You are on page 1of 5

Soft computing – Dr.

Sanjoy Mandol

Question 1:
Obtain the desired control force by the aircraft engine for an aircraft flying at a height of 1000 ft with a downward
velocity of -20ft/seconds for safe landing for four continuous seconds using fuzzy logic. BASIC FUNDAMENTAL of landing
principle is: Downward Velocity of aircraft is proportional to Square of height at landing. So, Velocity (v) is desired to be
LARGE at higher altitudes (h) and as height of aircraft reduces with its descend, the desired downward velocity also
reduces. Engine of the aircraft controls the reduction in speed through controlling force (f).
Universe of discourse for height of aircraft, downward velocity of aircraft and required control force of the engine is
given as per following.
height (ft) 0 100 200 300 400 500 600 700 800 900 1000
La rge (L) 0 0 0 0 0 0 0.2 0.4 0.6 0.8 1
Medium (M) 0 0 0 0 0.2 0.4 0.6 0.8 1 0.8 0.6
Small (S) 0.4 0.6 0.8 1 0.8 0.6 0.4 0.2 0 0 0
Nea r Zero (NZ) 1 0.8 0.6 0.4 0.2 0 0 0 0 0 0
Downward velocity (Ft/sec) -30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30
Up large (UL) 0 0 0 0 0 0 0 0 0 0.5 1 1 1
Up Small (US) 0 0 0 0 0 0 0 0.5 1 0.5 0 0 0
Zero (Z) 0 0 0 0 0 0.5 1 0.5 0 0 0 0 0
Down Small (DS) 0 0 0 0.5 1 0.5 0 0 0 0 0 0 0
Down la rge (DL) 1 1 1 0.5 0 0 0 0 0 0 0 0 0
Control force (F-lb) -30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30
Up large (UL) 0 0 0 0 0 0 0 0 0 0.5 1 1 1
Up Small (US) 0 0 0 0 0 0 0 0.5 1 0.5 0 0 0
Zero (Z) 0 0 0 0 0 0.5 1 0.5 0 0 0 0 0
Down Small (DS) 0 0 0 0.5 1 0.5 0 0 0 0 0 0 0
Down la rge (DL) 1 1 1 0.5 0 0 0 0 0 0 0 0 0

Solution:
Step 1: Draw the fuzzy set curves from the (given) universe of discourse for each input and output variables.
height (ft) 0 100 200 300 400 500 600 700 800 900 1000
La rge (L) 0 0 0 0 0 0 0.2 0.4 0.6 0.8 1
Medium (M) 0 0 0 0 0.2 0.4 0.6 0.8 1 0.8 0.6
Small (S) 0.4 0.6 0.8 1 0.8 0.6 0.4 0.2 0 0 0
Nea r Zero (NZ) 1 0.8 0.6 0.4 0.2 0 0 0 0 0 0
1
Fuzzy diagram
0.8 Height (h)
0.6 Large (L)
μ(h)

0.4 Medium (M)

0.2 Small (S)

Near Zero (NZ)


0
0 100 200 300 400 500 600 700 800 900 1000

Downward velocity (Ft/sec) -30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30


Up large (UL) 0 0 0 0 0 0 0 0 0 0.5 1 1 1
Up Small (US) 0 0 0 0 0 0 0 0.5 1 0.5 0 0 0
Zero (Z) 0 0 0 0 0 0.5 1 0.5 0 0 0 0 0
Down Small (DS) 0 0 0 0.5 1 0.5 0 0 0 0 0 0 0
Down la rge (DL) 1 1 1 0.5 0 0 0 0 0 0 0 0 0
1
Fuzzy diagram
0.8 Downward Velocity (v)
Up large (UL)
0.6
μ(v)

Up Small (US)
0.4
Zero (Z)
0.2
Down Small (DS)
0
-30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30 Down large (DL)

17KT000283 IIT-ISM-Dhanbad Sarad a Prasanna Routray sproutray@gmail.com


Soft computing – Dr. Sanjoy Mandol

Control Force (F-lb) -30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30


Up large (UL) 0 0 0 0 0 0 0 0 0 0.5 1 1 1
Up Small (US) 0 0 0 0 0 0 0 0.5 1 0.5 0 0 0
Zero (Z) 0 0 0 0 0 0.5 1 0.5 0 0 0 0 0
Down Small (DS) 0 0 0 0.5 1 0.5 0 0 0 0 0 0 0
Down la rge (DL) 1 1 1 0.5 0 0 0 0 0 0 0 0 0
1
Fuzzy diagram
0.8 Control force (F-lb)
Up large (UL)
0.6
μ(v)

Up Small (US)
0.4
Zero (Z)
0.2
Down Small (DS)
0
-30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30 Down large (DL)

Note: Triangular fuzzy diagram is considered as the data given satisfies triangular curves.

STEP 2: Define Rule base (if not given)


Following is the rule for control force fuzzy logic for set of height and downward velocity of the aircraft.
Downward velocity (v)
Height (h)
DL DS Z US UL
L Z DS DL DL DL
M US Z DS DL DL
S UL US Z DS DL
NZ UL UL Z DS DS

Step 3: Defining of equation for working principle of the aircraft:


The work done equation for the aircraft engine is m*Δv = f*Δt,
Considering change in time as unit (1 sec) and mass as unit mass (m=1), then the work done equation is Δv= f

So, the new velocity = v i+1 = vi + Δv = vi + f


And the updated height = h i+1 = hi + vi *Δt

Step 4: Analysis of input data with fuzzy diagram for EACH set of curve intersections
GIVEN: Aircraft flying at a height of 1000 ft with a downward velocity of -20ft/seconds.
For each combination of input curve intersections, one output curve needs to be drawn by r eferring above fuzzy set
curves and MINIMA of input curve intersection needs to be considered.
Large (L) and at 1000 ft, value is 1 from the curve
h = 1000 ft, satisfies fuzzy set
Medium (M) and at 1000 ft, value is 0.6 from the curve
Similarly
Down Large (DL) and at -20 ft/sec , value is 1 from the curve
Down Small (DS) and at -20 ft/sec , value is 0 from the curve.
v = -20 ft/sec, satisfies fuzzy set
(So this curve is not considered as its value is zero)
From above tables, four set of fuzzy interference curves are to be drawn for median calculations.

NOTES:
1. Minima of both input curves will be considered for output curve value.
2. If any output curve is symmetrical around zero axis, then the area of the curve will not be considered
for median calculations.
3. For median calculations, we will consider the distance of the central line of the output curve from the
zero axis in absolute terms (negative or positive).

17KT000283 IIT-ISM-Dhanbad Sarad a Prasanna Routray sproutray@gmail.com


Soft computing – Dr. Sanjoy Mandol

STEP-5: Input and output curve intersections:


There are total four intersection curves as per above data.

Intersection curve 1: Minima


1 1 1
Fuzzy diagram
0.8 0.8 0.8
Height (h) Downward Velocity (v) Zero (Z)
force (f)
0.6 0.6

μ(f)
Large (L) 0.6

μ(v)
Down large (DL)
0.4
μ(h)

0.4 0.4

0.2 0.2 0.2

0 0 0
1000
500

700

900
100
200
300
400

600

800
0

10

20
15

25
30
0
5

-25
-30

-20
-15
-10
-20

-10
-30
-25

-15

-5
-5

20
25
10
15

30
0
5
INPUTS OUTPUT
h = 1000 ft v = -20 ft/sec μ(f) will be intersected with Fuzzy diagram of
Large (L) Down Large (DL) Zero (Z)
μ(h) = 1 μ(v) = 1 Minima value = 1 Median axis location from zero axis of output curve = 0
Area under the curve is ZERO, because the curve is symmetrical around ZERO axis.
Remarks
In median calculation, this curve will not participate in numerator.

Intersection curve 2:
1 1 Downward Velocity (v) 1 force (f)
Fuzzy diagram Down Small (DS)
0.8 0.8 0.8
Height (h) Down Small (DS)
0.6 0.6 0.6 μ(f)
Large (L)
μ(v)

0.4 0.4 0.4


μ(h)

0.2 0.2 0.2

0 0 0 Minima
1000
500

700

900
100
200
300
400

600

800
0

-10
-30
-25
-20
-15
-10

-30

-20
-5

10
15
20
25
30

10

20

30
0
0
5

INPUTS OUTPUT
h = 1000 ft v = -20 ft/sec μ(f) will be intersected with Fuzzy diagram of
Large (L) Down-Small (DS) Down-Small (DS)
μ(h) = 1 μ(v) = 0 Minima value = 0 Median axis location from zero axis of output curve = -10
Remarks No area is covered under the curve, as minima is at 0.

Intersection curve 3:
1 1 1 force (f)
Fuzzy diagram Up Small (US)
0.8 0.8 0.8
Height (h) Downward Velocity (v)
μ(f)

0.6 0.6 0.6 Minima


Medium (M)
μ(v)

Down large (DL)


0.4 0.4
μ(h)

0.4

0.2 0.2 0.2

0 0 0
1000
300

800
100
200

400
500
600
700

900
0

-30

-20

-10
10

25

10

20

30
15
20

30

0
5
0
-30

-15
-10
-25
-20

-5

INPUTS OUTPUT
h = 1000 ft v = -20 ft/sec μ(f) will be intersected with Fuzzy diagram of
Medium (M) Down-Large (DL) Up-Small (US)
μ(h) = 0.6 μ(v) = 1 Minima value = 0.6 Median axis location from zero axis of output curve = 10
Remarks

17KT000283 IIT-ISM-Dhanbad Sarad a Prasanna Routray sproutray@gmail.com


Soft computing – Dr. Sanjoy Mandol

Intersection curve 4:
1 1 Downward Velocity (v) 1 force (f)
Fuzzy diagram Down Small (DS)
0.8 0.8 0.8
Height (h) Zero…
0.6 0.6

μ(f)
Medium (M) 0.6

μ(v)
0.4 0.4
μ(h)

0.4

0.2 0.2 0.2

0 0 0 Minima

1000
300

800
100
200

400
500
600
700

900
0

-10
-30

-20

10

20

30
0

-25

-15
-30

-20

-10
-5

15

25
10

20

30
5
0
INPUTS OUTPUT
h = 1000 ft v = -20 ft/sec μ(f) will be intersected with Fuzzy diagram of
Medium (M) Down-Small (DS) ZERO (Z)
μ(h) = 0.6 μ(v) = 0 Minima value = 0 Median axis location from zero axis of output curve = 0
Remarks No area is covered under the curve, as minima is at 0.

STEP 6: Median Calculation:


Note:
1. If minima of any of the output curve is at ‘0’, then ignore the graph for calculation of minima.
2. If any curve is symmetrical around ZERO axis, then curve is not considered in numerator part, where
as its minima value is considered in denominator part only.
3. So, only take the values of minima of each output curve and distance of central axis of the curve
from the ZERO axis for calculation of final output.

∑minima of curves with area x distance of central point of the area from ZERO axis
μ(f) =
∑ minima of curves

NOTE: Intersection curve 2 & 4 will not be considered as no area is under the intersection curve.

1 x 0 + 0 x (-10) + 0.6 x 10 + 0 x 0
μ(f) = = 3.75
1 + 0 + 0.6 +0

STEP 7: Next fuzzy control for after 1 second


Input variables for next second:
h new = 1000 ft + v old = 1000 -20 = 980 ft
v new = -20 ft/sec + f old = -20 + 3.75 = -16.25 ft/sec

Repeating steps 4 to 7 again.

STEP 4: Next fuzzy control for after 1 second


Large (L) and at 980 ft, value is 0.98 from the curve
h = 980 ft, satisfies fuzzy set
Medium (M) and at 980 ft, value is 0.64 from the curve

Similarly
Down Large (DL) and at -16.25 ft/sec , value is 0.58 from the curve
v = -16.25 ft/sec, satisfies fuzzy set
Down Small (DS) and at -16.25 ft/sec , value is 0.42 from the curve.
From above tables, four set of fuzzy interference curves are to be drawn for median calculations.

NOTE: By calculating above values, minima function can be found out very easily in the graph. And these
values will play their roles in DENOMINATOR of median of output curve.

17KT000283 IIT-ISM-Dhanbad Sarad a Prasanna Routray sproutray@gmail.com


Soft computing – Dr. Sanjoy Mandol

STEP-5: Input and output curve intersections:


There are total four intersection curves as per above data.

Critical note point:


All median axis distance values from ZERO axis of output curve will remain same as output curves are not changing.
So in the next steps only calculate their minima value s only.

Intersection curve 1:
INPUTS OUTPUT
h = 980 ft v = -16.25 ft/sec μ(f) will be intersected with Fuzzy diagram of
Large (L) Down Large (DL) Zero (Z)
μ(h) = 0.98 μ(v) = 0.58 Minima value = 0.58 Median axis location from zero axis of output curve = 0
Area under the curve is ZERO, because the curve is s ymmetri cal a round ZERO a xis.
Remarks In median cal cula tion, this curve will not pa rti cipa te in numerator.

Intersection curve 2:
INPUTS OUTPUT
h = 980 ft v = -16.25 ft/sec μ(f) will be intersected with Fuzzy diagram of
Large (L) Down-Small (DS) Down-Small (DS)
μ(h) = 0.98 μ(v) = 0.42 Minima value = 0.42 Median axis location from zero axis of output curve = -10
Remarks No area is covered under the curve, as minima is at 0.
Intersection curve 3:
INPUTS OUTPUT
h = 980 ft v = -16.25 ft/sec μ(f) will be intersected with Fuzzy diagram of
Medium (M) Down-Large (DL) Up-Small (US)
μ(h) = 0.64 μ(v) = 0.58 Minima value = 0.58 Median axis location from zero axis of output curve = 10
Remarks

Intersection curve 4:
INPUTS OUTPUT
h = 980 ft v = -16.25 ft/sec μ(f) will be intersected with Fuzzy diagram of
Medium (M) Down-Small (DS) ZERO (Z)
μ(h) = 0.64 μ(v) = 0.42 Minima value = 0.42 Median axis location from zero axis of output curve = 0
Remarks No area is covered under the curve, as minima is at 0.

STEP 6: Median Calculation:


0.58 x 0 + 0.42 x (-10) + 0.58 x 10 + 0.42 x 0
μ(f) = = 0.8
0.58 + 0.42 + 0.58 + 0.42

STEP 7: Next fuzzy control for after 1 second


h new = 980 ft + v old = 980 -16.25 = 963.75 ft
v new = -16.25 ft/sec + f old = -16.25 + 0.8 = -15.45 ft/sec

And so, on

Cycle 1 2 3 4
h 1000 980 963.75
v -20 -16.25 -15.45
f 3.75 0.8

17KT000283 IIT-ISM-Dhanbad Sarad a Prasanna Routray sproutray@gmail.com

You might also like