Spatio-Temporal Synergy with ViT: Enhancing Collaborative Perception and Object Detection for Heterogeneous Agents

Yuan Gao, Sicong Liu, Xiangrui Xu, Zhiyang Ding, Bin Guo, Zhiwen Yu

科研成果: 书/报告/会议事项章节会议稿件同行评审

摘要

To address the limitations of traditional heterogeneous agent cooperative sensing methods in terms of feedback latency and spatiotemporal dependencies, this paper proposes a heterogeneous agents enhancing cooperative perception and object detection system. The system is based on the Vision Transformer (ViT) model, leveraging its superior global context awareness and multimodal data fusion capabilities. Additionally, it incorporates the proposed adaptive delay position sensing module and spatiotemporal dependency dynamic modeling module, effectively resolving issues related to data transmission latency and complex spatiotemporal dependencies between agents. This significantly enhances the accuracy and timeliness of heterogeneous multi-agent collaborative sensing systems.

源语言英语
主期刊名RMELS 2024 - Proceedings of the 1st ACM International Workshop on Resource-efficient Mobile and Embedded LLM System in AIoT, Part of
主期刊副标题ACM Sensys 2024
出版商Association for Computing Machinery, Inc
3-5
页数3
ISBN(电子版)9798400712951
DOI
出版状态已出版 - 4 11月 2024
活动1st ACM International Workshop on Resource-efficient Mobile and Embedded LLM System in AIoT, RMELS 2024 - Hangzhou, 中国
期限: 4 11月 2024 → …

出版系列

姓名RMELS 2024 - Proceedings of the 1st ACM International Workshop on Resource-efficient Mobile and Embedded LLM System in AIoT, Part of: ACM Sensys 2024

会议

会议1st ACM International Workshop on Resource-efficient Mobile and Embedded LLM System in AIoT, RMELS 2024
国家/地区中国
Hangzhou
时期4/11/24 → …

指纹

探究 'Spatio-Temporal Synergy with ViT: Enhancing Collaborative Perception and Object Detection for Heterogeneous Agents' 的科研主题。它们共同构成独一无二的指纹。

引用此