Yingshuang Bai

Yingshuang Bai

Auto-generated profile, based on publicly available information

Yingshuang Bai

Wireless Network Research Department, SONY China Research Laboratory, Beijing, China

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

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

Representative Publications

View All Publications
30 November 2023 in IEEE Open Journal of the Communications Society
Verified
Journal Article
Open access
CSChen Sun
Chen Sun
LZLe Zhao
Le Zhao
TCTao Cui
Tao Cui
HLHaojin Li
Haojin Li
YBYingshuang Bai
Yingshuang Bai
QTQiang Tong
Qiang Tong
This paper investigates the application of artificial intelligence (AI) to wireless technology, specifically in the context of beam management (BM) in the advanced 5th-generation (5G) communication system. Our focus lies in aligning our study with the ongoing discussions within the Third Generation Partnership Project (3GPP) as of December 2022. Instead of evaluating the performance of specific AI models, we take user equipment (UE) receiver (Rx) beam prediction as an illustrative example of AI-based BM. We explore various aspects of AI model management, including model selection, monitoring, and activation/deactivation operations, from a 3GPP perspective. For model selection, we propose deploying distinct AI models for different propagation environments, categorized based on base station (BS) transmitter (Tx) beam measurement results. Reference Signal Received Power (RSRP) serves as a pivotal key performance index (KPI) for model performance monitoring. Our simulation results indicate that, instead of training one all-encompassing AI model with numerous layers for universal application, transitioning between domain-specific AI models with fewer layers yields superior performance. Model activation/deactivation procedures determine whether AI-based BM or traditional BM should be employed in a given scenario. We also introduce the use of AI for predicting the performance of both AI-based BM and traditional BM. By comparing the performance of these strategies, we can ascertain whether link performance degradation results from AI output errors or UE movement into challenging propagation environments. This approach enables the effective management of model switching between AI-based BM and traditional BM. The simulation shows that we can reduce the number of unnecessary switches by 10%.

7 October 2024 in IEEE Transactions on Wireless Communications
Verified
Journal Article
Closed access
PWPengyu Wang
Pengyu Wang
KMKe Ma
Ke Ma
YBYingshuang Bai
Yingshuang Bai
CSChen Sun
Chen Sun
ZWZhaocheng Wang
Zhaocheng Wang
SCSheng Chen
Sheng Chen
In non-cooperative communications, malicious electromagnetic interference attacks communication systems and causes higher probability of communication disruption. In order to address the challenges posed by electromagnetic interference, the wireless interference recognition technique has emerged, which identifies the interference signals without priori information. In recent years, the success of deep learning (DL) has sparked interest in introducing DL in the field of wireless interference recognition. However, most DL-based interference identification methods improve accuracy by dramatically increasing network sizes while ignoring the important effect of network inputs. For this reason, we extensively investigate the impact of different signal transformation forms of interference (called signal modalities) on performance. The artificial features of the interference signal are also utilized as one of the refined modalities, which breaks the inherent concept that artificial features are only used in the methods of feature extraction. Convolution and transformer are combined in the extraction of different modal features. In order to reduce the complexity of transformer, a dual transformer module (DTM) is proposed. Furthermore, to overcome the imbalance of modal optimization during the training process, an adaptive gradient modulation (AGM) strategy is proposed, which leads to better convergence for the multimodal training. Finally, modal information selection mechanism (MISM) selects the most appropriate modalities for each input sample, which saves computational costs. Extensive experiments demonstrate that combining multiple interference modalities is more effective than trying different networks.

29 November 2023 in IEEE Open Journal of Vehicular Technology
Verified
Journal Article
Open access
YBYingshuang Bai
Yingshuang Bai
JZJiawei Zhang
Jiawei Zhang
CSChen Sun
Chen Sun
LZLe Zhao
Le Zhao
HLHaojin Li
Haojin Li
XWXiaoxue Wang
Xiaoxue Wang
Artificial Intelligence (AI) has gained significant attention and extensive research across various fields in recent years. In the realm of wireless communication, researchers are exploring the use of AI to facilitate various physical layer (PHY) procedures. Within the standardization efforts of the Third Generation Partnership Project (3GPP), one prominent direction being explored is AI-based beam management (BM). The primary objective is to harness AI techniques for predicting optimal beams, thereby reducing measurement overhead and latency. This paper aims to discuss the progress made in AI-based beam management within the Release 18 standardization. Furthermore, through our research, we have identified the mobile speed of user equipment (UE) as a crucial factor that impacts the optimal time window size for collecting input data in AI models. We have observed an inverse correlation between UE speed and the time window size. Accordingly, to mitigate unnecessary measurement overhead and latency, we propose that the determination of the time window size for input data collection should be based on the UE speed. Additionally, we will present our simulation results and provide a comprehensive analysis of this relationship.

4 June 2025 in IEEE Transactions on Wireless Communications
Verified
Journal Article
Closed access
PWPengyu Wang
Pengyu Wang
KMKe Ma
Ke Ma
YBYingshuang Bai
Yingshuang Bai
CSChen Sun
Chen Sun
ZWZhaocheng Wang
Zhaocheng Wang
Benefiting from a large amount of unallocated bandwidth, millimeter-wave (mmWave) communications have been regarded as one of the most promising technologies. To overcome high pathloss of mmWave signals, the beamforming technique plays a fundamental role. In recent years, with the success of deep learning (DL), DL-based beam prediction methods have been widely studied to reduce the training overhead of traditional beam scanning methods. In this paper, a novel DL-based low-overhead beam prediction scheme is proposed, which is motivated by two important observations: (1) The optimal beam prediction is difficult for non-line of sight (NLOS) scenario, which limits the overall prediction accuracy. (2) On the contrary, the optimal beam can be precisely predicted with low computational costs under line of sight (LOS) scenario. Therefore, we propose a flexible network architecture, namely multi-stage network (MSN), to conduct the optimal beam prediction. Firstly, MSN contains multiple branches with gradually increasing computational complexity, and each branch carries with a classifier, which enables the MSN to have the capability of adaptively and dynamically allocating computational resources. Meanwhile, to combine the advantages of convolutional neural network (CNN) and transformer for feature extraction in MSN, we design joint CNN and transformer (JCT) module and its simplified module, namely Ghost-JCT. Secondly, we propose two pre-training strategies to effectively improve the performance of classifiers without additional computational costs. Finally, we propose confidence-based and Markov-based classifier selection strategies, which could select the appropriate classifier to strike a balance between accuracy and computational complexity. Simulation results demonstrate that MSN enjoys significant superiority in terms of computational complexity and prediction accuracy compared to its traditional counterparts.

17 October 2024 in IEEE Wireless Communications Letters
Verified
Journal Article
Closed access
PWPengyu Wang
Pengyu Wang
CSChen Sun
Chen Sun
KMKe Ma
Ke Ma
YBYingshuang Bai
Yingshuang Bai
ZWZhaocheng Wang
Zhaocheng Wang
To address the significant overhead associated with millimeter-wave (mmWave) beam management, deep learning (DL) techniques can be leveraged. However, existing DL-based beam prediction methods usually overlook the effects of hand blockage at the user equipment (UE) side, which may seriously impact the quality of beams. This paper aims to address the challenge of optimal beam prediction in the presence of hand blockage. Especially, we deploy a specific DL model at the UE side to handle the discovery of hand blockage and the corresponding beam prediction. Considering the computing resource constraints at the UE side, we propose a low-complexity shift transformer. It is similar to the traditional transformer structure, except that it replaces the attention operation with the shift operation, which could reduce the computational complexity while maintaining prediction accuracy. Furthermore, we introduce a novel beam recovery scheme under hand blockage. Simulation results show that the proposed method could efficiently predict hand blockage and the optimal beam.

15 May 2025 in IEEE Wireless Communications Letters
Verified
Journal Article
Closed access
YSYiliang Sang
Yiliang Sang
YBYingshuang Bai
Yingshuang Bai
KMKe Ma
Ke Ma
CSChen Sun
Chen Sun
PWPengyu Wang
Pengyu Wang
ZWZhaocheng Wang
Zhaocheng Wang
For millimeter-wave wireless systems, mobility management is crucial to facilitate high-quality services in dynamic scenarios, which generally includes the tasks of base station prediction, beam tracking, and positioning. Existing schemes usually adopt deep learning to separately implement these tasks to support mobility management, which suffer from the performance loss when training samples are inadequate. To address this issue, a dual-cascaded multi-task structure is proposed in this letter, which could simultaneously utilize the labeled data from various tasks to realize joint training based on their inter-task correlations. Specifically, the deep learning models for different tasks are connected through the dual-layer cascaded and residual design, enabling efficient information interaction among different tasks. Besides, a dedicated training scheme based on loss descending rate is proposed for effective multi-task optimization. Simulation results demonstrate that our proposed multi-task learning scheme significantly outperforms its conventional counterparts in mobility management, especially for dynamic wireless scenarios with inadequate training samples.