Details
| Originalsprache | Englisch |
|---|---|
| Titel des Sammelwerks | 2022 IEEE 25th International Conference on Intelligent Transportation Systems |
| Untertitel | ITSC 2022 |
| Herausgeber (Verlag) | Institute of Electrical and Electronics Engineers Inc. |
| Seiten | 3521-3528 |
| Seitenumfang | 8 |
| ISBN (elektronisch) | 9781665468800 |
| Publikationsstatus | Veröffentlicht - 2022 |
| Veranstaltung | 25th IEEE International Conference on Intelligent Transportation Systems, ITSC 2022 - Macau, China Dauer: 8 Okt. 2022 → 12 Okt. 2022 |
Publikationsreihe
| Name | IEEE Conference on Intelligent Transportation Systems, Proceedings, ITSC |
|---|---|
| Band | 2022-October |
Abstract
Urban road traffic is a highly dynamic environment due to its many rules, complex road layouts and road user interactions. Consequently, accurate prediction of future positions remains a challenging task. Inspired by achievements in the Natural Language Processing domain recent motion prediction models leverage the Transformer architecture. Although the models produce state-of-the-art results, they completely neglect the graph semantics inherent in the data. Graphormer is a promising architecture that was proposed recently for tackling this challenge in the field of molecule science. In this paper, we show how the Graphormer architecture can be leveraged for addressing the motion prediction task. We propose a novel encoding strategy to create a locality aware Graphormer. In addition, we extend the architecture to be able to handle edge features with the self-attention mechanism. To demonstrate the effectiveness of all components we evaluate our model on the publicly available urban motion forecasting dataset from Argoverse. Quantitative and qualitative evaluations show that our model is able to make precise state-of-the-art predictions.
ASJC Scopus Sachgebiete
- Ingenieurwesen (insg.)
- Fahrzeugbau
- Ingenieurwesen (insg.)
- Maschinenbau
- Informatik (insg.)
- Angewandte Informatik
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
2022 IEEE 25th International Conference on Intelligent Transportation Systems: ITSC 2022. Institute of Electrical and Electronics Engineers Inc., 2022. S. 3521-3528 (IEEE Conference on Intelligent Transportation Systems, Proceedings, ITSC; Band 2022-October).
Publikation: Beitrag in Buch/Bericht/Sammelwerk/Konferenzband › Aufsatz in Konferenzband › Forschung › Peer-Review
}
TY - GEN
T1 - Multi-Modal Motion Prediction with Graphormers
AU - Wonsak, Shimon
AU - Al-Rifai, Mohammad
AU - Nolting, Michael
AU - Nejdl, Wolfgang
N1 - Funding Information: ACKNOWLEDGEMENT The first author affiliated with Volkswagen AG has received funding from Volkswagen AG to conduct this research.
PY - 2022
Y1 - 2022
N2 - Urban road traffic is a highly dynamic environment due to its many rules, complex road layouts and road user interactions. Consequently, accurate prediction of future positions remains a challenging task. Inspired by achievements in the Natural Language Processing domain recent motion prediction models leverage the Transformer architecture. Although the models produce state-of-the-art results, they completely neglect the graph semantics inherent in the data. Graphormer is a promising architecture that was proposed recently for tackling this challenge in the field of molecule science. In this paper, we show how the Graphormer architecture can be leveraged for addressing the motion prediction task. We propose a novel encoding strategy to create a locality aware Graphormer. In addition, we extend the architecture to be able to handle edge features with the self-attention mechanism. To demonstrate the effectiveness of all components we evaluate our model on the publicly available urban motion forecasting dataset from Argoverse. Quantitative and qualitative evaluations show that our model is able to make precise state-of-the-art predictions.
AB - Urban road traffic is a highly dynamic environment due to its many rules, complex road layouts and road user interactions. Consequently, accurate prediction of future positions remains a challenging task. Inspired by achievements in the Natural Language Processing domain recent motion prediction models leverage the Transformer architecture. Although the models produce state-of-the-art results, they completely neglect the graph semantics inherent in the data. Graphormer is a promising architecture that was proposed recently for tackling this challenge in the field of molecule science. In this paper, we show how the Graphormer architecture can be leveraged for addressing the motion prediction task. We propose a novel encoding strategy to create a locality aware Graphormer. In addition, we extend the architecture to be able to handle edge features with the self-attention mechanism. To demonstrate the effectiveness of all components we evaluate our model on the publicly available urban motion forecasting dataset from Argoverse. Quantitative and qualitative evaluations show that our model is able to make precise state-of-the-art predictions.
UR - http://www.scopus.com/inward/record.url?scp=85141867909&partnerID=8YFLogxK
U2 - 10.1109/ITSC55140.2022.9921993
DO - 10.1109/ITSC55140.2022.9921993
M3 - Conference contribution
AN - SCOPUS:85141867909
T3 - IEEE Conference on Intelligent Transportation Systems, Proceedings, ITSC
SP - 3521
EP - 3528
BT - 2022 IEEE 25th International Conference on Intelligent Transportation Systems
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 25th IEEE International Conference on Intelligent Transportation Systems, ITSC 2022
Y2 - 8 October 2022 through 12 October 2022
ER -