Yali Yuan

Dr. Yali Yuan


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
28 May 2020 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
TWTong Wu
Tong Wu
PZPan Zhou
Pan Zhou
KLKai Liu
Kai Liu
YYYali Yuan
Yali Yuan
Xiumin WangXiumin Wang
Xiumin Wang
DWDapeng Oliver Wu
Dapeng Oliver Wu
As urban traffic condition is diverse and complicated, applying reinforcement learning to reduce traffic congestion becomes one of the hot and promising topics. Especially, how to coordinate the traffic light controllers of multiple intersections is a key challenge for multi-agent reinforcement learning (MARL). Most existing MARL studies are based on traditional Q-learning, but unstable environment leads to poor learning in the complicated and dynamic traffic scenarios. In this paper, we propose a novel multi-agent recurrent deep deterministic policy gradient (MARDDPG) algorithm based on Deterministic Policy Gradient algorithm for traffic light control (TLC) in vehiclar networks. Specifically, the centralized learning in each critic network enables each agent to estimate the policies of other agents in the decision-making process and each agent can coordinate with each other, alleviating the problem of poor learning performance caused by environmental instability. The decentralized execution enables each agent to make decisions independently. We share parameter in actor networks to speed up training process and reduce the memory footprint. The addition of LSTM is beneficial to alleviate the instability of the environment caused by partial observable state. We utilize surveillance cameras and vehicular networks to collect status information for each intersection. Unlike previous work, we have not only considered the vehicle but also considered the pedestrians waiting to pass through the intersection. Moreover, we also set different priorities for buses and ordinary vehicles. The experimental results in a vehicular network show that our method can run stably in various scenarios and coordinate multiple intersections, which significantly reduces vehicle congestion and pedestrian congestion.

15 May 2018 in IEEE Access
Verified
Journal Article
Open access
YYYali Yuan
Yali Yuan
LHLiuwei Huo
Liuwei Huo
ZWZhixiao Wang
Zhixiao Wang
DHDieter Hogrefe
Dieter Hogrefe
For location-aware applications in Wireless Sensor Networks (WSNs), it is important to ensure that sensor nodes can get correct locations in a hostile WSNs. Sybil attacks, which are vital threats in WSNs, especially in the distributed WSNs. They can forge one or multiple identities to decrease the localization accuracy, or sometimes to collapse the whole localization systems. In this paper, a novel lightweight SF-APIT algorithm is presented to solve the problem of sybil attacks in APIT localization scheme, which is a popular range-free method and performs at individual node in a purely distributed fashion. The proposed SF-APIT scheme requires minimal overhead for wireless devices and works well based on the Received Signal Strength (RSS). Simulations demonstrate that SF-APIT is an effective scheme in detecting and defending against sybil attacks with a high detection rate in distributed wireless localization schemes.

14 February 2020 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
YYYali Yuan
Yali Yuan
RTRobert Tasik
Robert Tasik
SASripriya Srikant Adhatarao
Sripriya Srikant Adhatarao
Yachao YuanYachao Yuan
Yachao Yuan
Zheli LiuZheli Liu
Zheli Liu
XFXiaoming Fu
Xiaoming Fu
With the rapid development of autonomous driving, collision avoidance has attracted attention from both academia and industry. Many collision avoidance strategies have emerged in recent years, but the dynamic and complex nature of driving environment poses a challenge to develop robust collision avoidance algorithms. Therefore, in this paper, we propose a decentralized framework named RACE: Reinforced Cooperative Autonomous Vehicle Collision AvoidancE. Leveraging a hierarchical architecture we develop an algorithm named Co-DDPG to efficiently train autonomous vehicles. Through a security abiding channel, the autonomous vehicles distribute their driving policies. We use the relative distances obtained by the opponent sensors to build the VANET instead of locations, which ensures the vehicle's location privacy. With a leader-follower architecture and parameter distribution, RACE accelerates the learning of optimal policies and efficiently utilizes the remaining resources. We implement the RACE framework in the widely used TORCS simulator and conduct various experiments to measure the performance of RACE. Evaluations show that RACE quickly learns optimal driving policies and effectively avoids collisions. Moreover, RACE also scales smoothly with varying number of participating vehicles. We further compared RACE with existing autonomous driving systems and show that RACE outperforms them by experiencing 65% less collisions in the training process and exhibits improved performance under varying vehicle density.

14 December 2018 in IEEE Transactions on Vehicular Technology
Verified
Journal Article
Closed access
YYYali Yuan
Yali Yuan
CLChencheng Liang
Chencheng Liang
MKMegumi Kaneko
Megumi Kaneko
XCXu Chen
Xu Chen
DHDieter Hogrefe
Dieter Hogrefe
The characteristics of mobile Underwater Sensor Networks (UWSNs), such as low communication bandwidth, large propagation delay, and sparse deployment, pose challenging issues for successful localization of sensor nodes. In addition, sensor nodes in UWSNs are usually powered by batteries whose replacements introduce high cost and complexity. Thus, the critical problem in UWSNs is to enable each sensor node to find enough anchor nodes in order to localize itself, with minimum energy costs. In this paper, an Energy-Efficient Localization Algorithm (EELA) is proposed to analyze the decentralized interactions among sensor nodes and anchor nodes. A Single-Leader-Multi-Follower Stackelberg game is utilized to formulate the topology control problem of sensor nodes and anchor nodes by exploiting their available communication opportunities. In this game, the sensor node acts as a leader taking into account factors such as 'two-hop' anchor nodes and energy consumption, while anchor nodes act as multiple followers, considering their ability to localize sensor nodes and their energy consumption. We prove that both players select best responses and reach a socially optimal Stackelberg Nash Equilibrium. Simulation results demonstrate that the proposed EELA improves the performance of localization in UWSNs significantly, and in particular the energy cost of sensor nodes. Compared to the baseline schemes, the energy consumption per node is about 48% lower in EELA, while providing a desirable localization coverage, under reasonable error and delay.

18 September 2020 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
Yachao YuanYachao Yuan
Yachao Yuan
SISaiful Islam
Saiful Islam
YYYali Yuan
Yali Yuan
SWShengjin Wang
Shengjin Wang
Thar BakerThar Baker
Thar Baker
LKLutz Maria Kolbe
Lutz Maria Kolbe
Road damages have caused numerous fatalities, thus the study of road damage detection, especially hazardous road damage detection and warning is critical for traffic safety. Existing road damage detection systems mainly process data at cloud, which suffers from a high latency caused by long-distance. Meanwhile, supervised machine learning algorithms are usually used in these systems requiring large precisely labeled data sets to achieve a good performance. In this article, we propose EcRD: an edge-cloud-based road damage detection and warning framework, that leverages the fast-responding advantage of edge and the large storage and computation resources advantages of cloud. There are three main contributions in this article: we first propose a simple yet efficient road segmentation algorithm to enable fast and accurate road area detection. Then, a light-weighted road damage detector is developed based on gray level co-occurrence matrix features at edge for rapid hazardous road damage detection and warning. Furthermore, a multitypes road damage detection model is introduced for long-term road management at cloud, embedded with a novel image generator based on cycle-consistent adversarial networks which automatically generates images with labels to further improve road damage detection accuracy. By comparing with the state-of-the-art, we demonstrate that the proposed EcRD can accurately detect both hazardous road damages at edge and multitypes road damages at cloud. Besides, it is around 579 times faster than cloud-based approaches without affecting users' experience and requiring very low storage and labeling cost.

1 December 2021 in Future Generation Computer Systems
Verified
Journal Article
Closed access
Yachao YuanYachao Yuan
Yachao Yuan
YYYali Yuan
Yali Yuan
Thar BakerThar Baker
Thar Baker
LKLutz Maria Kolbe
Lutz Maria Kolbe
DHDieter Hogrefe
Dieter Hogrefe