Hewei Li

Hewei Li

Auto-generated profile, based on publicly available information

Hewei Li

School of Electrical and Electronic Engineering, Changchun University of Technology, 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
3 March 2021 in Asian Journal of Control
Verified
Journal Article
Closed access
PDF file
HLHewei Li
Hewei Li
BZBin Zhao
Bin Zhao
WYWeiBo Yu
WeiBo Yu
NXNan Xu
Nan Xu
KGKongHui Guo
KongHui Guo

20 July 2023 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
HLHewei Li
Hewei Li
KLKeping Liu
Keping Liu
BZBin Zhao
Bin Zhao
NXNan Xu
Nan Xu
YHYANJUN HUANG
YANJUN HUANG
YYYuming Yin
Yuming Yin
Since the coupling of tire forces, the single longitudinal or lateral vehicle analysis method cannot adequately characterize the vehicle stability, especially the large tire slip ratio under the combined-slip conditions reduces the effective usage rate of tire force, which will increase the risk of vehicle instability. Thus, the effective quasi-usage rate of tire force based on the Monte Carlo method is proposed to solve the constraint of tire slip ratio that can maximize the quasi-usage rate. Firstly, the effective quasi-usage rate $P_{e\!f\!f}$ is proposed, which is a better comprehensive evaluation standard for vehicle stability and tire force usage rate under the combined-slip conditions by employing the Monte Carlo method and $\beta -\dot{\beta }-V_{x}$ phase space; Further, the trend of $P_{e\!f\!f}$ with tire slip ratio is divided into 3 characteristic points and 2 variation stages, the constraint of tire slip ratio is defined by analyzing the characteristics of tire force at different points and stages; Then, the neural network is used to train the data set of constraint to generate the dynamic constraint model, and the objective function of model predictive control strategy considering the constraint of tire slip ratio is established to verify the controller under the combined-slip conditions. The acceleration and deceleration-in-turn are conducted by 4WIMD-EVs on the Hardware-In-the-Loop (HIL) simulation platform. The results show that when the tire slip ratio is controlled under the dynamic constraint proposed in this paper, the vehicle has stronger stability and c.g. acceleration, which also proves the effectiveness of the proposed effective quasi-usage rate.

1 March 2024
Verified
Conference Paper
Open access
QWQingxuan Wang
Qingxuan Wang
BZBin Zhao
Bin Zhao
YGYongle Gao
Yongle Gao
HLHewei Li
Hewei Li
ZWZe Wang
Ze Wang
Data is the core part of the training model. In order to make the trained model have stronger performance, a substantial volume of data is frequently necessary. However, data collection and labeling are difficult, so the purpose of data enhancement technology is to expand the data set by transforming and learning existing data to generate new data samples. The current mainstream data enhancement model is Generative Adversarial Networks (GAN). However, this model is prone to overfitting problems during the image generation process, resulting in insufficient richness and low authenticity of the generated images. Therefore, this article proposes an Adaptive normalization data enhancement algorithm based on DCGAN. This model improves the richness of generated images by introducing a deep convolutional neural network, and equips each convolution block with adaptive normalization to ensure that the model will not overfit during the training process. We trained this model on the CIFAR100 dataset, and the experimental results show that the robustness of the model and the realism of the generated images are better than DCGAN.