search

hungpham2511 / toppra

robotic motion planning library
https://hungpham2511.github.io/toppra/index.html
MIT License
625 stars 170 forks source link

Can I use Cartesian Space path instead of Joint Space path as input for toppra？ #5

Closed Linjackffy closed 6 years ago

Linjackffy commented 6 years ago

I want to use toppra for my research. After I finish reading the documentation and tutorials, I found that the paths are represented by Joint Space.Can I use Cartesian Space path as input for toppra? Looking forward to your reply.

hungpham2511 commented 6 years ago

Hi @Linjackffy ,

In principle, it is possible

Suppose you want to limit the Cartesian and Velocity constraints of a point moving in 3D space. One thing you can do is to think of the point as being a 3-dof robot with translational movements along the X, Y and Z-axis respectively.

Rotational kinematic constraints will be somewhat more troublesome, but you should be able to re-formulate those constraints to the form accepted by TOPP-RA.

Hope that help. Please let me know if you have any further question.

JadBatmobile commented 6 years ago

Im doing something similar.. i simply translate the cartesian waypoints to joint waypoints tho, using the inverse kinematics of my model

Linjackffy commented 6 years ago

@hungpham2511 Thanks for your reply！ Your suggestion is instructive for me. I also need to think about joint constraints like velocity, acceleration, etc, while moving along Cartisian Space path. Can I use inverse kinematics to convert the joint constraint to the cartesian constraint?

Linjackffy commented 6 years ago

@JadBatmobile Your method may lose the precision of the cartesian path.

hungpham2511 commented 6 years ago

Hi @Linjackffy

Have you resolved your issue? In fact what @JadBatmobile suggest is what I would do if you also want to constraint joint velocity/accel limits. That is to use inverse kinematics to convert a path from Cartesian space to Joint space, then constraint w.r.t all constraints: Cart Accel, Joint Vel, Joint Accel, ....