Article
Engineering, Mechanical
Jia Ye, Zhifei Zhang, Jie Jin, Ruiqi Su, Bo Huang
Summary: The tire-road friction coefficient is crucial for vehicle safety systems. Existing methods have limited accuracy, while the proposed estimation method improves accuracy by adaptively adjusting tire stiffness and accurately identifies tire damage.
NONLINEAR DYNAMICS
(2023)
Article
Automation & Control Systems
Yan Wang, Chen Lv, Yongjun Yan, Pai Peng, Faan Wang, Liwei Xu, Guodong Yin
Summary: This article proposes an integrated scheme for estimating tire-road friction coefficient (TRFC) by combining a strong tracking unscented Kalman filter and an interactive multiple model unscented Kalman filter. Real-time experiments on a mass-produced vehicle demonstrate the feasibility and effectiveness of the proposed method. The results show that the proposed approach has better estimation accuracy than the existing ones under various driving scenarios.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2022)
Article
Mechanics
Elvis Villano, Basilio Lenzo, Aleksandr Sakhnevych
Summary: A novel method for estimating vehicle sideslip angle is proposed in this paper, utilizing a combination of kinematic and dynamic approaches with cross-feedback, and validated on experimental data obtained from different race tracks. The method shows promising results in improving both sideslip angle estimation accuracy and vehicle longitudinal velocity estimation compared to current state-of-the-art techniques.
Article
Engineering, Mechanical
Wei Liu, Xiaowei Wang, Shuisheng Yu, Zhihao Xu
Summary: This paper investigates the tire-road friction coefficient estimation using an adaptive singular value decomposition unscented Kalman filter (ASVD-UKF) with a noise estimator. The ASVD-UKF method significantly reduces the average absolute error compared to the traditional UKF method, improving estimation accuracy. Experimental results show that the proposed ASVD-UKF method is practical and can provide a theoretical basis and experimental foundation for tire-road friction coefficient estimation.
PROCEEDINGS OF THE INSTITUTION OF MECHANICAL ENGINEERS PART D-JOURNAL OF AUTOMOBILE ENGINEERING
(2023)
Article
Engineering, Mechanical
Hongyan Guo, Xu Zhao, Jun Liu, Qikun Dai, Hui Liu, Hong Chen
Summary: An estimation framework that combines vision and vehicle dynamic information is established to accurately obtain the peak tire-road friction coefficient. The framework collects information for the road ahead from an image captured by a camera and uses a lightweight convolutional neural network to identify the road type and its corresponding range of tire-road friction coefficients. An unscented Kalman filter (UKF) method is then used to estimate the tire-road friction coefficient value directly based on the dynamic vehicle states. The results from the road-type recognition and dynamic estimation methods are synchronized, and a confidence-based fusion strategy is proposed to obtain an accurate peak tire-road friction coefficient. Virtual and real vehicle tests confirm the effectiveness of the proposed fusion estimation strategy, which outperforms both general vision-based estimation methods and dynamic-based estimation methods.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2023)
Article
Engineering, Mechanical
Juqi Hu, Subhash Rakheja, Youmin Zhang
Summary: This paper proposes a two-stage TRFC estimation scheme based on longitudinal vehicle dynamics, which controls braking pressure pulses and estimates tire braking force using wheel rotational dynamics to achieve accurate estimation of road friction with minimal interference with vehicle motion.
PROCEEDINGS OF THE INSTITUTION OF MECHANICAL ENGINEERS PART D-JOURNAL OF AUTOMOBILE ENGINEERING
(2021)
Article
Materials Science, Multidisciplinary
Malal Kane, Ebrahim Riahi, Minh-Tan Do
Summary: This paper discusses the modeling of rolling resistance and the analysis of pavement texture effect, with experimental validation showing a good correlation between the model and actual results. The research also highlights the positive correlation between mean profile depth of surfaces and rolling resistance. Furthermore, it suggests the possibility of simplifying the model by neglecting the damping part in the constitutive model of rubber.
Article
Engineering, Civil
Qing Zeng, Xiaoyang Hu, Xiaodong Shi, Yiting Ren, Yuan Li, Zhongdong Duan
Summary: This study proposes a Kalman Filter-based scheme to indirectly evaluate road roughness using measurements of tire pressure. The scheme uses Extended Kalman Filter to calibrate parameters and solves unknowns in the vehicle's state-space equation using Discrete Kalman Filter. The results demonstrate the reliability of the proposed scheme and reveal the potential for better estimation of road roughness at lower running speeds.
INTERNATIONAL JOURNAL OF STRUCTURAL STABILITY AND DYNAMICS
(2022)
Article
Engineering, Mechanical
Chuanwei Zhang, Yansong Feng, Jianlong Wang, Peng Gao, Peilin Qin
Summary: This study proposes a new method for estimating the sideslip angle of vehicles using a radial basis function neural network and an unscented Kalman filter. The experiment shows that this method achieves optimal results in controlling the dynamic behavior of vehicles.
Article
Chemistry, Multidisciplinary
Yafei Li, Yiyong Yang, Xiangyu Wang, Yongtao Zhao, Chengbiao Wang
Summary: This study proposes a state observer based on the EKF method to estimate the vehicle sideslip angle using steering torque instead of steering angle. Transfer functions between the sideslip angle-steering torque and sideslip angle-steering angle are established, and the analysis shows that the steering torque signal reacts more rapidly and directly due to hydraulic pressure. Finally, the proposed method is validated through a simulation hardware-in-the-loop bench test, showing accurate reflection of the sideslip angle with good reliability and effectiveness.
APPLIED SCIENCES-BASEL
(2023)
Article
Engineering, Electrical & Electronic
Lei Zhang, Pengyu Guo, Zhenpo Wang, Xiaolin Ding
Summary: This article proposes a particle filter-based tire-road friction estimation method for four-in-wheel-motor-drive electric vehicles, using dual global positioning system and low-cost inertia measurement units. The method includes independent estimators for straight driving and cornering conditions, and a decision scheme to update the friction estimate based on tire dynamics states and force characteristics. The proposed method shows high accuracy, robustness, and computational efficiency.
IEEE TRANSACTIONS ON TRANSPORTATION ELECTRIFICATION
(2023)
Article
Automation & Control Systems
Amin Habibnejad Korayem, Ehsan Hashemi, Amir Khajepour, Baris Fidan
Summary: This paper introduces a new approach to estimate the lateral tire forces and hitch-forces of a vehicle-trailer system. The method is independent of trailer mass and geometry, and can be used for any ball type trailer. An observer is designed to estimate the hitch-forces and lateral tire forces, and simulations and experimental tests are used to validate the method, showing good accuracy.
JOURNAL OF DYNAMIC SYSTEMS MEASUREMENT AND CONTROL-TRANSACTIONS OF THE ASME
(2022)
Article
Automation & Control Systems
Yan Wang, Keke Geng, Liwei Xu, Yaping Ren, Haoxuan Dong, Guodong Yin
Summary: This article proposes a novel method, the fuzzy adaptive robust cubature Kalman filter (FARCKF), to accurately estimate the sideslip angle and tire cornering stiffness. The model parameters are dynamically updated and a fuzzy system is used to improve estimation accuracy.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2022)
Article
Engineering, Electrical & Electronic
Zhenqiang Quan, Bo Li, Shaoyi Bei, Xiaoqiang Sun, Nan Xu, Tianli Gu
Summary: This paper proposes a tire-road friction coefficient estimation method based on intelligent tire technology. Through finite element analysis and control variable method, the influence of sideslip angle on the voltage signal of each piezoelectric film under the rolling state of tire is analyzed, and the influence of load, tire pressure, vehicle speed, and slip ratio on the voltage signal of each piezoelectric film is also analyzed. Based on signal response analysis, prediction models are built and input into the brush tire model to solve the tire-road friction coefficient. The result shows that the estimation error percentage with genetic algorithm optimization is 5.14%, indicating the practicality of the friction coefficient estimation method.
SENSORS AND ACTUATORS A-PHYSICAL
(2023)
Article
Engineering, Multidisciplinary
Masahiro Higuchi, Yosuke Suzuki, Tomohiko Sasano, Hiroshi Tachiya
Summary: This study investigates a method for measuring road friction coefficients using strains on the sidewalls of tires. The proposed method is confirmed to be able to accurately measure the load acting on a tire and friction coefficient of the tire grounding surface at low speeds and under full-slip conditions.
PRECISION ENGINEERING-JOURNAL OF THE INTERNATIONAL SOCIETIES FOR PRECISION ENGINEERING AND NANOTECHNOLOGY
(2023)
Article
Engineering, Civil
Yaoyu Chen, Guofa Li, Shen Li, Wenjun Wang, Shengbo Eben Li, Bo Cheng
Summary: This study proposes an unsupervised method to extract and discover the behavioral patterns of lane change maneuvers, aiming to explore the composed behavioral patterns during lane changes. The method involves two phases: segmentation of lane change sequences into blocks and clustering of these blocks to find corresponding behavioral patterns. The results show that the method effectively mines descriptive behavioral patterns from lane change data, providing a promising data mining solution for deep understanding of driver lane change behaviors.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2022)
Article
Engineering, Civil
Ziyu Lin, Jun Ma, Jingliang Duan, Shengbo Eben Li, Haitong Ma, Bo Cheng, Tong Heng Lee
Summary: In the field of autonomous driving, it is challenging to solve the motion planning problem due to the nonlinearity of the vehicle model and the complexity of driving scenarios. To address this, a framework of integrated decision and control is investigated, using static path planning and an innovative constrained finite-horizon approximate dynamic programming algorithm. This algorithm effectively handles changing driving environments with varying surrounding vehicles and reduces computational loads through offline training and online execution.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Automation & Control Systems
Yang Guan, Yangang Ren, Qi Sun, Shengbo Eben Li, Haitong Ma, Jingliang Duan, Yifan Dai, Bo Cheng
Summary: This article presents an interpretable and computationally efficient framework called integrated decision and control (IDC) for automated vehicles. It decomposes the driving task into static path planning and dynamic optimal tracking that are structured hierarchically. The framework has been verified in both simulations and the real world, showing improved online computing efficiency, driving performance, as well as interpretability and adaptability in different driving scenarios and tasks.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Robotics
Yujie Yang, Yuxuan Jiang, Yichen Liu, Jianyu Chen, Shengbo Eben Li
Summary: This letter proposes a model-free safe reinforcement learning algorithm that achieves near-zero constraint violations with high rewards. By jointly learning a policy and a neural barrier certificate under stepwise state constraint setting, our algorithm balances the bias and variance of the barrier certificate and enhances both safety and performance.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Computer Science, Artificial Intelligence
Yongxin Zhu, Yongfu Li, Hao Zhu, Wei Hua, Gang Huang, Shuyou Yu, Shengbo Eben Li, Xinbo Gao
Summary: This article presents a jointly distributed adaptive sliding-mode controller (DASMC) and disturbance observer (DO) design for a heterogeneous platoon of connected vehicles (CVs) subject to disturbance and acceleration failure of neighboring vehicles under different common communication topologies. The effects of disturbance and acceleration failure are mitigated by jointly estimating the lumped disturbance arising from vehicle heterogeneity, nonlinearity, and neighboring vehicle acceleration using a DO. The DASMC, with an adaptive law based on defined lumped errors, is designed according to the vehicle dynamic model, and its stability is analyzed using the Lyapunov technique. The string stability of the CV platoon is also proven. The effectiveness of the proposed control scheme is demonstrated through simulations and experiments, comparing it with existing methods.
IEEE TRANSACTIONS ON INTELLIGENT VEHICLES
(2023)
Article
Computer Science, Artificial Intelligence
Jingliang Duan, Jie Li, Qiang Ge, Shengbo Eben Li, Monimoy Bujarbaruah, Fei Ma, Dezhao Zhang
Summary: This paper introduces the Relaxed Continuous-Time Actor-critic (RCTAC) algorithm, which aims to find the nearly optimal policy for nonlinear continuous-time systems. RCTAC has advantages over existing adaptive dynamic programming algorithms as it does not require specific conditions for convergence. The algorithm consists of a warm-up phase and a generalized policy iteration phase, where admissibility and convergence are achieved through minimizing the square of the Hamiltonian and relaxing the update termination conditions.
IEEE TRANSACTIONS ON INTELLIGENT VEHICLES
(2023)
Article
Computer Science, Artificial Intelligence
Feng Gao, Yu Han, Shengbo Eben Li, Shaobing Xu, Dongfang Dang
Summary: This paper presents an NMPC based motion planner for automated vehicles and introduces two techniques, adaptive Lagrange discretization and hybrid obstacle avoidance constraints, to accelerate numerical optimization. The techniques reduce optimization variables and simplify non-convex constraints. The Lagrange interpolation is adopted to ensure accuracy with fewer discretization points, and an adaptive strategy adjusts the order of Lagrange polynomials based on the discretization error. A hybrid strategy combines elliptic and linear time-varying methods to construct obstacle avoidance constraints. Comparative simulations and tests show that these techniques improve accuracy and efficiency by 74% and 60%, respectively.
IEEE TRANSACTIONS ON INTELLIGENT VEHICLES
(2023)
Article
Computer Science, Artificial Intelligence
Guofa Li, Yifan Qiu, Yifan Yang, Zhenning Li, Shen Li, Wenbo Chu, Paul Green, Shengbo Eben Li
Summary: End-to-end approaches are a promising solution for AV decision-making, but their deployment is often hindered by high computational burden. To address this, we propose a lightweight transformer-based end-to-end model with risk awareness for AV decision-making. We introduce a lightweight network with depth-wise separable convolution and transformer modules to extract image semantics from trajectory data. We then assess driving risk using a probabilistic model with position uncertainty and integrate it into deep reinforcement learning to find strategies with minimum expected risk. The proposed method is evaluated in three lane change scenarios to validate its superiority.
IEEE TRANSACTIONS ON INTELLIGENT VEHICLES
(2023)
Article
Computer Science, Artificial Intelligence
Lian Hou, Shengbo Eben Li, Bo Yang, Zheng Wang, Kimihiko Nakano
Summary: This paper presents a unified graphical representation method that takes into account the varying numbers and types of vehicles, various road structures, and traffic rules to improve trajectory prediction for autonomous vehicles in highway traffic scenarios. By integrating the constraints from vehicles and the collision risk implied behind road structures and traffic rules, this method provides a quantitative way to better utilize the influences of these elements on trajectory prediction.
IEEE TRANSACTIONS ON INTELLIGENT VEHICLES
(2023)
Article
Engineering, Civil
Ziqing Gu, Lingping Gao, Haitong Ma, Shengbo Eben Li, Sifa Zheng, Wei Jing, Junbo Chen
Summary: This paper proposes a state-based safety enhancement method for autonomous driving through direct hierarchical reinforcement learning. By integrating a dynamic module and generating future goals considering safety, temporal-spatial continuity, and dynamic feasibility, the proposed method shows better training performance, higher driving safety in interactive scenes, more decision intelligence in traffic congestions, and better economic driving ability on roads with changing slopes.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Ziyu Lin, Jingliang Duan, Shengbo Eben Li, Haitong Ma, Jie Li, Jianyu Chen, Bo Cheng, Jun Ma
Summary: The research addresses the challenge of solving the finite-horizon HJB equation, proposes a new algorithm, and validates its effectiveness through simulations.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Haotian Zheng, Chaoyi Chen, Shuai Li, Sifa Zheng, Shengbo Eben Li, Qing Xu, Jianqiang Wang
Summary: By using magnitude regularization techniques, this study proposes a method to enhance the efficiency of safe controllers by reducing conservativeness within the energy function while maintaining promising demonstrable safety guarantees.
IEEE OPEN JOURNAL OF INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Automation & Control Systems
Jingliang Duan, Jie Li, Xuyang Chen, Kai Zhao, Shengbo Eben Li, Lin Zhao
Summary: This article investigates the optimization landscape of policy gradient methods for static output feedback control in discrete-time LTI systems. By analyzing the crucial properties of the SOF cost, new findings on convergence and convergence rate for three policy gradient methods are derived, along with a proof of linear convergence for the vanilla policy gradient method when initialized near local minima. Numerical examples validate the theoretical findings. These results characterize the performance of gradient descent for optimizing the SOF problem and provide insights into the effectiveness of general policy gradient methods within the realm of reinforcement learning.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Xiaoyu Chen, Yao Mu, Ping Luo, Shengbo Eben Li, Jianyu Chen
Summary: Partially Observable Markov Decision Process (POMDP) is a generic framework for modeling real world sequential decision making processes, where the main challenge lies in accurately obtaining the belief state. This paper proposes a Flow-based Recurrent Belief State model that incorporates normalizing flows to learn general continuous belief states for POMDPs, and demonstrates its effectiveness in improving performance.
INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 162
(2022)
Article
Computer Science, Artificial Intelligence
Baiyu Peng, Jingliang Duan, Jianyu Chen, Shengbo Eben Li, Genjin Xie, Congsheng Zhang, Yang Guan, Yao Mu, Enxin Sun
Summary: In this article, the authors propose a separated proportional-integral Lagrangian (SPIL) algorithm to address the shortcomings of existing chance-constrained RL methods. The SPIL algorithm combines the advantages of penalty methods and Lagrangian methods while limiting the integral value to enhance safety and convergence. Experimental results demonstrate the effectiveness of the proposed method in car-following simulation and mobile robot navigation tasks.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)