liangyuwei / dual_ur5_arm

Employ Dynamic Movement Primitives (DMP) and PSO to find a time-optimal position for dual-arm peg-in-hole task.

dual_ur5_arm

A half-done implementation of an imitation learning approach to dual-arm peg-in-hole task, together with some other tools.

Only dual-arm movements are considered in learning, force is not considered.

Some useful tools

1. Time Optimal Path Parameterization that converts a path into time-optimal trajectory while considering velocity and acceleration limits.
   (1) TOTG: A service implemented in raw_totg folder, using code from ROS implementation.
   (2) TOPP and TOPP-RA: Services implemented in dual_ur5_control/script/ folder as TOPP_service.py and TOPPRA_service.py.