient feature fusion method to fuse semantic and interaction features into a unified feature space as the input to a novel hierarchical aggregation decoder for downstream prediction tasks is designed. Extensive experiments on the nuScenes data set have demonstrated that the proposed ST‐SIGMA achieves significant improvements compared to the state‐of‐the‐art (SOTA) methods in terms of scene perception and trajectory forecasting, respectively. Therefore, the proposed approach outperforms SOTA in terms of model generalisation and robustness and is therefore more feasible for deployment in real‐world AD scenarios. |
first_indexed | 2024-04-11T14:29:44Z |
format | Article |
id | doaj.art-4b19a0242195487986043e4fb589b832 |
institution | Directory Open Access Journal |
issn | 2468-2322 |
language | English |
last_indexed | 2024-04-11T14:29:44Z |
publishDate | 2022-12-01 |
publisher | Wiley |
record_format | Article |
series | CAAI Transactions on Intelligence Technology |
spelling | doaj.art-4b19a0242195487986043e4fb589b8322022-12-22T04:18:39ZengWileyCAAI Transactions on Intelligence Technology2468-23222022-12-017474475710.1049/cit2.12145ST‐SIGMA: Spatio‐temporal semantics and interaction graph aggregation for multi‐agent perception and trajectory forecastingYang Fang0Bei Luo1Ting Zhao2Dong He3Bingbing Jiang4Qilie Liu5School of Computer Science and Technology Chongqing University of Posts and Telecommunications Chongqing ChinaSchool of Computer Science and Technology Chongqing University of Posts and Telecommunications Chongqing ChinaSchool of Communication and Information Engineering Chongqing University of Posts and Telecommunications Chongqing ChinaSchool of Electrical Engineering Korea Advanced Institute of Science and Technology (KAIST) Daejeon Republic of KoreaSchool of Information Science and Technology Hangzhou Normal University Hangzhou ChinaSchool of Communication and Information Engineering Chongqing University of Posts and Telecommunications Chongqing ChinaAbstract Scene perception and trajectory forecasting are two fundamental challenges that are crucial to a safe and reliable autonomous driving (AD) system. However, most proposed methods aim at addressing one of the two challenges mentioned above with a single model. To tackle this dilemma, this paper proposes spatio‐temporal semantics and interaction graph aggregation for multi‐agent perception and trajectory forecasting (ST‐SIGMA), an efficient end‐to‐end method to jointly and accurately perceive the AD environment and forecast the trajectories of the surrounding traffic agents within a unified framework. ST‐SIGMA adopts a trident encoder–decoder architecture to learn scene semantics and agent interaction information on bird’s‐eye view (BEV) maps simultaneously. Specifically, an iterative aggregation network is first employed as the scene semantic encoder (SSE) to learn diverse scene information. To preserve dynamic interactions of traffic agents, ST‐SIGMA further exploits a spatio‐temporal graph network as the graph interaction encoder. Meanwhile, a simple yet efficient feature fusion method to fuse semantic and interaction features into a unified feature space as the input to a novel hierarchical aggregation decoder for downstream prediction tasks is designed. Extensive experiments on the nuScenes data set have demonstrated that the proposed ST‐SIGMA achieves significant improvements compared to the state‐of‐the‐art (SOTA) methods in terms of scene perception and trajectory forecasting, respectively. Therefore, the proposed approach outperforms SOTA in terms of model generalisation and robustness and is therefore more feasible for deployment in real‐world AD scenarios.https://doi.org/10.1049/cit2.12145feature fusiongraph interactionhierarchical aggregationscene perceptionscene semanticstrajectory forecasting |
spellingShingle | Yang Fang Bei Luo Ting Zhao Dong He Bingbing Jiang Qilie Liu ST‐SIGMA: Spatio‐temporal semantics and interaction graph aggregation for multi‐agent perception and trajectory forecasting CAAI Transactions on Intelligence Technology feature fusion graph interaction hierarchical aggregation scene perception scene semantics trajectory forecasting |
title | ST‐SIGMA: Spatio‐temporal semantics and interaction graph aggregation for multi‐agent perception and trajectory forecasting |
title_full | ST‐SIGMA: Spatio‐temporal semantics and interaction graph aggregation for multi‐agent perception and trajectory forecasting |
title_fullStr | ST‐SIGMA: Spatio‐temporal semantics and interaction graph aggregation for multi‐agent perception and trajectory forecasting |
title_full_unstemmed | ST‐SIGMA: Spatio‐temporal semantics and interaction graph aggregation for multi‐agent perception and trajectory forecasting |
title_short | ST‐SIGMA: Spatio‐temporal semantics and interaction graph aggregation for multi‐agent perception and trajectory forecasting |
title_sort | st sigma spatio temporal semantics and interaction graph aggregation for multi agent perception and trajectory forecasting |
topic | feature fusion graph interaction hierarchical aggregation scene perception scene semantics trajectory forecasting |
url | https://doi.org/10.1049/cit2.12145 |
work_keys_str_mv | AT yangfang stsigmaspatiotemporalsemanticsandinteractiongraphaggregationformultiagentperceptionandtrajectoryforecasting AT beiluo stsigmaspatiotemporalsemanticsandinteractiongraphaggregationformultiagentperceptionandtrajectoryforecasting AT tingzhao stsigmaspatiotemporalsemanticsandinteractiongraphaggregationformultiagentperceptionandtrajectoryforecasting AT donghe stsigmaspatiotemporalsemanticsandinteractiongraphaggregationformultiagentperceptionandtrajectoryforecasting AT bingbingjiang stsigmaspatiotemporalsemanticsandinteractiongraphaggregationformultiagentperceptionandtrajectoryforecasting AT qilieliu stsigmaspatiotemporalsemanticsandinteractiongraphaggregationformultiagentperceptionandtrajectoryforecasting |