Chan‐Won Park

Electronics and Telecommunications Research Institute

Papers

Total Citations

H-Index

About

Chan-Won Park is a researcher advancing the frontier of robot learning, with a focus on bridging the gap between simulation and real-world application. His work centers on three key areas: behavioral cloning from demonstration, deep reinforcement learning for control policy, and reward shaping for complex manipulation tasks. In his 2020 paper "Robotic Behavioral Cloning Through Task Building" (6 citations), Park explored how robots can efficiently learn policies by directly mapping demonstrations to actions, a paradigm that expands the deployability of robotic systems. His subsequent work, "Learning Control Policy with Previous Experiences from Robot Simulator" (4 citations), demonstrated cost-efficient training of physical robot actions by leveraging simulated environments, addressing the critical challenge of reward engineering. Park’s most specialized contribution, "Learning Robot Manipulation based on Modular Reward Shaping" (2 citations), tackles the complexity of discontinuous action spaces in deep reinforcement learning, building on advances that have surpassed human performance in domains like Atari. Together, these studies form a cohesive effort to make robot learning more accessible, efficient, and robust, with implications for manufacturing, service robotics, and autonomous systems.