Selected Projects
Arash Social Robot
Arash is a social companion robot for children with cancer. My focused research on the Arash project is utilizing SLAM techniques enable the robot to concurrently map the environment and localize itself in the map. I have also investigated augmenting the SLAM technique by a person-tracker module to achieve higher mapping and localization accuracies. In addition to this, I was the key person who was mainly in charge of Arash's software development. I have developed some variety of different software packages from low-level control modules to high-level-control modules utilizing the ROS ecosystem. Personally, Arash is the most amazing project that I have ever contributed to with a great team. Furthermore, since Arash is devoted to helping children with cancer, I found it a lovely philanthropic project. We have also tested Arash in some hospital environments and investigated its acceptability and aliveness in a storytelling scenario.
Painter Robot
During my stay at the University of Trieste in Italy, I have worked on the painter robot project under the supervision of Dr. Gallina. In this project, I have developed a novel optimized non-photorealistic rendering method which is inspired by Painterly algorithm. Utilizing the new method minimizes brush strokes overlapping on the canvas and decreases the cost of operation time. Additionally, I have also developed a python interface which gets brush strokes as the input data and paints them on the canvas by controlling a UR10 arm robot using MoveIt and ROS.
Ship/Iceberg classification using a deep neural network (Kaggle competition)
The above topic was the final project of my "Stochastic Pattern Recognition" course. The project marks of all of my classmate are evaluated based on both the final report and their ranking on the Kaggle competition. Fortunately, my CNN model could acquire the best classification accuracy on the test data in the class.
Magnetic Levitation
As you see in the following video clip, I and my colleague, Masoud Jafaripour, implemented a magnetic levitation experimental setup from scratch. In this project, we have utilized Arduino for running the control loop, a power electronic circuit, a copper coil, and two Hall effect magnetic sensors. The tricky part of controlling this system was the nonlinear nature of it.