Question about velocity calculation

[sritee]

Thanks for the great repository and clean code!

I was going through the code for educational purposes, and had a question about the following line.

//Include a velocity estimate for velocity-controlled robots
  Eigen::VectorXd joint_vel(delta_theta / parameters_.publish_period);

jog_arm/jog_arm/src/jog_arm/jog_arm_server.cpp

Line 552 in c88847c

Eigen::VectorXd joint_vel(delta_theta / parameters_.publish_period);

Multiplying the inverse of the jacobian with the end eff. vel should give us the joint velocities directly. Why do we need to divide this by publish time period? I am probably misunderstanding something.

Thank you!

[AndyZe]

I think what you propose sounds like a decent option.

I believe the way it's implemented now is more efficient, though. Since we already have delta_theta, and dividing by a scalar is much more efficient than a matrix multiplication.