Yiruo Lin

Dr. Yiruo Lin

orcid

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

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

Representative Publications

View All Publications
21 January 2022 in IEEE Transactions on Communications
Verified
Journal Article
Closed access
YLYiruo Lin
Yiruo Lin
KYKegen Yu
Kegen Yu
LHLianxiao Hao
Lianxiao Hao
JWJin Wang
Jin Wang
Jinwei BuJinwei Bu
Jinwei Bu
This paper focuses on improving indoor Wi-Fi localization by mitigating the effect of fluctuation of received signal strength indication (RSSI). The RSSI data collected at each reference point is first transformed through translation and scaling. The BP (Back Propagation) neural network is then used to construct the ranging model using the transformed RSSI to determine the distances between the target point and each reference point. A genetic algorithm (GA) is developed to optimize the initial values of weights and biases of the BP neural network. For convenience, our proposed ranging model is denoted as GTBPD. A new localization algorithm is then proposed, which uses the GTBPD model and the sequential quadratic programming (SQP, an iterative nonlinear optimization algorithm), and the algorithm is denoted as GTBPD-LSQP for simplicity. Experiments were conducted in three areas of two different teaching buildings with complex environments. The performance of the proposed GTBPD-LSQP algorithm is evaluated and compared with four existing algorithms. The experimental results show that our proposed GTBPD-LSQP algorithm achieves significantly higher location accuracy than the four existing algorithms.

29 July 2024 in IEEE Sensors Journal
Verified
Journal Article
Closed access
YLYiruo Lin
Yiruo Lin
KYKegen Yu
Kegen Yu
FZFeiyang Zhu
Feiyang Zhu
Jinwei BuJinwei Bu
Jinwei Bu
XDXiaoming Dua
Xiaoming Dua
Wi-Fi positioning has drawn great attention in the field of indoor positioning, due to its low cost, easy deployment and large positioning range. However, the Wi-Fi signal is highly volatile due to multi-path propagation in indoor environments, which seriously affects the positioning accuracy. Deep learning, a subset of machine learning, is particularly suited to handle the effect of signal fluctuation for Wi-Fi indoor positioning, due to its good nonlinear mapping and good fault tolerance. This paper presents a timely, systematic and comprehensive review on deep learning-based Wi-Fi indoor positioning. Specifically, this review mainly focuses on the basic theory of Wi-Fi indoor positioning, the organization of the latest literatures on deep learning-based Wi-Fi indoor positioning, and statistical analysis on the function of deep learning models in Wi-Fi positioning, measurement data for the deep learning models, the source of the test dataset, and the positioning accuracy under each deep learning model. We also present a generalization of the process of building a Wi-Fi positioning model through deep learning. Furthermore, we discuss the challenges of deep learning-based Wi-Fi indoor positioning and the trends of its future development.

5 May 2022 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
JWJin Wang
Jin Wang
KYKegen Yu
Kegen Yu
Jinwei BuJinwei Bu
Jinwei Bu
YLYiruo Lin
Yiruo Lin
SHShuai Han
Shuai Han
Due to the strong penetration ability and high transmission rate of ultra-wideband (UWB) signals, UWB technology plays a very significant role in the field of precise indoor positioning. However, the harsh and volatile indoor environment leads to non-line-of-sight (NLOS) propagation and severe attenuation of UWB signals, which may generate significant ranging and positioning errors. To mitigate NLOS effect and improve positioning accuracy, existing methods use ranging information and channel impulse response (CIR) to identify UWB signal propagation channels. However, these NLOS identification methods often require a priori knowledge and suitable thresholds, and most of them only perform a binary classification between LOS and NLOS in a particular scenario. To address these disadvantages, this paper proposes a novel multi-classification method to classify UWB signal propagation channels based on one-dimensional wavelet packet analysis (ODWPA) and convolutional neural network (CNN). This method first decomposes the CIR of known categories into colored coefficients images using a one-dimensional wavelet packet function, then uses these images to train CNNs with different structures, and finally selects CNN model with the best performance to identify the unknown category of UWB signal propagation channels. The experimental results show that the performance of the proposed method is always the best, regardless of the type of experimental scenario, and the identification accuracy is always higher than 90% and up to 100%.

18 March 2025 in IEEE Geoscience and Remote Sensing Magazine
Verified
Journal Article
Closed access
Jinwei BuJinwei Bu
Jinwei Bu
HLHuan Li
Huan Li
KYKegen Yu
Kegen Yu
Weimin HuangWeimin Huang
Weimin Huang
QWQiulan Wang
Qiulan Wang
XZXiaoqing Zuo
Xiaoqing Zuo
Spaceborne GNSS reflectometry (GNSS-R) missions have been successfully launched in recent years, such as Technology Demonstration Satellite (TDS-1) in 2014, Cyclone Global Navigation Satellite System (CYGNSS) in 2016, Bufeng (BF)-1 A/B in 2019, and Fengyun (FY)-3E/3F/3G and Tianmu-1, launched successively in 2021. They provide a large amount of data to support spaceborne GNSS-R remote sensing applications, and spaceborne GNSS-R technology has also been widely used in various remote sensing fields by virtue of its advantages. With the rise of artificial intelligence (AI), many machine learning (ML) models have been developed for GNSS-R observations to estimate geophysical parameters. In particular, deep learning (DL) techniques have proved to have great potential to improve the accuracy of retrieval models in spaceborne GNSS-R applications, including ocean, land, cryosphere, atmosphere, and environment monitoring. This article provides the first comprehensive review of the application of ML in GNSS-R for Earth observation and remote sensing. The article first summarizes common ML algorithms as well as their basic concepts and theories. It then thoroughly reviews the progress of ML methods in the field of spaceborne GNSS-R and discusses the advantages, disadvantages, and challenges of ML models applied to GNSS-R. More importantly, it is imperative to adopt DL into the field of GNSS-R remote sensing and use it as a general model to tackle unprecedented, large-scale, and impactful challenges in areas such as ocean, land, cryosphere, atmosphere, hydrology, and environment remote sensing.

16 May 2023 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
FZFeiyang Zhu
Feiyang Zhu
KYKegen Yu
Kegen Yu
YLYiruo Lin
Yiruo Lin
Changyang WangChangyang Wang
Changyang Wang
JWJin Wang
Jin Wang
MCMinghua Chao
Minghua Chao
Ultra-wideband (UWB) is a very promising indoor wireless positioning technology. However, in the harsh and volatile indoor environment, the propagation of UWB signals is vulnerable to non-line-of-sight (NLOS) conditions, and the contaminated range measurements will degrade the accuracy for UWB localization. Therefore, it is necessary to identify LOS/NLOS. Recent studies mainly focus on the identification of UWB signal propagation conditions by using channel impulse response (CIR) or extracted channel statistical features. However, these studies usually only focus on specific indoor environment or stable indoor conditions. In fact, the indoor environment is harsh and changeable. In order to deal with the dynamic and uncertain information of the indoor environments, this paper proposes a robust method to identify LOS/NLOS using fuzzy decision tree (FDT) based on Bayesian optimization. The proposed method first extracts the classification features from the CIRs, and then fuzzifies the features. Finally, it combines Bayesian optimization to construct the FDT, so as to identify the propagation conditions of UWB signals. The experimental results show that the identification accuracy of the proposed method is higher than 90 % in both static and dynamic experiments, and the overall performance is excellent. Compared with other methods, it has certain advantages and robustness.

20 April 2022 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
Jinwei BuJinwei Bu
Jinwei Bu
KYKegen Yu
Kegen Yu
SHShuai Han
Shuai Han
Nijia QianNijia Qian
Nijia Qian
YLYiruo Lin
Yiruo Lin
JWJin Wang
Jin Wang
Reliable rainfall information is very important for hydro climatology research, industrial and agricultural production, construction, and drought and flood prediction. As a new remote sensing technology, Global Navigation Satellite System-reflectometry (GNSS-R) can be used to retrieve geophysical parameters due to its unique advantages of high spatial and temporal resolution, low cost, large coverage, and all-weather operation. Existing studies have confirmed that GNSS-R technology can detect rainfall on the ocean surface, but there is no further study on the magnitude of rainfall intensity (RI). Therefore, this article aims to study the utilization of GNSS-R in the retrieval of sea surface RI based on the observables derived from GNSS-R delay-Doppler maps (DDMs). First, a wavelet denoising method is proposed for DDM denoising to improve the quality of DDM data. Then, three models were developed based on the four GNSS-R observables, which were divided into three categories according to the number of observables. The performance of RI retrieval based on the three models was evaluated against reference data IMERG-F extensively. The experimental results show that under very low wind speed (WS) (< 5 m/s), the following holds. First, the retrieval performance of the single observable model based on leading-edge slope of the normalized integrated delay waveform (LES-NIDW) proposed in the first category of model is the worst [root mean square error (RMSE) $=3.69$ mm/h and correlation coefficient (CC) $=0.71$ ]. However, right edge waveform summation of NIDW (REWS-NIDW) is the best, and RMSE and CC are better than 3.17 mm/h and 0.79, respectively. Second, compared with LES-NIDW observable, the model retrieval accuracy based on DDM Average (DDMA), leading-edge waveform summations of NIDW (LEWS-NIDW), and REWS-NIDW observables is improved by 7.05%, 7.96%, and 12.29%, respectively. Third, when using the combined model, compared with LES-NIDW single observable model, the retrieval accuracy of the model based on two observables (i.e., LES-NIDW and REWS-NIDW) combination and three observables (i.e., LES-NIDW, LEWS-NIDW, and REWS-NIDW) combination is improved by 10.34% and 13.60%, respectively. Therefore, considerable performance gain is obtained.