Recently I've started to read about BMS (Bondi-Metzner-Sachs), and I've encountered several statements such as the following (from ).
[I]t turned out that the asymptotic symmetry group at null infinity in four dimensions is not the Poincaré group, but an enhanced group where translations are replaced by supertranslations.
My question: In what sense is this BMS a symmetry? I know that Poincaré is a symmetry in the sense that the fundamental equations of motion of the fields are kept invariant (look the same) once all the objects in the equations are transformed by Poincaré. Is that the same sense in which BMS is a symmetry, when we consider only fields at null infinity?