CYChenguang Yao

Chenguang Yao

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Chenguang Yao has a total of 25 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • A Novel Domain-Adaptive Multiscale CNN-Transformer for Low-Cost MEMS-IMU Denoising

    by Chenguang Yao, Kaifei He, Xuchen Ma, Pengfei Su, Deyu Lin

    Low-cost Micro-Electro-Mechanical Systems Inertial Measurement Units (MEMS-IMUs) are widely employed in intelligent navigation systems. However, they are susceptible to complex noise, leading to a rapid degradation in navigation accuracy. In this study, a MEMS-IMU denoising method based on a Domain-Adaptive Multi-Scale CNN-Transformer (DAMCT) is proposed. First, the proposed method is trained on simulated MEMS-IMU data (source domain) by leveraging multi-scale convolution for spatial feature extraction and a Transformer architecture for temporal sequence modeling. Second, pseudo-labels are generated through inference on real measurement data (target domain), and the model is further fine-tuned using a weakly supervised strategy to enhance its generalization capability in real-world scenarios. This enables effective denoising of low-cost MEMS-IMU data without relying on noise-free labeled datasets. Finally, DAMCT achieves superior denoising performance compared with three state-of-the-art methods, yielding average reductions in position and velocity Root Mean Square Error (RMSE) of 68.0%, 82.9%, 60.1% and 22.4%, 80.6%, 43.9% along the East, North, and Up directions, respectively. These results provide a solid foundation for the future development of intelligent navigation systems.


  • A novel integrated navigation approach with SRCKF based on LSTM signal compensation

    by Chenguang Yao, Shengwei Zhang, Kaifei He, Yue Wang, Kai Ding

    In complex environments, the loss of Global Navigation Satellite System (GNSS) signals can lead to degraded performance or even failure of the GNSS/INS integrated navigation system. Therefore, a compensation method based on Rich Feature Long Short-Term Memory (RF-LSTM) is proposed to assist the GNSS/INS integrated navigation system. Firstly, a feature-rich dataset is constructed using the attitude, position and velocity data of inertial navigation system (INS) combined with the three-axis output data of gyroscope and accelerometer. Then, a time series neural network model based on LSTM is established. During the interruption of GNSS signals, this model can predict GNSS speed increment and position increment by using the input characteristics of the current and historical moments, so as to achieve the purpose of accurate compensation of GNSS signals, so that the integrated navigation system can operate normally without GNSS signals. Finally, through simulation experiments, the proposed method can compensate GNSS position and speed in real time when the GNSS signal is interrupted for as long as 30s, ensuring the continuous normal operation of the integrated navigation system. Compared with the traditional LSTM method, the average velocity measurement accuracy and positioning accuracy of this method are significantly improved. This research provides an effective solution to the navigation problem caused by GNSS signal interruption, and provides a new idea and direction for future research and application.


  • A novel facial age estimation through super-resolution based on deep learning

    by Chenguang Yao, Kaifei He, Kai Ding, Deyu Lin, Jinquan Yang

    Age estimation encounters challenges due to the low resolution (LR) of images captured in real-world scenarios, emphasizing the critical role of image quality in achieving accurate age estimations. Despite efforts to enhance age estimation through LR image training, existing methodologies often overlook the incorporation of super-resolution (SR) as a crucial pre-processing step for LR images. This study introduces an innovative approach that integrates SR of facial images with age estimation methodologies. Specifically, a hierarchical reconstruction network (HRN) structure is employed to train LR images, fully integrating shallow concrete features and deep abstract features. Validation of the training outcomes includes evaluation metrics such as peak signal-to-noise ratio and structural similarity index measurement, showing superior performance compared to several state-of-the-art approaches. Subsequently, an enhanced soft stage regression network is designed for facial age measurement and trained using the SR-IMDB dataset generated by HRN to minimize mean absolute error. This approach achieves a final result of 7.33, indicating an average performance improvement of 10.94% compared to five other state-of-the-art methods. Experimental results highlight the effectiveness of integrating SR into the age estimation process.


  • A Novel Fall Detection Framework With Age Estimation Based on Cloud-Fog Computing Architecture

    by Deyu Lin, Chenguang Yao, Weidong Min, Qing Han, Kaifei He, Ziyuan Yang, Xin Lei, Bin Guo

    Falls are one of the main causes of injuries and even fatalities among the elderly. However, few existing research on fall detection focuses on the supervision of the elderly. Furthermore, the face image of the target usually exhibits low resolution in the surveillance videos, posing a challenge in achieving a balance between the response latency and the computational cost. To this end, a novel fall detection framework with age estimation based on cloud-fog computing architecture is proposed in this paper. Specifically, an optimized Soft Stage Regression-Shallow (SSR-S) network is presented to achieve excellent performance on age estimation for the low-resolution images collected by the edge layer. A fall pre-judgment mechanism and a Shallow-Convolutional Neural Networks (S-CNN) are proposed to make a better judgment on fall behaviors among different age groups at the fog layer and the cloud layer respectively. Besides, an age estimation-based priority algorithm is presented to prioritize the people aged 60 or older in fall detection at the cloud layer, with the aim to make a trade-off between the response latency and the computational overhead. Finally, extensive simulations have been conducted to evaluate the performance of our proposal. Experimental results have shown that the minimum Mean Absolute Error (MAE) of SSR-S reaches 7.59. The fall prejudgment mechanism can achieve 0% miss rate, and the accuracy of S-CNN reaches 90.5%. The detection speed of the overall framework is 17.0 frames per second (FPS).


  • Preparation, structure, and properties of the PTT/SGF composites

    by Mingtao Run, Hongzan Song, Yanping Hao, Xiaomin Hu, Chenguang Yao

    Short glass fiber reinforced poly(trimethylene terephthalate) composites (PTT/SGF) were prepared by twin screw extruder. The structural feature and physical properties of these composites were studied by scanning electron microscope, differential scanning calorimetry, thermalgravimetric analyzer, capillary rheometer, universal tester, etc. The glass fiber was modified by the silane coupling agent before being blended with the polymer. The results suggest that there is strong interaction between SGF and PTT matrix, which leads to an increasing on the tensile strength, Young's modulus, impact strength and thermal stability of the composites with proper contents of SGF. Rheological behavior of the PTT/SGF composites melt is complicated, combining a dilate fluid at lower shear rate and a pseudo‐plastic fluid at higher shear rate. The melt apparent viscosity of composites decreases with increasing SGF content because of the rigid fibers improving the flow of the melt. Moreover, the flow activation energy of the composites suggests that the melt with more SGF has lower sensitivity to the processing temperature. In conclusion, the composite with 10–20 wt% content of SGF has better properties. POLYM. COMPOS., 2009. © 2008 Society of Plastics Engineers


  • Isothermal crystallization kinetics and melting behaviors of nanocomposites of poly(trimethylene terephthalate) filled with nano‐CaCO3

    by Mingtao Run, Chenguang Yao, Yingjin Wang, Jungang Gao

    The isothermal crystallization and subsequent melting behavior of poly(trimethylene terephthalate) (PTT) composites filled with nano‐CaCO3 were investigated at designated temperatures with differential scanning calorimetry. The Avrami equation was used to fit the isothermal crystallization. The Avrami exponents were determined to be 2–3 for the neat PTT and PTT/CaCO3 composites. The particles of nano‐CaCO3, acting as nucleating agents in the composites, accelerated the crystallization rate, with the half‐time of crystallization decreasing or the growth rate constant (involving both nucleation and growth rate parameters) increasing. The crystallization activation energy calculated from the Arrhenius formula was reduced as the nano‐CaCO3 content increased from 0 to 2%, and this suggested that nano‐CaCO3 made the molecular chains of PTT easier to crystallize during the isothermal crystallization process. Subsequent melting scans of the isothermally crystallized composites exhibited triple or double melting endotherms: the greater the content was of nano‐CaCO3, the lower the temperature was of the melting peak. The degree of crystallization deduced from the melt enthalpy of composites with the proper concentration of nano‐CaCO3 was higher than that of pure PTT, but it was lower when the nano‐CaCO3 concentration was more than 2%. The transmission electron microscopy pictures suggested that the dispersion state of nano‐CaCO3 particles in the polymer matrix was even when its concentration was no more than 2%, whereas some agglomeration occurred when its concentration was 4%. Polarized microscopy pictures showed that much smaller or less perfect crystals formed in the composites because of the interaction between the molecular chains and nano‐CaCO3 particles. © 2007 Wiley Periodicals, Inc. J Appl Polym Sci, 2007


  • Melting, crystallization behaviors, and nonisothermal crystallization kinetics of PET/PTT/PBT ternary blends

    by Mingtao Run, Aijun Song, Yingjin Wang, Chenguang Yao

    The melting, crystallization behaviors, and nonisothermal crystallization kinetics of the ternary blends composed of poly(ethylene terephthalate), poly(trimethylene terephthalate) (PTT) and poly(buthylene terephthalate) (PBT) were studied with differential scanning calorimeter (DSC). PBT content in all ternary blends was settled invariably to be one‐third, which improved the melt‐crystallization temperature of the ternary blends. All of the blend compositions in amorphous state were miscible as evidenced by a single, composition‐dependent glass transition temperature (Tg) observed in DSC curves. DSC melting thermograms of different blends showed different multiple melting and crystallization peaks because of their various polymer contents. During melt‐crystallization process, three components in blends crystallized simultaneously to form mixed crystals or separated crystals depending upon their content ratio. The Avrami equation modified by Jeziorny and the Ozawa theory were employed to describe the nonisothermal crystallization process of two selected ternary blends. The results spoke that the Avrami equation was successful in describing the nonisothermal crystallization process of the ternary blends. The values of the t1/2 and the parameters Zc showed that the crystallization rate of the ternary blends with more poly(ethylene terephthalate) content was faster than that with the lesser one at a given cooling rate. The crystal morphology of the five ternary blends investigated by polarized optical microscopy (POM) showed different size and distortional Maltese crosses or light spots when the PTT or poly(ethylene terephthalate) component varied, suggesting that the more the PTT content, the larger crystallites formed in ternary blends. © 2007 Wiley Periodicals, Inc. J Appl Polym Sci, 2007