Hadush Hailu - Academic Website

About Me

I’m Hadush Hailu, a Master’s student in Computer Science at Maharishi International University (USA), graduating December 2025. Under supervision of Prof Ankhtuya Ochirbat my research spans multimodal learning and deep metric learning, with prior work in multi-robot collision avoidance from my second Masters in robotics at the Intelligent Robot Laboratory , University of Tsukuba (Japan) under Prof. Akhisa Ohya and Prof. Aynori Yorozu .

I am seeking PhD opportunities starting in Spring 2026, with research interests in vision–language–action for generalist manipulation, computer vision, and applications in agricultural and service robotics. I am also interested about collaborative research and participation in international robotics competitions. You can reach me out at had.hailu [at] gmail.com

Selected Publications

Density-Aware Adaptive Line Margin Loss (DAAL) for Multi-Modal Deep Metric Learning

H Gebrerufael, A Tiwari, G Neupane, G Hailu, B Bora

IEEE/5th International Conference on Electrical, Computer and Energy Technologies (ICECET 2025), 03–06 July 2025, Paris, France.

Presentation

Video Presentation | GitHub | Slides

Experimental Investigation of Mutual Avoidance Behavior for Multiple Autonomous Robots

H Hailu, A Yorozu, A Ohya

The 38th Annual Conference of the Robotics Society of Japan (RSJ 2020), Vol. 38, October 9–11, 2020.

Demo | GitHub

Forecast-Driven MPC for Decentralized Multi-Robot Collision Avoidance

H Hailu, B Gebregziabher, P Raj

IEEE/The 8th International Conference on Intelligent Robotics and Control Engineering (IRCE 2025), Kunming, China, August 18–21, 2025.

Video Presentation | GitHub | Slides

Experimental investigation of mutual collision avoidance behavior for multiple mobile robots

A Yorozu, H Hailu, A Ohya

IEEE/ASME International Conference on Advanced Intelligent Mechatronics, 2021

Selected Projects

Phenotyping and Precision Agriculture (2025)

Leveraging ROS for robot control and TensorFlow for DL image processing to automate phenotyping in crop breeding applications.

Legged Robot (Boston Dynamics SPOT) for Industrial Site Inspection (2023)

Developed control and navigation algorithms for the Boston Dynamics SPOT robot to enable autonomous industrial site inspections. ROS for control and navigation with PyTorch x FastRCNN for grate and puddle detection.

Nursing care robot (2023)

In this project, I utilized the Sony Hanamoflor mobile humanoid robot to explore applications in elderly care assistance.

Showcasing Our Robot

Multi-robot coordination and control (2022)

Developed a multi-robot coordination system with collision avoidance by integrating dynamic obstacle detection via an Extended Kalman Filter (EKF) and Reciprocal Velocity Obstacles (RVO) for safe navigation among multiple robots without collisions.

Service Humanoid Robot (2023)

Developed a service robot capable of autonomously cleaning tables in a restaurant setting, using ROS for navigation and MoveIt for manipulation and motion planning.

Affordable 6-DOF Manipulator Design (2025)

A lightweight, 3D-printed 6-DOF robotic arm designed to be affordable, easy to build, and suitable for research and educational use.

Full Video

Research in Progress

Multi object tracking for predictive collision avoidance

Bruk Gebregziabher, Hadush Hailu

https://doi.org/10.48550/arXiv.2307.02161, 2023

Presentation

Demo | Simulation1 | Simulation2

Deep Learning Based Amharic Chatbot for FAQs in Universities

Goitom Ybrah, Hadush Hailu, Shishay Welay (PhD.)

https://doi.org/10.48550/arXiv.2402.01720, 2024

GitHub

Robotics Competition

iCog Makers RoboSoccer Cup, Ethiopia (2017)

We participated in Ethiopia’s first national RoboSoccer competition as Team Phoenix (Mekelle University & Mekelle Institute of Technology) and won 2nd place with our low-cost autonomous RoboSapien platform.

Full Video | Paper

Player Robot Competition, University of Tsukuba, Japan (2020)

We developed a robot that autonomously plays rock–paper–scissors using sensing and control algorithms. Although we did not win, the project offered valuable experience in real-time perception and human–robot interaction.

Full Video

Tsukuba Challenge, Tsukuba, Japan (2019)

As Team Aqua from the University of Tsukuba Intelligent Robot Laboratory, we designed a mobile robot and implemented the ROS Navigation Stack to compete in the outdoor autonomous navigation challenge. The project focused on robust localization and path planning in open-area environments.

Video | Learn about the challange, Paper

Robotic Platforms I’ve Worked With

Four Legged Boston Dynamics SPOT robot

Cobot 6DOF Manipulator Robot

Nyokkey kawasaki humanoid robot

HSR mobile manipulator robot

solinftec agronomy robot

vstone mobile robot

Sony HANAMOFLOR nursing care robot