This system adjusts the robot’s shape to minimize drag forces in environments with flowing currents. My team developed a robot system controlled by Reinforcement Learning (RL). Additionally, to address the challenges of real-world learning in aquatic environments, we implemented a fluid simulator for use in the learning environment.
This system places a stent in the target lesion of the coronary artery via a controlling manipulator made by Asan Medical Center. It is the world’s first autonomous PCI robot system. My team developed reinforcement learning-based control software. This project was awarded first place at Johnson & Johnson QuickFire Challenge.
The purpose of this system is to help doctors diagnose of lung cancer. It supports diagnosis by deep learning based on 3d detection and classification nodules.
My team participated in NeurIPS 2018 - AI for Prosthetics Challenge and took 17th place out of 401 teams. We were able to get valuable practical experiences with reinforcement learning and biomechanics domains.
This new type of product is a network camera working as an IoT gateway. It gathers sensor data and provides various services, such as social protection, fire monitoring and missing child prevention, by cooperating with adjacent IoT devices.
This service integrates the cloud system and network camera. It can provide necessary information from a network camera in the local network to the outside network user, all while maintaining security.
Brief descriptions of other projects in Samsung Techwin.