This step is also demoed with this video :

(ongoing...)

Step 3 : AUTO ASSEMBLY upon not empty architectures

As said in the previous step 2f, it might be useful now to auto assemble upon not empty architecture :

In our example, it has been said that can and analog measurements as well as targets and feedback modules can be organized in a more convenient way :

This is the role of the architect to organize calculus in an architecture.

Assume that the architect organizes a first version :

Architect decides to use can measurements only, ana measurements are decoded and transmitted but not used.

This architecture has been designed in MTX and is available as archi_training_vU0 :

and matches the previous sketch :

Remark : see Step 5 to specifically shows how to design an architecture :

This architecture is made of 4 rooms, i.e. SubSystems aiming to calculate by means of modules :

• targets_

• can_

• ana_

• FB_

The architecture explicitally defines routing between rooms :

• • FB_ uses calculation from targets_ and can_

  • targets_, can_ and ana_ use entries coming from the outside

  • calculations of targets_, can_, ana_ and FB_ are transmitted outside the architecture

As rooms are named by prefix (see step 5), they aim to receive modules of an assembly starting with this same prefix.

As an example, considering the assembly abl_training_vU0, the auto assembly process will allocate modules in rooms according to this figure :

As a result, an auto assembly of abl_training_vU0 upon its attached architecture archi_training_vU0 produces (press AUTO on ABL_training_vU0) :

As we can see, in the room FB_, signals mon_can_LatAcc and mon_can_YawRate are missing, they are then fed by 0 constant (we will get back to that later) :

Then, instead of having 3 assemblies participating :

we can have one assembly used in the this Simulink model :

Of course, the architecture can be modified all along the design process (to deal for example with the missed signals mon_can_LatAcc and mon_can_YawRate ot for any other purposes),

Next sub steps of step 3 illustrate this.

But before rearraging the architecture, one can notice the popupmenu of the attached architecture in the ABL view :

It enables to list modules :

• • outside of the architecture, i.e. modules that can not be inserted in the architecture because of their prefix does not belong to the architecture :

• • in the FB_ room of the architecture :

• • in the ana_ room of the architecture :

• • and inside the whole architecture :

etc...

Published with MATLAB® R2013a