State representation of a system made up of subsystems

Up to now, we've only been interested in the state representation of a "black box", i.e. in practice an element of a system known only by its input-output relationship and about which we have no structural information (which is why we had the choice between various possible state representations, obviously all of the same order).

In practice, a system is made up of a number of sub-systems, interconnected in a more or less complex but known way. We therefore have two levels of knowledge: at sub-system level, no structural information, and at system level, structural information.

In such a case, if we look for the global transfer function (or the input-output differential equation), we come up against three difficulties, of different kinds:

• firstly, the calculations can be tedious, especially if you have tangled...