Mingfeng Yuan

Mingfeng Yuan

Auto-generated profile, based on publicly available information

Mingfeng Yuan

Department of Earth and Space Science, Lassonde school of Engineering, York University, Toronto, ON, Canada

Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

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

Representative Publications

View All Publications
13 December 2021 in IEEE Robotics and Automation Letters
Verified
Journal Article
Closed access
MYMingfeng Yuan
Mingfeng Yuan
JSJinjun Shan
Jinjun Shan
KMKevin Mi
Kevin Mi
This letter presents an approach for implementing game-theoretic decision-making in combination with deep reinforcement learning to allow vehicles to make decisions at an unsignalized intersection by use of 2D Lidar to obtain their observations of the environment. The main novelty of this work is modeling multiple vehicles in a complex interaction scenario simultaneously as decision-makers with conservative, aggressive, and adaptive driving behaviors. The game model allows anticipating reactions of additional vehicles to the movements of the ego-vehicle without using any specific coordination or vehicle-to-vehicle communication. The solution of the game is based on cognitive hierarchy reasoning and it uses a deep reinforcement learning algorithm to obtain a near-optimal policy towards a specific goal in a realistic simulator (ROS-Gazebo). The trained models have been successfully tested on the simulator after training. Experiments show that the performance of the lab cars in the real-world is consistent with it in the simulation environment, which may have great significance to improve the safety of self-driving cars, as well as may reduce their dependence on road tests.


5 July 2023 in IEEE Transactions on Industrial Electronics
Verified
Journal Article
Closed access
MYMingfeng Yuan
Mingfeng Yuan
JSJinjun Shan
Jinjun Shan
HSHunter Schofield
Hunter Schofield
Sharing the road with human drivers requires autonomous vehicles to account for interactions between them. To resolve traffic conflicts in unsignalized intersections, a robust adaptive game-theoretic decision-making algorithm with scalability is proposed based on the receding horizon optimization, level-k game theory, and switching directed graph. A mismatch between the inherent (k-1) assumption of level-k theory and actual driver type may lead to unsafe action selection and reduce driving safety. To handle this problem, in this work, an autonomous vehicle would predict the driver types of surrounding vehicles based on historical interactive behaviors between them and utilize its trust in the driver types to achieve an adaptive driving strategy. Besides, switching interaction graph is incorporated into an adaptive level-k framework for the first time, so as to cutoff the connection between ego vehicle and nearby vehicles that do not affect driving behavior of the former, contributing to reducing the computing complexity. The feasibility, effectiveness, and real-time implementation of the proposed method are validated on both hardware and ROS-Gazebo platform.

30 August 2019 in IEEE Access
Verified
Journal Article
Open access
EDEnzeng Dong
Enzeng Dong
MYMingfeng Yuan
Mingfeng Yuan
FHFangfang HAN
Fangfang HAN
JTJigang Tong
Jigang Tong
Shengzhi DuShengzhi Du
Shengzhi Du
Scholars have done extensive research on dynamic analysis and analog circuits implementation of the classical fractional order chaotic system-Liu System (FOLS). However, they did not rigorously prove the existence of FOLS from the perspective of mathematics. And they also did not effectively design digital circuits to generate signals of the fractional order chaotic systems, especially the 2.7-order system. This paper selects an appropriate Poincaré section where a first return Poincaré map of FOLS was defined. Based on computer-assisted verification method, the conclusion is that the Poincaré map is semi-conjugate to a 2-shift map and the topological entropy of the map is no less than ln 2, which rigorously verifies the existence of chaotic behavior in the 2.7-order Liu system. This proof is necessary before the chaotic system is used for information encryption. The next and most significant task is to build a system model through DSP-Builder software and generate chaotic signals using Field Programmable Gate Array chip. The results of oscilloscope consistent with numerical simulations, which lays the foundation for image and video streaming encryption.

8 May 2023 in Chaos, Solitons & Fractals
Verified
Journal Article
Closed access
YLYue Li
Yue Li
MYMingfeng Yuan
Mingfeng Yuan
ZCZengqiang Chen
Zengqiang Chen

22 August 2023 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
MYMingfeng Yuan
Mingfeng Yuan
JSJinjun Shan
Jinjun Shan
KMKevin Mi
Kevin Mi
Reinforcement learning (RL) is a promising way to achieve human- like autonomous driving (HAD) in complex and dynamic traffic, but faces challenges such as low sample efficiency, partial observability, and sim2real transfer. In light of this, a comprehensive solution for RL-driven HAD is established. First, an efficient training scheme called Deep Recurrent Q-learning from demonstration algorithm (DRQfD) is proposed for lane-changing decision-making to address the low sample efficiency in RL and the poor generalization capability in Imitation Learning (IL). The inherent LSTM structure potentially learns to predict future states of surrounding vehicles, helping to address the partially observable problem in autonomous driving (AD). Second, to reduce the sim2real gap, a twin high-fidelity simulator is built based on ROS-Gazebo for simulating LiDAR sensing, model training, and evaluations. Domain randomization is used to improve the robustness and generalization ability, making it easier for the model to be transferred to real-world scenarios. In addition, for the multi-objective optimization and imbalanced data issues in this scenario, a hierarchical decision-making framework is proposed to decompose the complex decision-making problem into several subtasks, making the driving policies easier to converge. To avoid the excessive dependence of the decision-making module on the output of perception module in modular systems, we train each modularized skill in an end-to-end manner. Moreover, we compare our method with a vanilla RL method to show improvement in learning efficiency. Comparisons between RL-based model and IL baseline in terms of safety, travel efficiency, and human-likeness are also given. To further validate the generalization ability of our model, we test the model on real traffic dataset. Finally, we implement the RL model on physical cars to demonstrate the performance of sim2real transfer.