Possible wrong definition of computed torque control

[mrkjankovic]

Errors in books are a nightmare for the readers and the authors! If you spot something that is wrong, or something could be improved (clearer, less confusing, less ambiguous) please log it here. This will help fellow readers of the book, and will improve future editions of the book.

Where is the error? The possible error is either within Equation 9.16 on page 382 or within Figure 9.12 on page 381, both describing the computed torque controller.

What is the error? From Figure 9.21, preceding Equation 9.16, one would imply (in my opinion) that the coefficient matrices \hat{C} and \hat{g} are functions of the measured (not desired) joint positions and velocities, q and \dot{q}, respectively, as usually assumed in other robotics textbooks. On the other hand from Equation 9.16 one would imply that those same matrices are functions of desired joint positions and velocities, q^* and \dot{q}^*, respectively. With this in mind the "Config" and "JointVel" inputs of the "Inverse Dynamics" Simulink block in Figure 9.21 should (in my view) be the outputs of the "Polynomial Trajectory" block, as illustrated in Figure 9.21 of the previous version of this book. Therefore, I believe that there is an inconsistency between Equation 9.16 and Figure 9.12 for what concerns the representation of the computed torque controller and warrants in my view a second glance.