Mohamad Orabi

San Jose · California · +1 949 245 4310 · mohamad.orabi10@hotmail.com

A bit about my life .. Born and raised in Lebanon, my cousin taught me to code in 8th grade and we would have fun developing games together. I started by making python based games, and then moved on to javascript so I can publish them online, and then finally to untiy for real physics and 3D graphics.

Fast-forwarding to undergrad, I had a full ride scholarship to the top school in Lebanon, where I did some research in iterative learning control and applying ML for opportunistic navigation. I then moved on to graduate school where I matured my research interests in software defined receiver design, cellular signals (LTE & 5G), Low Earth Orbit (LEO) satellite signals, multipath mitigation, and deep/reinforcement learning.

In the last few years, I got my first taste of working with a professional team on a large codebase. My code matured to production level code with automated workflows for code building, testing, and quality checks. I also got my first taste of large scale gnss data, which helped me build an intuition and mature my understanding of the challenges facing accurate navigation. Want to test your intuition? Try the outlier detection game in the just for fun section.

Experience

Navigation/ML Engineer

OneNav, inc.

Optimized different blocks of the Java based Positioning Engine/Manager to improve the solution in various real-world scenarios (cold start - assisted - urban canyon - foliage - parking garage - tunnel). My activities included but where not limited to:

  • Research and development on outlier detection such as RAIM, RANSAC, and UKF innovations.
  • Research and development on ML models for measurement classification and bias estimation.
  • Estimating and characterizing measurement errors.
  • Root causing and fixing urgent issues in receiver performance.
  • Developing interactive dashboards for massive data analysis to aid and guide R&D.
  • Contribute to analysis, debugging, and regression tools.
Jan 2022 - Present

Graduate Student Researcher

ASPIN Lab - University of California Irvine

Conducted research which I published and presented in academic conferences. My research activities included but where not limited to:

  • Setting up hardware and conducting experiments and data collects.
  • Developing specialized software defined receivers for tracking various signals.
  • Research on exploiting cellular signals (LTE, 5G) for navigation purposes.
  • Research on exploiting low earth orbit (LEO) satellites for navigation. (Iridium and Orbcomm)
  • Research on the application of machine learning (ML) in tracking filters for multipath mitigation.
August 2020 - Jan 2022

Research Intern

ASPIN Lab - University of California Irvine

Designed and implemented a GPS receiver in C++:

  • Uses multiprocessing to achieve real-time performance.
  • Acquires and tracks satellite signals from collected IQ samples.
  • Decodes the navigation message
  • Computes the pseudorange, satellite positions, and correction parameters.
  • Computes a full position-time navigation solution.
June 2019 - August 2019

Education

M.S. in Electrical and Computer Engineer

University of California, Irvine

GPA: 3.80

  • Research Project: Opportunistic navigation exploiting LTE, 5G, and low Earth Orbit satellite signals.
  • Best Presentation Award at ION GNSS+ 2021 for my reinforcement learning for multipath mitigation work.
2020 - 2022

B.E. in Electrical Engineering

Lebanese American University

GPA: 3.91

  • Full scholarship by the University Scholarship Program (USP) hosted by the US Embassy.
2016 - 2020

Skills

This section lists the skills I have acquired during each of my experiences.

oneNav, Inc.

This experience helped me gain industry level coding skills necessary for collaboration on large projects.
Main languages used: Java - C++ - Python

  • Docker for creating isolated, reproducible, and scalable development environments.
  • YAML scripting for build and test automation (CI/CD).
  • Code testing and quality assurance for reliability and performance.
  • Training ML for resource constraint devices optimizing for performance, latency, and size.
  • Deploying ML on resource constraint devices using TensorFlow Lite for Microcontrollers.
  • Building interactive dashboards for massive data analysis.
  • Multiprocessing to speed up processing of massive data.

ASPIN Lab

This experience laid the foundational theoretical and mathematical frameworks crucial to my career.
Main languages used: Matlab - C++ - Python

  • Comprehensive understanding of Linear Algebra, Estimation Theory, Probability, and Stochastic Processes.
  • Theoretical and practical understanding of Kalman Filtering and Sensor Fusion.
  • Deep understanding of GNSS theory and design choices: PRN gold codes, spread spectrum signaling, code and phase tracking, navigation message, pseudorange models, and error sources.
  • In-depth knowledge of machine learning theory and techniques: supervised, unsupervised, and reinforcement learning.
  • Familiarity with cellular standards (LTE, 5G), specifically synchronization signals.
  • Familiarity with Low Earth Orbit (LEO) satellite signals, namely Orbcomm and Iridium.

Personal Projects

Includes everything from games to utility scripts, message spammers, data scrapers, and even an augmented reality (AR) function visualizer.
Main languages used: Python - Matlab - C++ - C# - HTML - JavaScript - Swift

  • Great problem solving skills and can-do attitude.
  • Exceptional debugging skills and proficiency in debugging tools and techniques.
  • Quickly learning new skills, programming languages, libraries, etc.
  • Swift for iOS development.
  • Unity for 3D games and physics simulations.
  • Image processing and filtering.
  • HTML for websites.
  • Intuitive UI design.

Publications

A Machine Learning Multipath Mitigation Approach for Opportunistic Navigation with 5G Signals

First Author

ION GNSS+

September 2021

Opportunistic Navigation with Doppler Measurements from Iridium Next and Orbcomm LEO Satellites

First Author

IEEE Aerospace

April 2021

Iterative Learning Control: Practical Implementation and Automation

Co-Author

IEEE Transactions on Industrial Electronics

March 2021

Carpe Signum: Seize the Signal

Co-Author

Inside GNSS Magazine

February 2021

A Machine Learning Approach for GPS Code Phase Estimation in Multipath Environments

First Author

IEEE/ION Position Location and Navigation Symposium PLANS

April 2020

Presentations

A Machine Learning Multipath Mitigation Approach for Opportunistic Navigation with 5G Signals
(Awarded best presentation)

First Author

ION GNSS+

September 2021

Opportunistic Navigation with Doppler Measurements from Iridium Next and Orbcomm LEO Satellites

First Author

IEEE Aerospace

April 2021

Just for Fun

Outlier Detection Game

Guess the outliers. The game simulates realistic gnss measurements and adds bias with a random probability based on the chosen difficulty.

How to Play:

  • Click on the satellites (0 or more) that you suspect are outliers based on the residual information in the plot.
  • Click Submit Guess to reveal the truth.

What You Need to Know:

  • Minimum bias value added is 10 meters.
  • All measurements have equal weights in the estimator.
  • Residuals are calculated with respect to the estimated position using the biased measurements.
  • Realistic satellite positions are generated by fetching TLE files.

Interactive GNSS Dashboard Demo