Zhaoxuan Zhu

University of Electronic Science and Technology of China

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Trends

Author Order

Document Type

Co-Authors

Marcello Canova

The Ohio State University

Houjun Wang

University of Electronic Science and Technology of China

Shobhit Gupta

The Ohio State University

Nicola Pivaro

The Ohio State University

Zijun Wang

Nanjing University of Aeronautics and Astronautics

Abhishek Gupta

The Ohio State University

Lin Cai

Huazhong University of Science and Technology

Shreshta Rajakumar Deshpande

Southwest Research Institute

Stephanie Stockar

The Ohio State University

Jingmin Dai

Harbin Institute of Technology

Cooperative Institutions

University of Electronic Science and Technology of China

The Ohio State University

Chinese Academy of Sciences

Peking University

Huazhong University of Science and Technology

Nanjing University of Aeronautics and Astronautics

Shanghai Jiao Tong University

Chinese Academy of Medical Sciences & Peking Union Medical College

Jilin University

Soochow University

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Field

Safe Model-based Off-policy Reinforcement Learning for Eco-Driving in Connected and Automated Hybrid Electric Vehicles

arXiv (Cornell University) (2021)

Zhaoxuan Zhu Nicola Pivaro Shobhit Gupta Abhishek Gupta Marcello Canova

Connected and Automated Hybrid Electric Vehicles have the potential to reduce fuel consumption and travel time in real-world driving conditions. The eco-driving problem seeks to design optimal speed and power usage profiles based upon look-ahead information from connectivity and advanced mapping features. Recently, Deep Reinforcement Learning (DRL) has been applied to the eco-driving problem. While the previous studies synthesize simulators and model-free DRL to reduce online computation, this work proposes a Safe Off-policy Model-Based Reinforcement Learning algorithm for the eco-driving problem. The advantages over the existing literature are three-fold. First, the combination of off-policy learning and the use of a physics-based model improves the sample efficiency. Second, the training does not require any extrinsic rewarding mechanism for constraint satisfaction. Third, the feasibility of trajectory is guaranteed by using a safe set approximated by deep generative models. The performance of the proposed method is benchmarked against a baseline controller representing human drivers, a previously designed model-free DRL strategy, and the wait-and-see optimal solution. In simulation, the proposed algorithm leads to a policy with a higher average speed and a better fuel economy compared to the model-free agent. Compared to the baseline controller, the learned strategy reduces the fuel consumption by more than 21\% while keeping the average speed comparable.

Baseline (sea)

10.48550/arxiv.2105.11640

Cite

Citations (1)

Reconstruction of sampling in shift-invariant space using generalized inverse

Journal of Systems Engineering and Electronics (2011)

Zhaoxuan Zhu

A method that attempts to recover signal using generalized inverse theory is presented to obtain a good approximation of the signal in reconstruction space from its generalized samples. The proposed approaches differ with the assumptions on reconstruction space. If the reconstruction space satisfies one-to-one relationship between the samples and the reconstruction model, then we propose a method, which achieves consistent signal reconstruction. At the same time, when the number of samples is more than the number of reconstruction functions, the minimal-norm reconstruction signal can be obtained. Finally, it is demonstrated that the minimal-norm reconstruction can outperform consistent signal reconstruction in both theory and simulations for the problem.

10.3969/j.issn.1004-4132.2011.02.004

Cite

Citations (0)

A Physics-Based Three-Way Catalytic Converter Model for Real-Time Prediction of Temperature Distribution

IFAC-PapersOnLine (2018)

Zhaoxuan Zhu Marcello Canova Shawn Midlam‐Mohler

Thermal management of aftertreatment devices is a critical requirement to comply with the stringent emission standards. A recent challenge in aftertreatment control and diagnostics is the need to monitor the temperature distribution along the entire length of the catalyst, rather than relying on a single-point measurement. To this extent, this work focuses on the preliminary development of a control-oriented, physics-based Three-way Catalytic Converter (TWC) model for the purpose of real-time thermal monitoring. Starting from the governing equations for the TWC thermal model in Partial Differential Equation (PDE) form, a model order reduction technique that combines Proper Orthogonal Decomposition and Collocation is proposed. The reduced order model executes much faster than using standard numerical methods, and meanwhile provides comparable accuracy. The model is also validated against GT-Power in the ability to predict the temperature distribution during cold-start conditions.

Collocation (remote sensing)

Catalytic converter

Orthogonal collocation

Model order reduction

10.1016/j.ifacol.2018.10.031

Cite

Citations (3)

A GPU Implementation of a Look-Ahead Optimal Controller for Eco-Driving Based on Dynamic Programming

2022 European Control Conference (ECC) (2021)

Zhaoxuan Zhu Shobhit Gupta Nicola Pivaro Shreshta Rajakumar Deshpande Marcello Canova

Predictive energy management of Connected and Automated Vehicles (CAVs), in particular those with multiple power sources, has the potential to significantly improve energy savings in real-world driving conditions. In particular, the eco-driving problem seeks to design optimal speed and power usage profiles based upon available information from connectivity and advanced mapping features to minimize the fuel consumption between two designated locations.In this work, the eco-driving problem is formulated as a three-state receding horizon optimal control problem and solved via Dynamic Programming (DP). The optimal solution, in terms of vehicle speed and battery State of Charge (SoC) trajectories, allows a connected and automated hybrid electric vehicle to intelligently pass the signalized intersections and minimize fuel consumption over a prescribed route. To enable real-time implementation, a parallel architecture of DP is proposed for an NVIDIA GPU with CUDA programming. Simulation results indicate that the proposed optimal controller delivers more than 15% fuel economy benefits compared to a baseline control strategy and that the solver time can be reduced by more than 90% by the parallel implementation when compared to a serial implementation.

Solver

Model Predictive Control

State of charge

10.23919/ecc54610.2021.9655197

Cite

Citations (18)

Model Order Reduction of the Doyle-Fuller-Newman Model via Proper Orthogonal Decomposition and Optimal Collocation

2021 IEEE Vehicle Power and Propulsion Conference (VPPC) (2023)

Gianluca Manduca Zhaoxuan Zhu Polina Brodsky Ringler Guodong Fan Marcello Canova

Adapting physics-based models for digital twin frameworks, estimation and control of Lithium ion batteries requires a significant reduction of the computational complexity of electrochemical models without sacrificing accuracy. In this study, we propose a model order reduction technique for the Doyle-Fuller-Newman model of a battery cell that integrates Proper Orthogonal Decomposition with an optimized orthogonal collocation method. Simulation results indicate that the reduced-order model is faster than the full-order model, reducing computation time by up to 99% while maintaining negligible voltage discrepancy with a root mean square error (RMSE) of less than 0.005V. The proposed approach provides a powerful framework for improving the computation time of complex, physics-based models based on coupled partial differential equations, leading to more accurate predictions and better design and optimization of complex components and processes.

Model order reduction

Orthogonal collocation

Collocation (remote sensing)

Proper orthogonal decomposition

10.1109/vppc60535.2023.10403283

Cite

Citations (0)

Research on CFRP Defects Recognition and Localization Based on Metamaterial Sensors

Symmetry (2024)

Zhaoxuan Zhu Rui Han

In the paper, the concept of symmetry is utilized to detect internal defects in Carbon fiber reinforced polymer (CFRP), that is, the reconstruction and localization methods for internal defects in CFRP are symmetrical. CFRP is widely used in industrial, biological and other fields. When there are defects inside the composite materials, its dielectric constant, magnetic permeability, etc. change. Therefore, metamaterial sensors are widely used in non-destructive testing of CFRP Defects. This paper proposes a defect identification and location method based on principal component analysis (PCA) and support vector machine (SVM). The trained model is used to classify the dimensionally reduced data, and the reconstructed defect binary image is obtained. Simulation and physical experiment results show that the method used in this article can effectively identify and locate defects in carbon fiber composite materials.

10.3390/sym16121706

Cite

Citations (0)

Road to chaos for a soft spring system under weak periodic disturbance

Applied Mathematics and Mechanics (1986)

Zhengrong Liu Weiguo Yao Zhaoxuan Zhu

Subharmonic

Constant (computer programming)

10.1007/bf01897053

Cite

Citations (6)

Controlling Hamiltonian chaos by medium perturbation in periodically driven systems

Physical review. E, Statistical physics, plasmas, fluids, and related interdisciplinary topics (1998)

Zhihua Wu Zhaoxuan Zhu ZHANG CHENG-FU

By employing another external field with intensity not larger than 60% of the original driving force, the Hamiltonian chaos in a driven Morse oscillator could be controlled in the sense that the bound state regions can be changed at will. Based on our understanding about the complex dynamics of a periodically driven Hamiltonian system with one degree of freedom by two external fields, an idea of decreasing bound regions is proposed. A formula for island width, being helpful for selecting a proper controlling field, is derived. The mechanism in increasing bound regions extracted from a large number of numerical experiments is still not so clear at present, but we have gotten a good grasp of the roles of controlling parameters with suppression of Hamiltonian chaos. Though no small perturbations with this method can have a satisfactory effect, as previous studies, some remarks about the magnitude of perturbation are presented.

Hamiltonian (control theory)

10.1103/physreve.57.366

Cite

Citations (18)

Research on CFRP Defects Recognition and Localization Based on Metamaterial Sensors

Zhaoxuan Zhu Rui Han

The wide application of carbon-fiber-reinforced polymer(CFRP) has brought with it an urgent demand for their reliability and stability. However, considering factors such as the manufacturing process, material characteristics, and use environment, defects may exist in composite materials, posing a threat to engineering safety. Metamaterial sensors are widely used in non-destructive testing of microwave sensors due to their high resolution and high sensitivity. This paper proposes a defect identification and location method based on principal component analysis and support vector machine. The trained model is used to classify the dimensionally reduced data, and the reconstructed defect binary image is obtained. Simulation and physical experiment results show that the method used in this article can effectively identify and locate defects in carbon fiber composite materials.

10.20944/preprints202412.0093.v1

Cite

Citations (0)

Angle of Arrival Estimation with Transformer: A Sparse and Gridless Method with Zero-Shot Capability

arXiv (Cornell University) (2024)

Zhaoxuan Zhu Chulong Chen Bo Yang

Automotive Multiple-Input Multiple-Output (MIMO) radars have gained significant traction in Advanced Driver Assistance Systems (ADAS) and Autonomous Vehicles (AV) due to their cost-effectiveness, resilience to challenging operating conditions, and extended detection range. To fully leverage the advantages of MIMO radars, it is crucial to develop an Angle of Arrival (AOA) algorithm that delivers high performance with reasonable computational workload. This work introduces AAETR (Angle of Arrival Estimation with TRansformer) for high performance gridless AOA estimation. Comprehensive evaluations across various signal-to-noise ratios (SNRs) and multi-target scenarios demonstrate AAETR's superior performance compared to super resolution AOA algorithms such as Iterative Adaptive Approach (IAA). The proposed architecture features efficient, scalable, sparse and gridless angle-finding capability, overcoming the issues of high computational cost and straddling loss in SNR associated with grid-based IAA. AAETR requires fewer tunable hyper-parameters and is end-to-end trainable in a deep learning radar perception pipeline. When trained on large-scale simulated datasets then evaluated on real dataset, AAETR exhibits remarkable zero-shot sim-to-real transferability and emergent sidelobe suppression capability. This highlights the effectiveness of the proposed approach and its potential as a drop-in module in practical systems.

Zero (linguistics)

10.48550/arxiv.2408.09362

Cite

Citations (0)