fixed restoring forces formulation #4
Conversation
src/hydrodynamics.cpp
Outdated
| @@ -155,8 +155,8 @@ Eigen::Vector6d RestoringForces::calculateRestoringForcesVector(const Eigen::Mat | |||
|
|
|||
| Eigen::Vector6d g_rb; | |||
|
|
|||
| g_rb.topRows(3) = rot * (fg + fb); | |||
| g_rb.bottomRows(3) = center_of_gravity_.cross(rot * fg) + center_of_buoyancy_.cross(rot * fb); | |||
| g_rb.topRows(3) = rot.transpose() * (fg + fb); | |||
There was a problem hiding this comment.
I think there's two potential fixes here:
- The expected transform can be left as is and we can modify the rotation (what you've proposed)
- We update the documentation to state that users should provide the
R_I_Binstead ofR_B_I
I'm currently leaning toward option 2 here, to simplify the implementation and avoid placing additional requirements on the function, but I could be swayed otherwise. What do you think @Eddy-Morgan? If we opt to modify the code, I think we could clean it up a bit to be:
const Eigen::Matrix3d rot_t = rot.transpose()
g_rb.topRows(3) = rot_t * (fg + fb);
// etcThere was a problem hiding this comment.
Both approaches are viable. However, to maintain adherence to the Fossen formulation, I recommend modifying the function to mirror the transformation method detailed in his book. This adjustment ensure consistency with his established standards and enhance the clarity of the code.
The suggested code snippet looks excellent:
const Eigen::Matrix3d rot_t = rot.transpose();
g_rb.topRows(3) = rot_t * (fg + fb);
// etcThere was a problem hiding this comment.
Great lets go with that approach then
|
Looks good! Thanks for contributing @Eddy-Morgan! |
Changes Made
The rotation matrix
rotis now transposed before being applied to the force vectorsfgandfb, ensuring correct force calculations as per guidelines in Thor Fossen's Handbook of Marine Craft.Associated Issues