Circuits and Systems, 2016, 7, 718-725
Published Onlin e May 2016 in Sci Res. http://www.sci rp.org/journal/cs
http://dx.doi.org/10.4236/cs.2016.76061
How to cite this paper: Ramesh, P. (2016) Certain Algebraic Test for Analyzing Aperiodic Stability of Two-Dimensional Li-
near Discrete Systems. Circuits and Systems, 7, 718-725. http://dx.doi.org/10.4236/cs.2016.76061
Certain Algebraic Test for Analyzing
Aperiodi c Stability of Two-Dimensional
Linear Discrete Systems
P. Ramesh
Depart ment of Electrical and Electronics Engineering, Anna Uni versi ty , Un i vers ity College of Engineering,
Ramanathapuram, India
Received 21 March 2016; accepted 9 May 2016; published 12 May 2016
Copyright © 2016 by author and Scientific Res earch Publishing Inc.
This wor k is lic ensed under the Creati ve Commons Attribution International License (CC B Y).
http://creativecommons.org/licenses/by/4.0/
Abstract
This paper addresses the new alge b rai c tes t to che ck th e ape riodic st abili ty of tw o dim ensi onal li-
near time invariant disc re te sy stems. Initial ly, the two d imensi on al charac teristics equ ations are
converted into equival ent one-dime nsi onal equati on. Furth er Fuller’s idea is applied on the equi-
valent one-d imensi onal ch ar acte ris tics equ ation. Th en usi ng th e co-efficient of the char acteri stics
equation, the r outh tab le is formed to asce rt ain th e aperi odic st ability of the given tw o-d imen-
sional line ar dis cre te s ystem . The illu stratio ns wer e prese nted to show th e app licabilit y of the
proposed techniqu e.
Keywords
Routh Table, Aperiodic Stability, Characteristics Equation, Two-Dimensional,
Linear Discrete Systems
1. Introduction
Stability is the unique and basic property to be possessed by all kinds of systems. To investigate this property,
various graphical and analytical methods are available. For a given absolutely stable linear time invariant dis-
crete system represented by its characteristics equation f(Z) = 0, with all the roots having z < 1, the aperiodic
stability can be obtained in the given stable system. If all the roo ts are simple and lie on the negative real axis, it
represents ap eriodica lly stable condition. Usin g fuller’s conc ept routh table is formed to check the aperiodic sta-
bility of the system. Information about the aperiodic stability of a control system is of paramount importance for
any design problem. This is generally used for the design of instrumentation systems, network analysis and au-
P. Ramesh
719
tomatic controls. The existence of real and distinct roots in the negative real axis determines the aperiodic beha-
vior of a linear system. The presence of any complex roots shows that the system is aperiodicallly stable. To
analyze the aperiodic stability, a generalized method was inv estigated in the literature by szaraniec. A thre e-step
transformation procedure is presented by fuller, which develops a polynomial whose number of right hand poles
equals the number of complex roots present in the original polynomial.
In the Rout h-Hurwitz stabilit y criterion, the co efficient of the p olyno mial i s arra nged i n two ro ws. W hen n i s
even, the Sn row is formed by coefficients of even order terms (i.e., coefficients of even power of S) and Sn1 row
is formed by coefficients of odd order terms (i.e., coefficients of odd power of S). When n is odd, the Sn ro w is
formed by coefficients of odd terms (i.e., coefficients of odd power of S) and Sn1 row is formed by coefficients
of even order terms (i.e., coefficients of even powers of S). In our proposed scheme, the analysis of aperiodic
stability of a given stable linear discrete system is presented with a help of fuller s e quation a s follows it accounts
all the coefficients of the equivalent one dimensional characteristics equations in the Sn row when n is either
even or odd and Sn1 row is formed using the coefficients by differentiating the equivalent one dimensional cha-
racteristics equation. The other rows of routh array up to S0 ro w can be formed by nor mal Ro uth -Hurwitz stabil-
ity criterion. I llust rative exam ples s how the applicability of the proposed scheme.
2. Literature Survey
The stabilit y analysis is the mo st important test that should be considered in LTIDS analysis and desi gn. There-
fore stability determination is very important. Hence many researches had been conducted to ascertain, the sta-
bility in last few years. A new for mulation of the cr itical constrai ns for stabilit y limits matr ix and its bi-alterna-
tive product was discussed by Jury i n [1]. Bistritz had introduced in [2] the efficient stability test that involves
real univariate polynomials and real arithmetic only. It also examines the doubling degree technique in the de-
velopment of two dimensional stability test discussed by Bistritz in [3]. The possible root locations of two di-
mensional polynomials were proposed by Khargonekerin [4] had proved the critical constrains of obtained from
the polyno mial and also ob tained the s tabilit y within the un it c ircle. A new method to c o mpute the stab ilit y mar-
gin of two dimensional continuous system was provided by Mastorakis in [5] and illustrated and discussed. A
discussion of stability test was obtained for two dimensional and sufficient conditions for asymptotic stability
were easy to checked by Anderson et al. in [6] [7]. Results o btained in [8] proved that double bi-line ar tr a nsfor-
mation does not pr eserve the stability in either direc tion. i.e. Continuous to discrete domain and vice versa pro-
posed by Jury et al. The algorithm was proposed by Bose et al. in [9 ] explains the positive d efiniteness of arb i-
trary quadratic forms that is expressed in terms of inner wise number. The necessary and sufficient conditions
are us ed to form t he inne r of square matrix dis cussed by Jury in [10]. A new algebraic procedure was solving the
prob lem of stability in very low co unt of arit hmetic oper ations was given by B istritz and Ahmed in [11] [12] . A
simplify algebraic equations were Presented by Jury in [13] it was analyzed for its consistency and stability.
Goodman had focused in [14] the suitable difference between the one dimensional and two dimensional test
cases were presented. Two dimensional recursive filtering were proposed by Huang in [15]. It derived and sim-
plified version of stability theorem. The Hurwitz character of the system was determined by Bauer in [16] and
Also proved a relationship between the Hurwitz character of the denominator polynomial of the two dimensional
transfer function. Vimalsingh had proposed in [17] a new criterion for the general asymptotic stability of two
dimensional discrete system discussed by the roesser model using saturation arithmetic was proposed. The im-
plementation procedure for stab ility test for two di mensional digital filter was presented by katbab et al. in [18]
it had reduced the computations, to determine a conservative coefficient space within the coefficient of a real
two dimensional d igital filter. Kamat et al. in [19] had presented the root distr ibution polyno mial with real coef-
ficients with respect to the unit circle is equivalent to one proposed with the same degree of complexity. The
condition for aperiodicity was discussed by szaraniec in [20]. Fullers proposed methodology in [21] to check
whet her a c ontr ol syst em wa s ha ving the de ad-beat conditio n (Aperiod ic stability) a nd revealed that the charac-
teristics equation with real coefficients and can transformed into characteristics equation with complex coeffi-
cients. Later, szaraniec methodology was analyzed and modified by jury in [22] and it exp lains the ne w theo-
rem for the aperiodicity. Jury et al. in [23] proposed a simple stability test similar to routh table was being in-
troduced for linear discrete systems. Bose et al. in [24] presented a procedure to determine whether or not a
polynomial in several real varia bles is globally positive. A method for determining stability b y use of a table
form has been presented by jury in [25], the table had also been used in determining the roots distribution within
the unit circle.
P. Ramesh
720
3. Proposed Method
The t wo di mensional discrete system is repre sented in transf e r fu nctio n [5] by Mastorakis (1998) as,
() ( )
( )
12
12 12
,
,,
AZ Z
HZZ BZ Z
=
(1)
where A and B are non-cancellable polynomials in
1
Z
and
2
Z
. Meth od proposed i n [5] is more com plicated an d
requires more computations to check the aperiodic stability of the given system. In this paper to avoid more
computations a simple algebraic procedure is presented
To convert t wo-dimensional characteristics equation in to one-dimensional c har a c te r istics equation:
In general the following for m of Two-dimensional (2D) [11] equation is chosen:
()( )( )( )
1
12012 1121
,0
nn
n
BZZTZZ TZZTZ
=+++ =
(2)
The reciprocals of
1
Z
and
2
Z
are
and
2
1
Z



respectively are utilized so that the Equation (2) is re-
written as:
1
01
12 12121
11 11111
,0
nn
n
BT TT
ZZ ZZZZZ
 
=+++ =
 
 
(3)
Again Eq uation (3) is represented as,
( )( )
12
12
,0
ZZZ
MZZ fZ
= =
= =
(4)
This
( )
0
fZ=
is one dimensional equatio n and for aperio dic stability
1
Z<
.
Then
( )
fZ
can be analyzed by any algebraic method for checking aperiodic stability of the equivalent one-
dimensional characteristics equation from two-di mensional charac te r istics equation.
Aperiodic Stability Test
If the characteristics roots of
( )
0fZ=
lie in the sector region
01Z≤<
and all are sim ple (distinct) and all the
roots are positive then the system becomes aperiodically stable and the characteristic equation of the system can be
written as,
() ()()
( )
12
0
n
fzzx zxzx
=− −−=
(5)
whe re,
i
x
are distinct roots of
( )
0
FZ=
.
In general, the Equa tion (5) can be arranged as
( )
12
12 0
0
nn n
nn
fzzaza za
−−
−−
= −+−+ =
(6)
The coefficients of
( )
0fZ=
should alternate in sign. It is obse rved in Equation (6).
With
ZZ= −
, the coefficients of Equation (6) will become positi ve in the sector region,
10Z−< ≤
.To test
the aperiod ic stabilit y of a give n line ar t ime-invariant continuous system represented in the for m of its characte-
ristics equation.
( )
12
12 0
0
nn n
nn
fZZaZa Za
−−
−−
= ++++=
(7)
Fuller (1955) formulated [20] a transformed equation of
( )
0Fs=
as,
( )( )( )
2
2
d0
d
ss ss
fs
Fsfss s
=
=
=+=
(8)
Routh’s test referred in Anderson et al. (1987) and Byrne (1975) is applied for the Equation (8). If the first
column of Routh’s table does not possess any sign change, then the system represented by the Equation (8) is
P. Ramesh
721
aperiodically stable. Thus, extending Fuller’s idea, the following transformed equation is written for the Equation
(9) as,
( )( )( )
2
2
d0
d
ZZ zZ
fZ
FZfZZ Z
=
=
=+=
(9 )
The transformed Equation (9) can be handled by Routh’s test verify the sufficiency condition for aperiodic
stability. Thus, above proposed procedure is applied for the following illustrative examples.
4. Illustrations
Example 1: [15]
Consider a two dimensional characteristic equation,
( )
1 2 12
1 0.750.50.3f zzzzz=− −+
Convert the two d imensional characteristic equation i n to single dimensio nal character istic equation,
1 2 12
1 0.750.5.30
0
z zzz−+ =
Taking the inverse of variables
1
z
and
2
z
12 12
0.75 0.50.3
10
zz zz
−−+ =
12
zzx
= =
2
0.75 0.5 0.3
10
xx
x
− −+=
The required one dimensional equation is,
2
1.250.3 0xx− +=
( )
2
1.250.3 0fx xx=− +=
The coefficient of
has alternate in sign; then the necessary condition is sa tisfied.
Then Substitut e
xx= −
in
( )
2
1.25 0.3fxx x−= ++
( )
2 1.25fx x
−= +
Applyin g Fuller’s C onc ept
( )()( )
Fxf xfx
= −+−
( )
( )
( )
2
1.250.321.25Fx xxx=++ ++
Routh Table :
Fro m Table 1, all the elements in the first column are po sitive then the s ystem is aperiod ic a lly stable.
Table 1. Routh table using fullers concept for Example 1.
1 1.25 0.3
2 1.25
0.625 0.3
0.288
0.3
P. Ramesh
722
Output is v erified using MATLAB
()( )()
;;p fxqfxrfx
= =−=−
p = 1.0000 1.2500 0.3000
q = 1.0000 1.2500 0.3000
r = 2.0000 1.2500
sol mat =
2.000 1.2500 0
0.625 0.300 0 0
0.29000 0
0.30000 0
0.29000 0








Example 2: [25]
Consider the following two dimensional characteristic equation
( )
22 2222
221121211212
1 1.20.31.51.80.750.60.720.290f zzzzzzzzzzzzz=−+−+−+−+=
Convert the two d imensional characteristic equation i n to sin gle dimensional characteristic equation,
22 2222
22112121121 2
1 1.20.31.51.80.750.60.720.290zzz zzzzzzzzz−+ −+−+ −+=
Taking inver s e of variables
1
z
and
2
z
,
222222
21 12
212112 12
1.2 0.31.51.80.75 0.6 0.720.29
10
zz zz
zzzzzz zz
−+−+ −+−+=
12
zzx==
22323 4
1.2 0.31.51.80.750.6 0.72 0.29
10
xx
xxxxx x
−+−+− +− +=
The required one dimensional equation is,
432
2.72.71.470.29 0
xxx x− +− +=
( )
432
2.72.71.47 0.29fx xxxx=−+−+
The coefficient of
has alternate sign; then the necessary condition is satisfied.
Then Substitut e
xx= −
in
The transformed equation is formed as,
( )
432
2.72.71.47 0.29fxxx xx−= ++++
( )
32
48.15.4 1.47fx xxx
−=+++
Using Fuller ’s Concept,
( )()( )
Fxf xfx
= −+−
( )
() ()
432 32
2.72.71.470.2948.15.4 1.47Fxfxxxxf xxx
=++++++ ++
Routh Table :
Fro m Table 2, there is two sign change in the first column therefore the given system is unstable.
Table 2. Routh table using fullers concept for Example 2.
1
2.7
2.7
1.47
0.29
4 8.1 5.4 1.47
0.68
1.35
1.10
0.29
0.1
1.11
0.25
8.84 2.77 0.29
1.14 0.25
0.833 0.29
0.15
0.29
P. Ramesh
723
Output is v erified using MATLAB
()( )()
;;p fxqfxrfx
= =−=−
p = 1.0000 2.7000 2.7000 1.4700 0.2900
q = 1.0000 2.7000 2.7000 1.4700 0.2900
r = 4.0000 8.1000 5.40 00 1.4700
sol_mat =
1.0002.7002.7001.4700.2900
4.0008.1005.400 1.47000
0.6751.3501.102 0.29000
0.100 1.1330.248 00 0
0.900 2.780 0.290000
1.164 0.2520000
0.833 0.2900000





−−



−−



Example 3: [15]
Consider a two dimensional characteristic equation,
() ()
1 12
1 0.950.95 0.50z zz
− −−=
Convert the two d imensional characteristic equation i n to single dimensio nal character istic equation,
12 12
1 0.950.950.50
zz zz
−−+ =
Taking inver s e of variables
1
z
and
2
z
,
12 12
0.95 0.950.5
10
zzzz
−−+=
12
zzx==
2
0.95 0.95 0.5
10
x xx
−−+=
The required one dimensional equation is,
2
1.900.5 0xx−+=
( )
2
1.90 0.5fxxx=−+
The coefficient of
has alternate sign; then the necessary condition is satisfied.
Then Substitut e
xx= −
in
The transformed equation is formed as,
( )
2
1.90 0.5fxx x−=++
( )
2 1.90fx x
−= +
Using Fuller ’s co ncept
( )()( )
Fxf xfx
= −+−
( )
( )
( )
2
1.900.52 1.90Fx xxx=++ ++
Routh Table :
Fro m Table 3, all the ele ments in the fir st column of routh ta ble are positive then the s ystem is aperiodica lly
stable.
Output is v erified using MATLAB
( )()()
;;p fxq fxr fx
= =−=−
P. Ramesh
724
Table 3. Routh table using fullers concept for Example 3.
1 1.90 0.5
2 1.90
0.95 0.5
0.848
0.5
p = 1.0000 1.9000 0.5000
q = 1.0000 1.9000 0.5000
r = 2.0000 1.9000
sol_mat =
1.0001.900 0.500 0
2.000 1.90000
0.950 0.50000
0.8474 000
0.500000








5. Conclusion
The main contrib ution of this p a per is to ascertain the aperiodic stability o f t wo dimensional linear time in var iant
discrete s ystems with the help o f Routh table [24] by applying fuller’s concept [21] to form the Sn1 row i n the
Routh table which reduces the order of the characterist ics equation and hence reduces the num ber of comput ations.
The result of the analysis is verified by using MATLAB. In the view of conceptual simplicity of these tests, it
would appear that development of dedicated implementation procedures were simple and worthy subject of fur-
ther investiga tion. T he co ntri buti on made in this paper can be extended to characteristics polynomial containing
complex coefficients.
References
[1] J ur y, E.I. and Gutman, S. (1975) On the Stability of the A Matrix inside the Unit Circle. IEEE Tr an sacti on on Auto-
matic Control, 533 -535. http://dx.doi.org/10.1109/TAC.1975.1100995
[2] Bistriz, Y. (2004) Immitance and Telepolation-Based Procedures to Test Stability of Continuous-Discrete Bivariate
Polynomials. IEEE Transaction on Circuits and Systems , 3, 2 93 -296 .
[3] Bistritz, Y. (2004) Testing Stability of 2-D Discrete Systems by a Set of Real 1-D Stability Tests. IEEE Transact i ons
on Circuits and Systems I: Regu l ar Papers, 51, 1312 -13 20 . http://dx.doi.org/10.1109/TCSI.2004.830679
[4] Khargoneker, P .P . and Bruc e Lee, E. (1986) Further Results on Possible Root Locations of 2-D Polynomials. IEEE
Transactions on Circuits & Systems, 33, 566-569. http://dx.doi.org/10.1109/TCS.1986.1085944
[5] Mast orakis, N.E. (1998) A Method for Computing the 2-DStability Mar gi n . IEEE Transactions on Circuits & Sys-
tems-II, Analog & Digital Sign al Processing, 45, 376-378. http:// dx. d oi.o rg/ 1 0. 1109/ 82 . 66 4243
[6] Anderson, B.D.O. and Jur y, E.I. (1973) A Simplified Schur-Cohn Test. IEEE Transactions on Automatic Control,
AU-31, 157-163. http://dx.doi.org/10.1109/TAC.1973.1100253
[7] Anderson, B.D.O. and Ju r y, E.I. (1973) Stability Test for Two-Dimensional Recursive Filter. IEEE Transactions on
Audio Electroacoustics, AU-21, 366-372. http://dx.doi.org/10.1109/TAU.1973.1162491
[8] J ur y, E.I. and Bauer, P. (1988) On the Stability of Two-Dimensional Continuous Systems. IEEE Transactions on Cir-
cuits and Systems, 35, 1487-1500. http://dx.doi.org/10.1109/31.9912
[9] Bos e, N.K. and Jury, E.I. (1975 ) Inn er Algorithm to Test for Positive Definiteness of Arbitrary Binary Forms. IEEE
Transaction on Automatic Control, 16 9 -170. http://dx.doi.org/10.1109/TAC.1975.1100875
[10] J ury, E.I. (1971) Inners-Approach to Some Problems of Syst em Theory. IEEE Transactions on Automatic Cont rol,
AC-16, 233-240. http://dx.doi.org/10.1109/TAC.1971.1099725
[11] Bistritz, Y. (2001) Stability Testin g of 2-D Discrete Linear Sys tem s by Telepolation of an Immittance-T ype Tabular
Test . IEEE Transactions on Circuits and Systems I: Fundamental Theory and Applications, 48, 840-846.
P. Ramesh
725
http://dx.doi.org/10.1109/81.933325
[12] Ahmed, A. (1980) On the Stability of Two-Dimensional Discrete Systems. IEEE Transactions on Automatic Control,
25, 551-552. http://dx.doi.org/10.1109/TAC.1980.1102352
[13] J ury, E.I. (1970) A Note on the Analytical Absolute Stability Test. Proceeding of the IEEE, 58, 82 3 -824.
http://dx.doi.org/10.1109/PROC.1970.7763
[14] Goodman, D. (1977) Some Stability Properties of Two-Dimensional Linear S hi ft-Invariant Digital Filters. IEEE
Transactions on Circuits and Systems, 24, 201-20 8. http://dx.doi.org/10.1109/TCS.1977.1084322
[15] Huang, T.S. (1972) Stability of Two-Dimensional Recursive Filters. IEEE Transactions on Audio and El ectroacoustics,
20, 158-163. http://dx.doi.org/10.1109/TAU.1972.1162364
[16] Bauer, P. and J ury, E.I. (1988) Stability Analysis of Multidimensional (m-D) Direct Realization Digital Filters under
the Influence of Nonlinearities. IEEE Transactions on Acoustics, Speech & Signal Processing, 36, 1770-1780.
http://dx.doi.org/10.1109/29.9014
[17] S ingh, V. (2012) New Approach to Stability of 2-D Discrete Systems with State Saturation. Signal Processing, 92,
240-247. http://dx.doi.org/10.1016/j.sigpro.2011.07.012
[18] Katbab, A., J ury, E.I. and Mansour, M. (1992) On Robust Schur Property of Discrete-Time P olynomial s. IEEE Trans-
actions on Circuits Systems I: Fundamental Theory and Applications, 39, 467-470.
http://dx.doi.org/10.1109/81.153641
[19] Kamat, P.S . and Zwass, M. (1985) On Zero Location with Respect to the Un it Circle of Discrete-Ti m e Linear S ys tems
Polynomials. Proceedings of the IEEE, 73, 1686-16 87 . http://dx.doi.org/10.1109/proc.1985.13351
[20] Szaraniec, E. ( 19 73 ) Stability, Instability and Aperiodicity Test for Linear Discrete Systems. Automatica, 9, 513-516.
http://dx.doi.org/10.1016/0005-1098(73)90097-6
[21] Fuller, A.T. (1955) Conditions for Aperiodicity in Linear Systems. British Journa l of Applied Physics, 5, 174-179.
http://dx.doi.org/10.1088/0508-3443/5/5/304
[22] J ury, E.I. (1985) A Note on Aperiodicity Condition for Linear Discrete Systems. IEEE Transactions on Automatic
Control, 30, 1100-1101 . http://dx.doi.org/10.1109/tac.1985.1103843
[23] J ury, E.I. and Blanchard, J. (1961) A Stability Test for Linear Discrete Systems in Table Form. Proceedings of the IRE,
21, 1947-1948.
[24] Bose, N.K. and Modarressi, A.K. (1976) General Procedure for Multivariable Positivity Test with Control Applications.
IEEE Transactions on Automatic Control, 21, 696 -701. http://dx.doi.org/10.1109/TAC.1976.1101356
[25] J ury, E.I. (1988) Modified Stability Table for 2-D Digital Filters. IEEE Transactions on Circuits and Systems, 35 , 116-
119. http://dx.doi.org/10.1109/31.1707