Skip to content

Expressing frame transformations in code

multiagent-mapping-sheep edited this page Nov 27, 2017 · 1 revision

We will use the minkindr library to represent robot poses and we will follow the notational conventions both in documentation and in code.

The diagrams below specify our notational conventions for vector quantities, rotations, transformations, and points. A coding example with coding standards is at the bottom of the page.

Vector quantities

Rotation matrices

Transformations and points

///
/// \brief A frame transformation example appending an odometric transform onto a robot pose
///
/// Coordinate frames:
///   - W : the world coordinate frame
///   - B : the robot body coordinate frame at time k
///
/// @param T_W_Bk  The robot pose at time k.
/// @param T_Bk_Bkp1 The odometric transform from odometry between time k and time k+1
/// @param out_T_W_Bkp1 The robot pose at time k+1
void frameTransformationExample(const Transformation& T_W_Bk, 
                                const Transformation& T_Bk_Bkp1, 
                                Transformation* out_T_W_Bkp1) {
//
}

Because we are using capital letters for the frame tags, we will reserve R and T as special values that should not be used as frame tags. R will represent rotation matrices (R_A_B) and T will represent transformation matrices (T_A_B).