Skip to content
CV
Index

RoboArm

Reinforcement learning to control a bio-inspired arm driven by cables and springs.

Role
Student researcher, SRTP
Period
2026 – present
Status
Early stage
Advisor
Prof. Liangjing Yang, ZJUI
Illustration · the idea, not measurements90° of flexion
spoolelasticcable

Or drag the forearm. As the elbow folds, the cable winds onto the spool and the elastic behind the elbow stretches.

Energy in the elastic

It grows with the square of the stretch, so the last part of the fold stores the most.

Release the cable and the elastic gives its energy back, returning the arm. RoboArm studies how a learned controller works with this.

Results so far

  • No experimental results yet.

Limits

  • Early stage: the mechanism and controller are still being built.