Jianhu Liu

Mr. Jianhu Liu

orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
20 March 2025 in Applied Sciences
Verified
Journal Article
Open access
PDF file
JLJianhu Liu
Jianhu Liu
PWPing Wang
Ping Wang
XWXia WU
Xia WU
Vehicle-infrastructure cooperative perception enhances the perception capabilities of autonomous vehicles by facilitating the exchange of complementary information between vehicles and infrastructure. However, real-world environments often present challenges such as differences in sensor resolution and installation angles, which create a domain gap that complicates the integration of features from these two sources. This domain gap can hinder the overall performance of the perception system. To tackle this issue, we propose a novel vehicle–infrastructure cooperative perception network designed to effectively bridge the feature integration between vehicle and infrastructure sensors. Our approach includes a Multi-Scale Dynamic Feature Fusion Module designed to comprehensively integrate features from both vehicle and infrastructure across spatial and semantic dimensions. For feature fusion at each scale, we introduce the Multi-Source Dynamic Interaction Module (MSDI) and the Per-Point Self-Attention Module (PPSA). The MSDI dynamically adjusts the interaction between vehicle and infrastructure features based on environmental changes, generating enhanced interacting features. Subsequently, the PPSA aggregates these interacted features with the original vehicle–infrastructure features at the same spatial location. Additionally, we have constructed a real-world vehicle–infrastructure cooperative perception dataset, DZGSet, which includes multi-category annotations. Extensive experiments conducted on the DAIR-V2X and our self-collected DZGSet datasets demonstrate that our proposed method achieves Average Precision (AP) scores at IoU 0.5 of 0.780 and 0.652, and AP scores at IoU 0.7 of 0.632 and 0.493, respectively. These results indicate that our proposed method outperforms existing cooperative perception methods. Consequently, the proposed approach significantly improves the performance of cooperative perception, enabling more accurate and reliable autonomous vehicle operation.

24 June 2025 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
JLJianhu Liu
Jianhu Liu
XWXia WU
Xia WU
PWPing Wang
Ping Wang
Cooperative perception via Vehicle-to-Everything (V2X) wireless communication technology shares complementary perception information among autonomous vehicles and infrastructure, significantly enhancing perception capabilities by addressing limitations such as restricted range and occlusions inherent in on-board sensing. However, the limited bandwidth of V2X communication poses challenges for real-time transmission of the large data volumes required for early- and mid-level cooperative perception, necessitating consideration of communication efficiency alongside perception performance. To overcome this challenge, we propose a novel communication-efficient cooperative perception framework that achieves a better tradeoff between perception performance and communication cost. By exploiting the static deployment of infrastructure sensors and the heterogeneity of the perception region, the framework performs high-precision foreground segmentation on static infrastructure-side point clouds. A feature importance map is then constructed to extract sparse yet critical features for sharing, significantly reducing communication overhead. Additionally, we design a multiscale weighted attention feature fusion (MWAF) module to effectively integrate vehicle- and infrastructure-side features. The experimental results demonstrate that the proposed framework outperforms existing cooperative perception methods in terms of perception performance, while reducing communication traffic by approximately 60 times and 90 times on the DAIR-V2X and DZGSet dataset, respectively, without considering additional feature compression.