Sungwook Moon

Ph.D. Candidate
Mobile Networking Laboratory
Dept. of Computer & Information Science & Engineering
University of Florida

Home

CV/Resume

Research

Links

Personal

Embedding Community Contact Profiles on Autonomous Mobile Nodes

(Sungwook Moon, Ahmed Helmy)

We propose a profile embedding of contact-based rule decision on autonomous mobile nodes. In this decision model, each autonomous node is an entity that can make a mobility decision based on given behavioral rules. In addition, we design a scheduler to emulate real-world contact behavior and apply the power-law distribution, which we discover from our own analysis of real-world group sizes according to contact pattern. We provide various knobs to tweak the parameters to emulate different aspects of real world contact pattern. Our evaluation shows that power-law distribution, frequent contact with community members and periodical contact pattern that are similar to real-world distribution.

Memory mode

Random mobility (random walk)

[Figure 1. Youtube videos that show the community convergence by autonomous nodes with a memory mode (top) vs random mobility (bottom).]

Presented at MobiCom ’11 SRC (Student Research Competition).
Accepted to publish at Mobile Computing and Communications Review ’11.
Under review for journal publication (Elsevier Adhoc Networks).

Included in my Ph.D. Dissertation.

Mobile Networking Testbed with an Attitude

(Sungwook Moon, Ahmed Helmy)

This is a prototype of mobile networking testbed that can emulate community-based human contact pattern. It has two components: 1) networks of robots and 2) participatory testing. Firstly, the networks of robots have robots (we use iRobot Create) that emulate human mobility with embedded personality profile. Secondly, participatory testing is a concept which volunteers can participate in testing communication protocols by downloading and installing the protocols in their mobile devices. We provide a prototype implementation for proof of concept for the network of robots with personality profile; thereby, blending the network of robots with participatory testing via emulated mobility of human. Detailed study for participatory testing is out of scope of our work though we show part of studies with Bluetooth discovery measurements.

[Figure 2. Youtube video that shows a prototype mobile networking testbed for proof of concept.]

Presented at IEEE GlobeCom’10, ACM MobiCom’10, WINTECH workshop’10.
Published at the proceedings of IEEE IWCMC’11.

Included in my Ph.D. Dissertation.

Understanding Periodicity and Regularity of Nodal Encounter

(Sungwook Moon, Ahmed Helmy)

Understanding human encounter pattern is an important problem space to the design of message transfer between human carrying mobile devices in mobile networks. We analyze the periodicity of human encounter pattern by spectral analysis. We investigate large and extensive datasets - WLAN traces from two campuses (UF, USC), city of Montreal and Bluetooth encounter traces we collected in college class that contain over 50,000 users and millions of encounter pairs. By looking at the periodicity of encountered pairs, we discover that most of the mobile nodes show a strong trend of periodical encounters, particularly weekly encounter, with an exception of Montreal trace. We also observe that individual nodes’ encounter pattern and rarely encounter pairs show even stronger periodicity.

[Figure 3. Periodicity graphs of encounter days in frequency domain for rarely encountering pairs over 128 days of traces. Frequency peak at 18 shows the presence of strong weekly encounter pattern as 18 times over the period of 128 days yields the repetition of the pattern every 7 days (18*7.xx = 128). ]

Presented at IEEE GlobeCom’10, ACM MobiCom’10.
Published at the proceedings of IEEE GlobCom ’10.

Included in my Ph.D. Dissertation.

Visualization and Representation of Mobile Users

(Sungwook Moon, Udayan Kumar, Wei-jen Hsu, Ahmed Helmy)

We develop a framework to visualize the mobile users on Google Earth. Our proposed framework can extract mobile users’ data from database, parse the data and convert them to kml file that can display the mobile users on Google Earth.

[Figure 4. Youtube videos that show the animation of grouped WLAN users’ activity over time with different presentations. Groups are defined by clustering similar location visiting preference.]

Presented at IEEE INFOCOM’08, SECON’08, ICNP’08.
Published at the proceedings of IEEE SECON ‘08

NOT Included in my Ph.D. Dissertation for coherency of story line. Contributions between authors are equal and the sequences of authorships changed in different presentations.

Modeling Mobile Users Based on Periodicity Profile

(Sungwook Moon, Ahmed Helmy)

With our analysis of periodicity for mobile users’ encounter, we propose a methodology to profile mobile users based on their encounter (contact) behavior and generate synthetic trace using the profiles. Our process is useful in modeling behaviors of target profile users for the experimentation of communication protocols and prediction of mobile encounter based on periodicity.

Manuscript in preparation.

Partially included in my Ph.D. Dissertation.

Analysis of Bluetooth trace with WiFi association for mobile users over time and various user groups

(Sungwook Moon, Udayan Kumar, Ahmed Helmy)

We collect Bluetooth and WiFi encounter of various user groups (class students, faculty, and lab members) over time. We analyze large quantity of these traces and highlight important trends.

Work in progress.

Partially included in my Ph.D. Dissertation.

Last update: Jun 6, 2011