5 Commits

Author	SHA1	Message	Date
Derek Witcpalek	ba6901d32a	Modified arm controller node to handle all arm control. It receives desired end effector poses and actuates the arm. Also includes some clean up.	2020-04-15 15:52:49 -04:00
Derek Witcpalek	ef26d8ffef	Working arm movement to specific pose (x,y,z,roll,pitch,yaw). There are a lot of other changes too. ... -Removed gravity (temporary, I hope). I was struggling to get inverse dynamics working correctly to provide a torque to compensate for gravity. Eventually, I think we want to have a PID, plus gravity compensation. -Created a demo world with setpoints that the robot heads to. -Arm kinematic representation is working. -PID control of each joint is working to a basic extent. Some segments are more responsive than others. Still to do: -reading in SDF should be more robust, and should support all joint types -better control -obstacle avoidance in arm movement path -control with gravity -arm joint limits, and shortest path to angle (ie considers that theta wraps at 2 pi	2020-04-12 14:18:01 -04:00
Derek Witcpalek	6acb7158a0	Working kinematic representation. One caveat: there is no link representation between the world frame and the robot base frame. Sometimes when the simulation starts, the arm slides slightly off of (0, 0, 0) which causes joint positions to be slightly off (by the same amount).	2020-04-09 17:05:06 -04:00
Derek Witcpalek	6a73cd2371	Forward kinematics verification is mostly working for individual joints. The axis of rotation is correct for all joints, but some cause incorrect amounts of motion. Specifically, moving the wrist_pivot joint causes a circular motion of the end effector, when (I believe) it should cause no movement, just rotation, of the end effector.	2020-04-08 23:06:53 -04:00
Derek Witcpalek	8e25bb09a3	Implemented initial dynamics model which is read in from an sdf file. There are currently some issues with coordinate transforms.	2020-04-07 22:45:22 -04:00