Add feature to have multiple transforms when parsing YAML actions

[pfankhauser]

Example:

adapt_coordinates:
  - transform:
      source_frame: footprint
      target_frame: odom
  - transform:
      source_frame: <target_frame>
      target_frame: odom
      transform_in_source_frame:
        position: [<position_x>, <position_y>, <position_z>]
        orientation: [0.0, 0.0, <orientation_yaw>]