Loading...
  • Roozbeh Ketabi, Ph.D.

About Me.


I received my Ph.D. in computer engineering from University of Florida in 2020 amid the Covid pandemic. I was a member of Mobile Networking Laboratory under supervision of Prof. Ahmed Helmy in Computer and Information Science and Engineering department of University of Florida. During my graduate program, I have been UF graduate school and UF Informatics Institute graduate fellows. Prior to joining UF, I received my B.Sc. in Computer Engineering - Information Technology from department of Computer Engineering at Sharif University of Technology.

My general research direction is mobility modeling with a focus on vehicular mobility (in particular applications of machine learning in mobility). I've also been a part of a series of collaborations on mobility modeling in pedestrian settings.

You can find my CV here.

Updated April 2021.

Research.


As part of MobiBench, we explore how people move in different settings and how it affects potential applications (i.e. networking, ride-sharing, etc.)

This includes systematic analysis, statistical and machine learning modeling for various goals including forecasting and prediction, from different spatio-temporal perspectives, simulation and scenario generation, and these kind of stuff!

Vehicular Mobility: En Route to Urban-scale Simulations

Using estimates of traffic density extracted from traffic cameras around the globe, we propose a framework to estimate traffic demands in the form of an OD matrix. Then we examine different routing ideas to finally generate simulation scenarios using SUMO.

For more details including data and generated simulation scenarios (for London and Washington DC. for now) visit our En Route page.

Vehicular Mobility: Playing with Matches

Using data of kolntrace we explored the trips happening in city of Cologne from a similarity point of view. By proposing a spatio temporal measure of similarity we find clusters that are spatially and/or temporally distinguishable. Then we explore the applications of it in different formulations of ride or car sharing (carpooling, catch-a-ride, and car sharing or minimum path coverage in ride request graph).

For more details, see its repo on my github Playing with Matches page.

Vehicular Mobility: Through the Eyes of Traffic Cameras

In this investigation we designed and evaluated various timeseries models including recurrent neural networks (deep learning) to forecast density of traffic (and/or count of cars) against the estimates from the traffic camera images.

We have also collected newer images, for a longer timespan and higher spatial resolution, to process through more advanced image processing methods to create a more accurate & recent basis for the study.

Pedestrian Mobility: Flutes vs Cellos

Huge traces of wireless lan access point logs (100s of gb) and network flows (terabytes) are the data foundation and we lay our work on. We have investigated the behavioral differences of highly mobile users vs those of stop to use nature, taking steps toward integrated modeling human mobility and network usage.

Pedestrian Mobility: Predictability

Work on predicting next building or access point where users are most likely to go in our campus setting. Building on processed data from our wireless lan traces and using machine/deep learning models in addition to theoretical predictors we investigated the matter in different spatio-temporal granularities.

Side Projects

Collaborating with colleagues in other departments on problems that require computer engineering expertise. So far we've worked on a blockchain prototype and a finite difference time domain method based 3d simulation of sound waves on cuda nvidia gpus.

Publications.


  • Roozbeh Ketabi, Mimonah Al-Qathrady, Babak Alipour, and Ahmed Helmy. "Vehicular traffic density forecasting through the eyes of traffic cameras; a spatio-temporal machine learning study." ACM DIVANET, 2019.
  • Babak Alipour, Leonardo Tonetto, Roozbeh Ketabi, Aaron Yi Ding, Jörg Ott, and Ahmed Helmy. "Where are you going next? A practical multi-dimensional look at mobility prediction." ACM MSWIM, 2019.
  • Babak Alipour, Leonardo Tonetto, Roozbeh Ketabi, Aaron Yi Ding, Jörg Ott, and Ahmed Helmy. "Practical Prediction of Human Movements Across Device Types and Spatiotemporal Granularities." arXiv:1903.00951, 2019.
  • Roozbeh Ketabi, Babak Alipour, and Ahmed Helmy. "Playing with matches: vehicular mobility through analysis of trip similarity and matching." ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems, 2018.
  • Babak Alipour, Leonardo Tonetto, Aaron Ding, Roozbeh Ketabi, Jörg Ott, Ahmed Helmy. "Flutes vs. Cellos: Analyzing Mobility-Traffic Correlations in Large WLAN Traces." IEEE INFOCOM, 2018.
  • Azad, H., Ketabi, R., & Siebein, G. (2018, December). "A Study of Diffusivity in Concert Halls Using Large Scale Acoustic Wave-Based Modeling and Simulation." INTER-NOISE and NOISE-CON, 2018
  • Roozbeh Ketabi, Babak Alipour, and Ahmed Helmy. "En Route Towards Vehicular Mobility Scenario Generation at Scale." IEEE INFOCOM SmartCities, 2017.
  • Roozbeh Ketabi, Babak Alipour, and Ahmed Helmy. "Poster abstract: En route towards trace-based simulation of vehicular mobility." IEEE INFOCOM Poster, 2017. [Best Poster Award]
  • Alireza Shojaei, Roozbeh Ketabi, Mohamad Razkenari, Hamed Hakim, Jun Wang. "Enabling a circular economy in the built environment sector through blockchain technology." Journal of Cleaner Production, 2021

Teaching.


  • Assisted Peter Dobbins (as teaching TA) in Computer Organization (CDA3101) and Discrete Math. (COT3100)
  • Taught COP3275: Programming using C, for three semesters in 2018-2019
  • Assisted Dr. Ye Xia in Computer Networks (multiple semesters) and Dr. Timothy Davis in Discerete Math.
  • Assisted in Computer Networks by design and implementation of proof of concept of programming assignements (NAT, Traffic Controller and IPv4 Tunneling) on PARTOV framework to facilitate understanding of network concepts.
  • Assisted in Systems Analysis and Design (and also Application [Re]Engineering) by guiding groups of students on different stages of their project (definition, methodology, technologies and implementation, documentation, etc.). The projects were Auto Insurance Information System, Paper Evaluation Information System and Inventory Management Information System.
  • Assisted in many more courses such as Database Design, Computer Organization and Language, Fundamentals of IT, Computer Simulation.
  • Experience with teaching C++ to groups of junior students.

Skills.


Data Science and Machine Learning:

Python's Pandas, Numpy and Scikit-learn, Hadoop MadReduce in Java. Tensorflow and Keras for deep learning. Some experience in R.

Programming:

Mainly C/C++, Python. Comfortable with Java.

I have done some R, C#. Net, Php, and Matlab

Also have done some MIPS programming and Verilog HDL.

Not afraid of Unix and Linux! Have done some Bash scripting in my time.

Next to learn: What is functional programming :thinking: (Scala and Haskell I guess)

Have done some HTML, CSS, Javascript and JQuery, ASP .Net, Php Symfony and Python Django.

One day I'll learn NodeJs and AngularJs! (but then probably there are newer frameworks for both web and data science!).

Database:

[Theoretical] SQL, MS SQL Server (T-SQL and familar with Entity Framework), Oracle (PL/SQL) and MySql. Have worked with warehousing (i.e. Hive). Some familiarity with Apache ecosystem.

Next to learn: NOSQL and other types of persistent storage (document storage, key-value, graph, etc.)

Simulation:

SUMO for vehicular mobility, Cisco Packet Tracer, OMNet++, NS2 (Otcl), ONE simulator, and Arena (minor experience).

Software tools, IDEs, cloud systems and etc.:

VS Code, MS Visual Studio, Android Studio (IntelliJ), PhpStorm, PyCharm, Eclipse (C++ and Java), VIM.

Adobe Photoshop, Adobe Lightroom, Autodesk Autocad (LoL!).

Amazon AWS services including: S3, EC2, EMR, EBS, Elastic Beanstalk (web app deployment)

Wireshark and packet inspection.

Hardware modelers such as XilinX ISE, Altera Quartus and Mentor Graphic's Modelsim

Projects.


Below is a short description list of some of project that I can remember that I've done:

  • Seq-to-Seq deep learning model on Tensorflow.
  • LSTM based timeseries predictors.
  • Anomaly detection framework
  • Analysis of health status of elderly community
  • Datascience NIST Prepilot Evaluation
  • PageRank MadReduce
  • BerryDroid (app, cloud, and raspberry based smart shop. cart)
  • VehiBench/MobiTrace
  • A Video Streaming Android App Core
  • SHARK Messenger
  • Implementation of Djkstra Algorithm with Fibonacci Heap
  • BaaZaaR, a simple ebay like market
  • Personal Cinema
  • News Show
  • OSPF Implementation and Simulation
  • Packet Filter/Firewall Implementation and Simulation
  • NAT and Traffic Controller Gateway
  • Simplified BitTorrent Client
  • Command Line Based 2 Player Online Board Game!
  • FreeBSD Kernel Scheduler Modification
  • 3D Checkers Board Game!
  • 2D Chess Game!
  • Museum Management Information System
  • Simulation and Optimization of a Workshop
  • Polynomial Multipliers in HDL
  • A lot of Algorithm Problems

Contact Me.


Address

E401, CISE, University of Florida, 32611

Links

Linkedin

Github

Email

myfirstname at$i$n { cise.ufl.edu , ufl.edu }

Phone

Ask!