Xiaoyi Fan

Xiaoyi Fan

Auto-generated profile, based on publicly available information

Xiaoyi Fan

Tsinghua University

Research Focus & Metrics

Academic Career Length

12Years (2015-Present)

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

Representative Publications

View All Publications
27 February 2018 in IEEE/ACM Transactions on Networking
Verified
Journal Article
Closed access
wei gongwei gong
wei gong
HLHaoxiang Liu
Haoxiang Liu
JLJiangchuan Liu
Jiangchuan Liu
XFXiaoyi Fan
Xiaoyi Fan
KLKebin Liu
Kebin Liu
XJXiaoyu Ji
Xiaoyu Ji
Backscatter communication networks receive much attention recently due to the small size and low power of backscatter nodes. As backscatter communication is often influenced by the dynamic wireless channel quality, rate adaptation becomes necessary. Most existing approaches share a common drawback: they fail to take both spatial and frequency diversity into consideration at the same time. Consequently, the transmission rate may be improperly selected, resulting in low network throughput. In this paper, we propose a channel-aware rate adaptation framework (CARA) for backscatter networks. CARA incorporates three essential modules, a lightweight channel probing scheme that differentiates collisions from packet losses, a burstiness-aware channel selection mechanism benefiting as many backscatter nodes as possible, a rate selection method choosing the optimal rate, and a mobility detection that discovers location changes. We implement CARA on commercial readers, and the experiment results show that CARA achieves up to 4× goodput gain compared with the state-of-the-art rate adaptation scheme.

1 June 2018 in IEEE Transactions on Mobile Computing
Verified
Journal Article
Closed access
FWFangxin Wang
Fangxin Wang
XFXiaoyi Fan
Xiaoyi Fan
FWFeng Wang
Feng Wang
JLJiangchuan Liu
Jiangchuan Liu
Base stations have been widely deployed to satisfy the service coverage and explosive demand increase in today's cellular networks. Battery groups are installed as backup power in most of base stations in case of power outages. The limited numbers and capacities of batteries, however, can hardly sustain a long power outage. As a result, the service interruption occurs along with an increasing maintenance cost. Meanwhile, a deep discharge of battery in such case can also accelerate the battery degradation and eventually contribute to a higher battery replacement cost. In this paper, we closely examine the base station features and backup battery features from a 1.5-year dataset of a major cellular service provider, including 4,206 base stations and more than 1.5 billion records. We build up a deep learning based model to estimate the remaining lifetime of backup batteries. We then develop BatAlloc, a battery allocation framework to address the mismatch between the battery supporting ability and diverse power outage incidents. We present an effective solution that minimizes both service interruption time and the overall cost. Our real world trace-driven experiments show that BatAlloc cuts down the average service interruption time from 4.7 hours to nearly zero with only 85% of the overall cost compared to the current practical allocation.

7 October 2019 in IEEE Transactions on Intelligent Transportation Systems
Verified
Journal Article
Closed access
FWFangxin Wang
Fangxin Wang
YZYifei Zhu
Yifei Zhu
FWFeng Wang
Feng Wang
JLJiangchuan Liu
Jiangchuan Liu
XMXiaoqiang Ma
Xiaoqiang Ma
XFXiaoyi Fan
Xiaoyi Fan
The explosion of online shopping brings great challenges to traditional logistics industry, where the massive parcels and tight delivery deadline impose a large cost on the delivery process, in particular the last mile parcel delivery. On the other hand, modern cities never lack transportation resources such as the private car trips. Motivated by these observations, we propose a novel and effective last mile parcel delivery mechanism through car trip sharing, to leverage the available private car trips to incidentally deliver parcels during their original trips. To achieve this, the major challenges lie in how to accurately estimate the parcel delivery trip cost and assign proper tasks to suitable car trips to maximize the overall performance. To this end, we develop Car4Pac, an intelligent last mile parcel delivery system to address these challenges. Leveraging the real-world massive car trip trajectories, we first build up a 3D (time-dependent, driver-dependent and vehicle-dependent) landmark graph that accurately predicts the travel time and fuel consumption of each road segment. Our prediction method considers not only traffic conditions of different times, but also driving skills of different people and fuel efficiencies of different vehicles. We then develop a two-stage solution towards the parcel delivery task assignment, which is optimal for one-to-one assignment and yields high-quality results for many-to-one assignment. Our extensive real-world trace driven evaluations further demonstrate the superiority of our Car4Pac solution.

27 December 2019 in IEEE Transactions on Mobile Computing
Verified
Journal Article
Closed access
XFXiaoyi Fan
Xiaoyi Fan
FWFeng Wang
Feng Wang
DSDanyang Song
Danyang Song
YLYuhe Lu
Yuhe Lu
JLJiangchuan Liu
Jiangchuan Liu
Automobiles have become one of the necessities of modern life, but also introduced numerous traffic accidents that threaten drivers and other road users. Most state-of-the-art safety systems are passively triggered, reacting to dangerous road conditions or driving maneuvers only after they happen and are observed, which greatly limits the last chances for collision avoidances. Timely tracking and predicting the driving maneuvers calls for a more direct interface beyond the traditional steering wheel/brake/gas pedal. In this paper, we argue that a driver's eyes are the interface, as it is the first and the essential window that gathers external information during driving. Our experiments suggest that a driver's gaze patterns appear prior to and correlate with the driving maneuvers for driving maneuver prediction. We accordingly present GazMon, an active driving maneuver monitoring and prediction framework for driving assistance applications. GazMon extracts the gaze information through a front-camera and analyzes the facial features, including facial landmarks, head pose, and iris centers, through a carefully constructed deep learning architecture. Both our on-road experiments and driving simulator based evaluations demonstrate the superiority of our GazMon on predicting driving maneuvers as well as other distracted behaviors. It is readily deployable using RGB cameras and allows reuse of existing smartphones towards more safely driving.

13 June 2017 in IEEE Communications Magazine
Verified
Journal Article
Closed access
XFXiaoyi Fan
Xiaoyi Fan
JLJiangchuan Liu
Jiangchuan Liu
ZWZhi Wang
Zhi Wang
YJYong Jiang
Yong Jiang
XLXue Liu
Xue Liu
Map services and applications have been studied extensively within the mobile computing community in the past two decades, starting from standalone GPS receivers and then moving toward connected smart terminals with live digital maps of transportation networks and even real-time traffic. More recently, with the deep penetration of modern 3G/4G networking and social networking, the crowd intelligence from the social community has been explored toward crowdsourced navigation. In this article, we first provide an overview of the past and present road navigation technologies. We then discuss very recent advances in crowd intelligence, identifying the unique challenges and opportunities therein. We further present a case study that utilizes the crowdsourced driving information to combat the last mile puzzle for road navigation.

18 September 2017 in ACM Transactions on Intelligent Systems and Technology
Verified
Journal Article
Open access
PDF file
XFXiaoyi Fan
Xiaoyi Fan
WGWei Gong
Wei Gong
JLJiangchuan Liu
Jiangchuan Liu
Many radio frequency identification (RFID) applications, such as virtual shopping cart and tag-assisted gaming, involve sensing and recognizing tag mobility. However, existing RFID localization methods are mostly designed for static or slowly moving targets (less than 0.3m/sec). More importantly, we observe that prior methods suffer from serious performance degradation for detecting real-world moving tags in typical indoor environments with multipath interference. In this article, we present i 2 tag, an intelligent mobility-aware activity identification system for RFID tags in multipath-rich environments (e.g., indoors). i 2 tag employs a supervised learning framework based on our novel fine-grain mobility provile, which can quantify different levels of mobility. Unlike previous methods that mostly rely on phase measurement, i 2 tag takes into account various measurements, including RSSI variance, packet loss rate, and our novel relative phase--based fingerprint. Additionally, we design a multidimensional dynamic time warping--based algorithm to robustly detect mobility and the associated activities. We show that i 2 tag is readily deployable using off-the-shelf RFID devices. A prototype has been implemented using a ThingMagic reader and standard-compatible tags. Experimental results demonstrate its superiority in mobility detection and activity identification in various indoor environments.