Why is a joint space trajectory unable to deal with operational space obstacles?

Question

I have found several sources that state that a Joint space trajectory model is not capable of dealing with obstacles in the manipulator workspace. These sources include:

slide 10 of "Manipulator Trajectory Planning":

URL: https://cw.fel.cvut.cz/old/_media/courses/a3m33iro/080manipulatortrajectoryplanning.pdf

and also Mathworks in their presentation on Trajectory Planning for Robot Manipulators on Youtube at timestamp 3:16.

URL: https://www.youtube.com/watch?v=Fd7wjZDoh7g&ab_channel=MATLAB.

The problem I am facing is that these sources do not explain why joint space models can not deal with obstacles.

Is there somebody here that can shed some light on this?

Answer 1

The main difference between task space control and joint space control are :

Joint space : You know start and last position of joint configuration from IK(Inverse kinematic), and interpolate it

Task space : You know start and last position of end-effector, then interpolate it and IK every point in path to get joint configuration for every point in path.

Obstacle are known in cartesian space (x,y,z, and orientation). The other component we know in cartesian space is end-effector. Obstacle avoidance only able to be performed on task space control because on task space control we have information position and orientation of end-effector which may collide with obstacle, thus we can plan the path of end-effector. While in joint space we didnt have that information because we didnt plan the path of end-effector(only take the first and last position and orientation of end-effector).