Digital Signal Processing Reference
In-Depth Information
1FUN
validateClassCand
2I
:
c
// candidate
3I
:
// actor FSM
4 BEGIN
5
create a queue Q
6
enqueue
q
0
on to Q
7
annotate
r
0
to
q
0
8
WHILE
Q is not empty
9
dequeue an item from Q into
q
n
10
FOREACH
transition
t
∈{
t
∈
T
|
t
.
q
=
q
n
}
11
RETURN f
i f
t
does not conform to both transition criteria
12
calculate
r
n
+
1
from
r
n
annotated to
q
n
and
t
13
IF
q
n
+
1
=
t
.
q
has no tuple
r
annotated
14
annotate
r
n
+
1
to
q
n
+
1
15
enqueue
q
n
+
1
onto Q
16
ELSE
17
RETURN f
i f
r
=
r
n
+
1
18 RETURN t
19 END
Fig. 18
Algorithm to validate a
classification candidate c
for the
actor FSM
R
4
Design Methodologies
In this section, we will discuss design flows based on integrated Finite State
Machines (
FSM
) and dataflow modeling. A focus is thereby given to the exploitation
of knowledge about models of computation for analyzability and, hence enabling
subsequent model refinements and optimization of the implementations.
4.1
Ptolemy II
infrastructure for modeling, analysis, and simulation of embedded systems. The
project focuses on the integration of different Models of Computation (
MoC
)by
so-called
hierarchical heterogeneity
. Currently, the supported
MoCs
are continuous
time, discrete event, synchronous dataflow, finite state machines, concurrent sequen-
tial processes, process networks, etc. Ptolemy II supports
domain polymorphism
,
i.e., reuse of actors in different MoCs. The coupling of different MoCs is realized
abstract syntax based on the concept of so-called
clustered graphs
.Aclustered
graph consists of entities and relations. Entities have ports and relations connect to
ports. In general, entities correspond to actors and relations to edges. This enables
the designer to model, analyze, or simulate heterogeneous systems. Each hierarchy
