Tung Do

Tung Do

M.Sc. Electrical and Computer Engineering - Robotics candidate | Research Assistant | Robotics Engineer

Publications

Enhancing Vision based SLAM through Shadow Removal Processing (Unpublished manuscript)

Published in , 2024

  • Abstract—In general, vision-based SLAM struggles to detect dynamic objects, which complicates tracking for Unmanned Ground Vehicles (UGVs). This issue arises because vision-based SLAM is susceptible to environmental factors such as shadows or significant changes in illumination, which can affect object detection. Particularly, raw data that excludes dynamic objects but includes shadows does not accurately represent the real environment. Our objective is to implement a shadow removal algorithm that addresses both static and dynamic objects at the front end of the SLAM pipeline to see whether it improves our SLAM accuracy results.
  • Index Terms: Ground Vehicle, Vision-based SLAM, Shadow Removal

Recommended citation: Wan, H., Kusumadjaja, K., Lee, S.H., & Do, T. (2024). Enhancing Vision-based SLAM through Shadow Removal Preprocessing. Unpublished manuscript, University of Michigan, Ann Arbor.

Enhancing Monocular 3D Object Detection in Foggy Conditions: An Adapted MonoCon Approach for Autonomous Vehicles (Unpublished manuscript)

Published in , 2023

  • Abstract: This paper explores advancements in monocular 3D object detection, a pivotal aspect of autonomous vehicle technology. We focus on enhancing detection accuracy and robustness in diverse weather conditions, specifically addressing the challenges in foggy scenarios. Implementing the MonoCon model, our methodology includes transfer learning, image augmentation techniques, and pre-processing strategies to improve visibility in foggy images. Challenges such as fluctuating Average Precision (AP) values and inefficient detection of distant or small vehicles in fog are addressed through a revised evaluation strategy and targeted image processing. Results showed an increase in AP from 7.05% to 17.67% for the normal dataset after training to more epochs and up to 25.82% for foggy conditions after training to 300 more epochs and applying CLAHE and blur. These findings underscore the model’s adaptability and effectiveness in diverse environments.
  • Index Terms: Monocular 3D Object Detection, Autonomous Vehicles, Deep Neural Networks (DNN), Deep Learning in Computer Vision

Recommended citation: Do, T., Liu, X., & Swayampakula, R. (2023). Enhancing Monocular 3D Object Detection in Foggy Conditions: An Adapted MonoCon Approach for Autonomous Vehicles. Unpublished manuscript, University of Michigan, Ann Arbor.

Autonomous/Remote Control Mecanum Wheels Tesla Roadster in real-world (Unpublished manuscript)

Published in , 2022

  • Abstract: This project aims to apply everything the student has learned in engineering about mechanics, machine elements, programming, and electronics to build a Mecanum Wheels Tesla Roadster to research the possibility of flexible movement in developing a smart vehicle network.
  • Keywords: Mecanum, Robot, Wheel, Tesla, smart car, innovative vehicle, network

Recommended citation: Do, T. (2023). Autonomous/Remote Control Mecanum Wheels Tesla Roadster in real-world. Unpublished manuscript, California State Polytechnic University, Pomona.