Ya Wu

Ya Wu

Auto-generated profile, based on publicly available information

Ya Wu

Department of Automotive Engineering, Tongji University, Shanghai, China

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

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

Representative Publications

View All Publications
27 September 2021 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
YWYa Wu
Ya Wu
GCGuang Chen
Guang Chen
Zhijun LiZhijun Li
Zhijun Li
LZLijun Zhang
Lijun Zhang
LXLu Xiong
Lu Xiong
AKAlois Christian Knoll
Alois Christian Knoll
Predicting the future trajectories of surrounding agents has become an crucial problem to be solved for the safety of autonomous vehicles. Recent studies based on Long Short Term Memory (LSTM) networks have shown powerful abilities to model social interactions. However, many of these approaches focus on spatial interactions of the neighborhood agents but ignore temporal interactions that accompany spatial interactions. In this paper, we propose a Hierarchical Spatio-Temporal Attention architecture (HSTA), which activates the utilization of spatial interactions with different weights, and jointly considers the temporal interactions across time steps of all agents. More specially, the graph attention mechanism (GAT) is presented to capture spatial interactions, the multi-head attention mechanism (MHA) is conducted to encode temporal correlations of interactions and a state gated fusion (SGF) layer is used to integrate spatial and temporal interactions. We evaluate our proposed method against baselines on both pedestrian and vehicle datasets. The results show that our model is effective and achieves state-of-the-art achievements.

23 December 2022 in IEEE Robotics and Automation Letters
Verified
Journal Article
Closed access
YWYa Wu
Ya Wu
ABAriyan Bighashdel
Ariyan Bighashdel
GCGuang Chen
Guang Chen
GDGijs Dubbelman
Gijs Dubbelman
PJPavol Jancura
Pavol Jancura
Learning to predict the trajectories of pedestrians is essential for improving safety and efficiency of mobile robots. The prediction is challenging since the robot needs to operate in multiple environments in which the motion patterns of pedestrians are different between environments. Existing pedestrian trajectory prediction models heavily rely on the availability of representative data samples during training. In the presence of additional training data from a new environment, these models must be retrained on all datasets to avoid catastrophic forgetting of the knowledge obtained from the already supported environments. In this paper, we address this catastrophic forgetting problem in the context of learning to predict the trajectories of pedestrians. We propose a pseudo-rehearsal approach based on a novel Generative Replay (GR) model, referred to as Social-GR . The proposed method is consistent with crowd motion patterns and is free of any explicit reference to past experiences. To demonstrate the problem of catastrophic forgetting and evaluate our solution, we develop the Continual Trajectory Prediction Benchmark , which consists of four tasks, each representing a real-world pedestrian trajectory dataset from a different environment. By conducting several experiments, we show that our proposed Social-GR approach significantly outperforms other continual learning methods that depend on explicit experience replay, including the state-of-the-art conditional-GR model. We further illustrate the robustness of our proposed approach to mitigating catastrophic forgetting by switching the order of environments and employing a more complex prediction model.

23 December 2024 in IEEE Signal Processing Letters
Verified
Journal Article
Closed access
ZLZhengfa Liu
Zhengfa Liu
YWYa Wu
Ya Wu
DZDequan Zeng
Dequan Zeng
SDShihang Du
Shihang Du
BPBoyang Peng
Boyang Peng
In cross-domain pedestrian trajectory prediction, most existing methods usually focus on learning entangled spatialtemporal domain-invariant features, while ignoring the different contributions of spatial and temporal shifts to the prediction model. To address this issue, we propose a novel gate-Calibrated Double Disentangled Distribution Matching Network (CD3MN) that can effectively eliminate cross-trajectory domain shifts at both the spatial and temporal levels while learning robust prediction using a calibrated gated-fusion. The key idea of CD3MN is to model domain-invariant features across trajectories as a calibrated gated-fusion of disentangled domain-invariant features at the temporal and spatial levels. We first introduce a spatial-temporal disentanglement module to disentangle the spatial-temporal properties of pedestrian trajectories from the spatial-level and temporal-level. Secondly, we design a domaininvariant disentanglement module for learning domain-invariant sample-level transferable feature representations at the spatial and temporal levels. Finally, to effectively fuse these disentangled temporal and spatial features, we design a calibrated gatedfusion module where both inter-level and intra-level knowledge are introduced to calibrate the fusion gate. Extensive experiments on real datasets demonstrate the effectiveness of CD3MN.

22 September 2022 in Journal of Advanced Transportation
Verified
Journal Article
Open access
PDF file
ZLZhengfa Liu
Zhengfa Liu
GCGuang Chen
Guang Chen
YWYa Wu
Ya Wu
JDJiatong Du
Jiatong Du
JCJörg Conradt
Jörg Conradt
AKAlois Christian Knoll
Alois Christian Knoll
An important aspect of the perception system for intelligent vehicles is the detection and signal measurement of vehicle taillights. In this work, we present a novel vision-based measurement (VBM) system, using an event-based neuromorphic vision sensor, which is able to detect and measure the vehicle taillight signal robustly. To the best of our knowledge, it is for the first time the neuromorphic vision sensor is paid attention to for utilizing in the field of vehicle taillight signal measurement. The event-based neuromorphic vision sensor is a bioinspired sensor that records pixel-level intensity changes, called events, as well as the whole picture of the scene. The events naturally respond to illumination changes (such as the ON and OFF state of taillights) in the scene with very low latency. Moreover, the property of a higher dynamic range increases the sensor sensitivity and performance in poor lighting conditions. In this paper, we consider an event-driven solution to measure vehicle taillight signals. In contrast to most existing work that relies purely on standard frame-based cameras for the taillight signal measurement, the presented mixed event/frame system extracts the frequency domain features from the spatial and temporal signal of each taillight region and measures the taillight signal by combining the active-pixel sensor (APS) frames and dynamic vision sensor (DVS) events. A thresholding algorithm and a learned classifier are proposed to jointly achieve the brake-light and turn-light signal measurement. Experiments with real traffic scenes demonstrate the performance of measuring taillight signals under different traffic conditions with a single event-based neuromorphic vision sensor. The results show the high potential of the event-based neuromorphic vision sensor being used for optical signal measurement applications, especially in dynamic environments.

4 November 2025 in IEEE Transactions on Circuits and Systems for Video Technology
Verified
Journal Article
Closed access
TWTianhang Wang
Tianhang Wang
FLFan Lu
Fan Lu
SQSanqing Qu
Sanqing Qu
BLBin Li
Bin Li
YWYa Wu
Ya Wu
GCGuang Chen
Guang Chen
Collaborative perception seeks to mitigate the limitations of single-vehicle perception, such as occlusions, by facilitating communication and information sharing among connected vehicles. However, most existing works assume a homogeneous scenario where all vehicles share identity sensor types and perception model architectures. In contrast, real-world systems often involve heterogeneous agents with diverse sensor configurations and independently developed models. In such settings, directly exchanging features without proper alignment can significantly degrade performance and hinder effective collaboration. While some methods have been proposed to address heterogeneity, they typically require retraining or access to internal model parameters, making them impractical for scalable deployment. To address these challenges, we propose DiffAlign, a plug-and-play adapter that enables feature alignment across heterogeneous agents in a training-free and model-agnostic manner. DiffAlign treats received BEV features as noisy latent representations and progressively refines them through a pretrained diffusion process. This alignment strategy does not require access to model internals or any retraining, which makes it both scalable and privacy-preserving while supporting diverse sensor modalities and perception backbones. Extensive experiments on simulated OPV2V and real-world V2V4Real datasets demonstrate that DiffAlign consistently improves detection performance in heterogeneous settings, improving CoBEVT by 132.01% and 91.95%, respectively. Our method provides a practical path toward scalable, generalizable, and deployment-ready collaborative perception.