Undereractuated Astrobee Hand 💫🐝Robotic Manipulation and Mobility (ROAM) Laboratory, Columbia University | Spring 2025
Advisors: Eric Chang, Matei Ciocarlie


During this project, I made hardware and control updates to an exisiting robotic hand to introduce tactile sensing capabilities. The underactuated, single-DOF, tendon-driven hand was originally desgined for use on the Astrobee platform aboard the ISS, but has since become a research platform in the lab. It features three fingers, each with two flexion/extension joints. Two of the fingers also have abduction/adduction capabilities, totaling eight joints controlled by a single motor. The whole hand is actuated by a system of tendons and pulleys that run througout.



My first task was to assemble the hand, evaluate how it moved, and make needed improvements. Here, I focused on ensuring the proximal and distal flexion/extension joints in each finger moved together in sync. In other words: the outer finger joints tended to bend before the inner ones, resulting in a dysfunctional grasp. To fix this, I changed the pulley design to reduce friction, re-calibrated the pulley sizes, and redesigned the finger housing to accomodate the changes.





Original Motion

Improved Motion

Using the improved design, I then worked on integrating tactile sensing into the hand. In doing so, we hoped to create a hand that could grasp an object lightly, attempt to lift it, determine if it was slipping through the fingers, and adjust the grasp accordingly. This would make the hand functional for a variety of objects, from fragile items like oragami or balloons (which require a gentle grasp!) to robust, heavy pieces like ceramic cups and soup cans.

To do so, I mounted a tactile sensor to the thumb, then developed a control system to communicate over ROS2 between the Teensy board running the hand and a controller script on my laptop. The controller script implemented a state machine with five states that used measured changes in capacitive feedback to determine when to open and close the hand. 

Finally, we conducted a variety of tests with the hand, demonstrating some success in adaptive grasping!



Empty Cup


Heavy Cup


Pouring Beads
Get in touch! 
kathrynlampo@gmail.com |