Convergent Manufacturing (ARM 22-01-F-07)

Developed a kinematic calibration framework for robot arms, leveraging motion capture systems to achieve sub-millimeter accuracy in TCP position estimation, significantly surpassing baseline performance by sixfold.

Developed a scan-n-print framework for wire arc additive manufacturing, resulting in a 66% smoother deposition than the baseline, while integrating multiple sensor interfaces to enhance manufacturing cell capabilities.

Emotional Expressive Robot

Designed a control strategy for a 10-degree-of-freedom mobile manipulator to express internal information, focusing on emotions, through motion behaviors. Utilized crowdsourcing to learn the best parameter combinations for expressing different emotions.

DARPA Subterranean Challenge

I am the Team Lead of Team NCTU from January, 2019. Team NCTU have been competing in SubT Challenge in Tunnel Circuit and Urban Circuit, where we got #7 and #8 among all team and #2 among all self-funded team.

Watchtowers - Enable Localization of Duckiebot in Duckietown.

Duckietown has been a great platform to learn and conducting reseach in robotics and autonomous driving car. To keep advancing the platform, the ability of localizing those Duckiebot is crucial. Here, we present the watchtower solution. We chose watchtower is a tower-like infrastructure inside Duckietown. They survellent over the whole city and try to localize Duckiebots using the Apriltag on top. Every watchtower is equiped with a RPI 3B+. They detect Apriltag and send the localization results back to the server computer where optimization is performed. The reason we chose watchtower over over-head cameras is that we would like these little guy to be a part of the town.