11 / 15
Table 6
Values
p
tc
in the “Test” mode
Values
p
tc
at
p
r
= 0
.
1
,
n
= 16
,
b
= 145
и
ε
s
= 1
ε
v
w
v
= 1
w
v
= 2
w
v
= 3
w
v
= 4
w
v
= 5
0
0.135
0.025
0.004
0.0008
0.0001
1
0.375
0.193
0.099
0.051
0.026
2
0.576
0.454
0.358
0.283
0.223
3
0.680
0.633
0.602
0.549
0.512
4
0.717
0.705
0.693
0.681
0.670
5
0.727
0.725
0.722
0.720
0.718
6
0.729
0.729
0.728
0.728
0.728
by 0.079, and the increase from the value (4, 4) to (5, 5), it increases
p
tc
by
0.038, etc. Based on this, we selected
(
ε
v
, w
v
) = (3
,
3)
.
Thus, the selection of parameters
(
ε
v
, w
v
)
for “Test” mode consists of
two steps. First, the required value
R
T
is given and two sets
m
1
, m
2
are
selected according the following conditions:
m
1
=
{
(
ε
v
, w
v
) :
R
(
ε
v
, w
v
)
≤
R
T
}
;
(15)
m
2
=
{
p
tc
{
m
1
}}
=
{
p
tc
1
(
m
1
{
1
}
)
, p
tc
2
(
m
1
{
2
}
)
, . . . , p
tcl
(
m
1
{
l
}
)
}
,
(16)
where
l
is the number of elements
m
1
.
Then, parameters
(
ε
v
, w
v
)
for the “Test” mode are calculated by the
formula
(
ε
v
, w
v
) = arg max
p
{
d
p
:
p
tc
(
ε
v
, w
v
)
∈
m
2
, ε
v
> ε
s
}
,
(17)
where
d
p
=
p
tci
−
p
tci
−
1
,
i
= 2
, . . . , l
.
The results of the experiments with telemetry data are shown in Table 7.
The experiments were performed at
p
r
= 0
.
1
, the selected value
ε
s
, and for
different values of the probability of bit missing
p
d
=
{
0
,
10
−
4
,
10
−
3
,
10
−
2
}
and for different values
ε
v
and
w
v
.
Table 7
Parameters
ε
v
and
w
v
in the “Test” state
Parameters
Theoretical values
p
tc
Experimental values
p
tc
p
d
= 0
p
d
= 0
p
d
= 10
−
4
p
d
= 10
−
3
p
d
= 10
−
2
ε
v
= 2
, w
v
= 2
0.454762
0.7972 0.7777 0.6613 0.1203
ε
v
= 3
, w
v
= 3
0.60248714
0.9308 0.923 0.7916 0.124
ε
v
= 4
, w
v
= 4
0.68165127
0.9876 0.967 0.8112
—
ε
v
= 5
, w
v
= 5
0.718099482
0.9900 0.9752 0.838
—
ISSN 0236-3933. HERALD of the BMSTU. Series Instrument Engineering. 2015. No. 2 127