Professional Documents
Culture Documents
on steady-state residual
convergence ratio
Abstract
In the feedback control robotic systems, the repetitive control method has a high control performance for the track or
elimination of the periodic signals. The promotion of the plug-in type configuration of the controller broadens the application
range and applicability of the control method. In this article, a novel design algorithm based on the steady-state residual
convergence ratio of the repetitive control system is proposed to improve the performance of the stabilized platform to
resist the periodic perturbation. The basic structure and stable condition of the plug-in type repetitive control method are
first introduced by applying the small gain theorem and the stability theorem for time-lag systems. Then the analysis of the
convergence rate is utilized in constructing the basic index of the design algorithm of a plug-in type repetitive control system
based on a steady-state residual convergence ratio. The parameters of the designed controller are checked by the validity
condition of the plug-in type repetitive control system, and a simulation example is given to verify the effectiveness of the
design algorithm. The article provides basic design guidelines and schemes for the design of the periodic disturbance sup-
pression performance of the feedback control system. In the final physical prototype experiment, the prospective steady-
state residual convergence ratio is basically achieved within the allowable range of error.
Keywords
Repetitive control, steady-state residual convergence ratio, convergence rate, periodic perturbation
Creative Commons CC BY: This article is distributed under the terms of the Creative Commons Attribution 4.0 License
(http://www.creativecommons.org/licenses/by/4.0/) which permits any use, reproduction and distribution of the work without
further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/
open-access-at-sage).
2 International Journal of Advanced Robotic Systems
Im 1 QðsÞeLs
X ðsÞ ¼ ð9Þ
1 A0 ðsÞQðsÞeLs
a (s =1
( where X ðsÞ represents the effect caused by the PITRC
compared with the original stable system. Since the refer-
Re ence signal is supposed to be periodic signals, the X ðsÞ will
–1
be examined only at the frequency of k!L which represents
the frequency of the reference signal and its multiplier. To
Q ( jw
(
ensure that plug-in repetitive control system is effective at
k!L , jX ðjk!L Þj 1 should be satisfied which is described
as effectiveness conditions in this article.
Figure 3. Stability region of the PITR. PITR: plug-in type repeti-
tive control system. According to jX ðjk!L Þj 1, we obtain
1 Qðk!L Þ
1 A ðk! ÞQðk! Þ 1
0 L L
dB
|I + G(s)|
) jQðk!L Þj2 fRe2 ½A0 ðk!L Þ þ Im½Qðk!L Þ2 g
fRe½Qðk!L Þ2 þ Im½A0 ðk!L Þ2 g þ 2Im½Qðk!L ÞIm½A0 ðk!L Þ
þ 2Re½Qðk!L Þ 2Re½A0 ðk!L ÞRe½Qðk!L Þ
s
0 ð10Þ
Ws O Hz
The derivation process is described in Appendix 1. It
| Q (s) | is shown that the effectiveness of the PITR can be
satisfied as long as the designed low-filter satisfies the
last formula.
Figure 4. The frequency responses of jI þ Gðj!Þj and QðsÞ.
tion of the steady state of the first k-1 time domain func-
•
convergence ratio
1
tions and the transient response of the kth function during
the kth repetitive control period.
According to (b), we have jjQF jj1 ¼ jjQðsÞA0 ðsÞjj1 < 1. •
2
ð13Þ ss
Since the reference input or the perturbation is supposed where hk js¼j!L represents the convergence residual ratio of
to be periodic signal, the real-time convergence residual the kth repetitive control period.
ratio of the residual tracking error to the reference signals Note that ½A0 ðsÞQðsÞkþ1 ! 0 is tenable according to
of the kth repetitive control period is given by jjQF jj1 ¼ jjQðsÞA0 ðsÞjj1 < 1. Therefore, hk ! hss exists
E ðsÞ when k ! 1. Figure 5 depicts how the value of each resi-
k
hk ¼ dual convergence ratio changes on the controlling process.
Rk ðsÞ s¼j!
It implies that the real-time residual convergence ratio of
1 the kth repetitive control period is close to the steady-state
¼ A0 ðsÞ ð16Þ
1 A0 ðsÞQðsÞeLs residual convergence ratio of the system as the repetitive
n control period increases, which verified the convergence of
X
k 1
þ ð1ÞQðsÞ½A0 ðsÞQðsÞ eðnþ1ÞLs the PITR indirectly. At the same time, it reveals how fast
n¼0 s¼j! the convergence rate can be with different control
parameters.
6 International Journal of Advanced Robotic Systems
First-order low-pass filter. The structure of the first-order low- b1 ¼ 2!L ðhss 0 Þ 2 ImðA0 Þ
pass filter is given by c 1 ¼ ½2!L ðhss 0 Þ 2 ReðA 0 Þ 2!L jA0 j 2
e Lq s d 1 ¼ ðhss 0 Þ 2 jA0 j2 2jA0 j2 þ ðhss 0 Þ 2
QðsÞ ¼ ð19Þ
tq s þ 1 e 1 ¼ ½2jA0 j 2 2ðhss 0 Þ 2 ReðA 0 Þ
t q j!s þ 1 eL q s
QðsÞ ¼ s2 2& q s
ð24Þ
þ þ1
jcosðLq !s Þ þ jsinðLq !s Þj !q2 !q
¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðt q !s Þ 2 þ 1 where !q represents the undamped natural frequency of the
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi low-pass filter, & q represents the damping ratio, and L0 q
ð20Þ
½cosðLq !s Þ 2 þ ½sinðLq !s Þ2 represents the lead compensation coefficient used to com-
¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
pensate the lag caused by the denominator. To obtain the
ðt q !s Þ 2 þ 1 robust frequency domain characteristics, generally & q is
1 selected to be 0.707.
¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi < W s Since the damping ratio is set to 0.707, jjQðsÞjj1 1 is
ðt q !s Þ 2 þ 1
established for any !q and L0 q . The upper bound of the
undamped natural frequency can be obtained through
After transforming the formula of (20), we obtain
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi jQðj!s Þj < W s . Then we can get
1 1 vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
tq > 1 ð21Þ u sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
!s W s 2 u W 2 ð1 2& 2 Þ W s 2 ð2& q 2 1Þ2 1
!q < !s t
s q
2
þ W s 2
2
We can get the range of the value of the time constant t q ðW s 1Þ 2
ðW s 1Þ ðW s 1Þ
by (21). Noted that t q has a lower bound that reveals that ð25Þ
the non-damp natural frequency of the filter has an upper
bound. To ensure the stability of the PITR, the cut-off To ensure that the sensitivity function gets the minimum
frequency of the filter should not reach infinity. In view value right at the reference frequency, the selection of the
of the good linearity of the phase-frequency characteristic leading compensation coefficient of the second-order filter
of the first-order filter, the lead compensation coefficient should be adapted to the lag phase of the denominator.
can be selected as pt q =4 in general. Calculate L0 q based on !q got at step (3) or use 2& q =!q as
When designing based on the steady-state residual conver- a rough estimate of L0 q .
gence ratio of the system, it can be obtained from (15) that By solving (26), we can get the equivalent result of the
designed cut-off frequency with (25). The derivation pro-
Eðj!Þ
hss 0 ¼ ¼ A0 ðsÞ 1 QðsÞ ð22Þ cess is described in Appendix 3. The purpose of the trans-
Rðj!Þ s¼j!L 1 A 0 ðsÞQðsÞ s¼j!L
formation is to simplify the process of calculation and make
where hss 0 represents the expected steady-state residual the equation that can be solved in Matlab.
convergence ratio.
a2 !q 2 b2 !q 2 cosð2& q !L =!q Þ c 2 !q2 sinð2& q !L =!q Þ
By solving (23), we can get the equivalent result of the
designed cut-off frequency with (22). The derivation pro- þd 2 !q sinð2& q !L =!q Þ þ e 2 !q cosð2& q !L =!q Þ þ f 2
cess is described in Appendix 2. The purpose of the trans-
þg 2 cosð2& q !L =!q Þ þ h2 sinð2& q !L =!q Þ ¼ i 2 =!q 2
formation is to simplify the process of calculation and make
the equation that can be solved in Matlab. ð26Þ
Jiang and Hong 7
where
a2 ¼ ½2jA0 j 2 hss 0 jA 0 j 2 hss 0
0.1 = 24.74 Hz
b2 ¼ ½2jA0 j 2 þ 2hss 0 ReðA0 Þ s
c 2 ¼ 2hss 0 ImðA 0 Þ 0
0 2
d 2 ¼ ½4& q !L hss ReðA 0 Þ 4& q !L jA0 j Ws = – 0.1793dB
-0.1
e 2 ¼ 4& q !L hss 0 ImðA0 Þ
f 2 ¼ ½4& q 2 !L 2 jA0 j 2 4& q 2 !L 2 hss 0 2!L 2 jA0 j 2 2!L 2 hss 0 -0.2
30 60 90 120
g 2 ¼ ½2!L 2 jA0 j 2 2hss 0 !L 2 ReðA0 Þ
25
h2 ¼ 2!L 2 hss 0 ImðA0 Þ
i 2 ¼ hss 0 !L 4 !L 4 jA0 j 2 20
15
Magnitude(dB)
10
The design algorithm for the PITR based on a steady-
state residual convergence ratio 5
0
Residual convergence ratio
0.04 H (s)
-1
-2
10 11 12 13 14 15
Figure 9. Experimental model of PITR. PITR: plug-in type repe-
0 titive control system.
0 5 10 15 Experimental setup
Time(s) Figure 9 depicts the experimental model for the PITR, where
HðsÞ and GðsÞ represent the PITRC and the open-loop trans-
Figure 7. Residual convergence ratio. fer function of the stable loop, respectively. In the experiment,
the input command is set to zero and the perturbation dðtÞ of 1
-1
Hz or 2 Hz is imposed to the platform. The platform is stabi-
10
Original lized under the control of the PITRC, and its residual motion
PITR error eðtÞ is recorded by the gyroscope whose measured
amplitude reflects the stability precision of the platform and
-2
10 the control accuracy of the PITRC. In addition, the value of L
is the reciprocal of the frequency of the perturbation signal
.The lead compensation coefficient L0 q and the cut-off fre-
X(f)
-3
quency !q are the controller parameters to be designed.
10
The experimental device is shown in Figure 10. The
swing table provides a sine swing motion that can cause
a perturbation of the same frequency to the stable platform.
-4
10 The DC power supplies power to the gyro, the driver, and
0 1 2 3 other instruments on the platform. The dSPACE (Germany)
Frequency(Hz) module with the sampling time of 0.001 s is used to acquire
the sensor signals and achieve digital to analog conversion.
Figure 8. Spectrum of the residual convergence ratio. The computer and its display are helpful to the realization
of the controller and the operation of the experiment. The
Figure 8 shows the comparison of spectrums of the resi- gyro is used to sense the angular velocity of the platform
dual convergence error before and after the insertion of the around the pitch axis in the inertial space.
PITRC. As we can see from the figure, the origin amplitude Based on the identification model whose recognition
is 4.7 102 and then reduces to 1.867 104 in the PITR. accuracy is 87%, the PITRC of the physical prototype sys-
The decreasing of value by two orders of magnitude not only tem is constructed using the simulation parameters calcu-
shows the correctness of the design algorithm based on lated in the third section. The stability of the system will be
steady-state residual convergence ratio but also proves that verified by observing the steady-state accuracy of the sys-
the PITRC is much more effective than the traditional pro- tem under the external perturbation of 1 Hz and the correct-
portional–integral controller in suppressing the perturbation. ness of the proposed design algorithm will be verified by
calculating the steady-state residual convergence ratio of
the system at the frequency of 1 Hz.
Experiment
To verify the correctness of the design algorithm proposed Validation experimental results
in the third section, a physical prototype system is set up to To demonstrate the perturbation of 1 Hz dðtÞ as shown in
verify the simulation results. A stabilized platform is uti- Figure 9, the swing table is set to perform a rocking motion
lized as the control object and a swing table is used to of 1 Hz with an amplitude of 1 . Then, the pitch axis of the
produce the perturbation to the platform. Perturbation– stable platform is subject to a perturbation of the same
rejection experiments are designed to realize the certifica- frequency. Since the perturbation caused by the swing table
tion of the simulation example and to explore the is not a single-frequency disturbance and is easy to derive
Jiang and Hong 9
Perturbation
Platform
Driver
Swing table
DA out
AD in
DC power
The computer
Dapace1103
domain are shown in (b), (d), and (f) and the views of 0–2 Hz 3
are enlarged in the dashed boxes. The smooth residual signals
2
reveal that the basic stability of the system under the control of
the PITRC can be guaranteed. By comparing the frequency 1
spectrums between Figures 11 and 12, it can be found that the
perturbation is greatly attenuated which reveals an excellent 0
perturbation rejection performance of the PITR. 0 10 20 30
Then, the amplitudes of 1 Hz in the frequency spectrum of Frequency(Hz)
the steady-state residue signal of each experiment are recorded
in Table 1, which is used for calculating the steady-state resi- Figure 11. The time domain and frequency domain characteris-
dual convergence ratio of the system under the PITRC. tics of the perturbation.
10 International Journal of Advanced Robotic Systems
(a) 6 (b) 6 -3
(c) 6
10 x 10
4 5 8 4
Angular velocity(deg/s)
Angular velocity(deg/s)
2 4 6 2
4
X(f)
0 3 0
2
-2 2 0 1 2 -2
-4 1 -4
-6 0 -6
0 2 4 6 8 0 10 20 30 0 2 4 6 8
Time(s) Frequency(Hz) Time(s)
6 4 6
4 2
4 4
X(f)
X(f)
3 3
2 0 2
2 -2 2 0 1 2
0 1 2
1 -4 1
0 -6 0
0 10 20 30 0 2 4 6 8 0 10 20 30
Frequency(Hz) Time(s) Frequency(Hz)
Figure 12. Time domain and frequency domain characteristics of steady-state residue. The wave forms of three experiments are
shown in (a),(c) and (e). Their features in frequency-domain are shown in (b), (d) and (f) respectively.
work was supported in part by the National Basic Research Prio- 17. Savkin AV, and Petersen IR. Robust control with rejection of
rities Program of China (grant number A0920132001) and (17- harmonic disturbances. IEEE Trans Automat Contr 1995;
163-11-ZT-005-034-01). 40(11): 1968–1971.
18. Byrnes CI, Gilliam DS, Isidori A, et al. Internal model based
References design for the suppression of harmonic disturbances.
1. Hara S, Yamamoto Y, Omata T, et al. Repetitive control Directions in mathematical systems theory and optimization.
system: a new type servo system for periodic exogenous sig- Berlin, Heidelberg: Springer, 2003, pp. 51–70.
nals. IEEE Trans Automat Contr 1988; 33: 659–668. 19. Lindquist A, and Yakubovich VA. Universal regulators for opti-
2. Li CX, Gu GY, Yang MJ, et al. High-speed tracking of a mal tracking in discrete-time systems affected by harmonic dis-
nanopositioning stage using modified repetitive control. turbances. IEEE Trans Automat Contr 1999; 44(9): 1688–1704.
IEEE Trans Autom Sci Eng 2017; 14: 1467–1477. 20. Marino R, and Tomei P. Robust adaptive compensation of
3. Lin C-Y, and Zeng Y-S. Synergetic repetitive control—a periodic disturbances with unknown frequency. IEEE Trans
plug-in hybrid control method for precision motion control. Automat Contr 2014; 59(10): 2760–2765.
Int J Adv Manuf Technol 2016; 87: 19–24. 21. Bodson M, Jensen JS, and Douglas SC. Active noise control
4. Lai J, Lv P, Liu J, et al. Noncommutativity error analysis of for periodic disturbances. IEEE Trans Control Syst Technolo
strapdown inertial navigation system under the vibration in 2001; 9: 200–205.
UAVs. Int J Adv Robot Syst 2012; 9: 136. 22. Koren Y. Cross-coupled biaxial computer control for manufac-
5. Lan Y-H, and Liu X. Infinite horizon optimal repetitive con- turing systems. J Dyn Syst Meas Control 1980; 102: 265–272.
trol of fractional-order linear systems. J Vib Control 2016; 23. Inoue T, Nakano M, and Iwai S. High accuracy control of
22: 2083–2091. servomechanism for repeated contouring. In: Proceedings of
6. Gao M, Zhang H, Zhou Y, et al. Ground maneuvering target the 10th annual symposium on incremental motion control
tracking based on the strong tracking and the cubature Kal- system and devices. Champaign, Illinois, USA, 1981, pp.
man filter algorithms. Electim 2016; 25: 023006. 285–292. Champaign, Illinois, USA: Incremental Motion
7. Yang G, Yao J, Le G, et al. Adaptive integral robust control of Control Systems Society.
hydraulic systems with asymptotic tracking. Mechatronics 24. Sun J, Gong J, Chen B, et al. Analysis and design of repetitive
2016; 40: 78–86. controller based on regeneration spectrum and sensitivity
8. Perez-Arancibia NO, Gibson JS, and Tsao TC. Observer- function in active power filter system. IET Power Electron
based intensity-feedback control for laser beam pointing and 2014; 7: 2133–2140.
tracking. IEEE Trans Contr Syst Trans 2012; 20: 31–47. 25. Guvenc L. Repetitive controller design in parameter space.
9. Fanzhi CAO, Fanzhi CAO, Shulong ZHU, et al. Tracking In: 2001 Proceedings of the American control conference,
road centerlines from remotely sensed imagery using mean 2001, Vol. 4, 2001, pp. 2749–2754.
shift and Kalman filtering. Acta Geodaetica et Cartographica 26. Mi-Ching T, and Wu-Sung Y. Design of a plug-in type repe-
Sinica. 2016; 45: 205–212. titive controller for periodic inputs. IEEE Trans Contr Syst
10. Li S, and Zhong M. High-precision disturbance compensation Trans 2002; 10: 547–555.
for a three-axis gyro-stabilized camera mount. IEEE Trans
Mech 2015; 20: 3135–3147. Appendix 1
11. Jun-Keun J, and Seung-Ki S. Kalman filter and LQ based speed
controller for torsional vibration suppression in a 2-mass motor
1 Qðk!L Þ
drive system. IEEE Trans Ind Electron 1995; 42: 564–571.
1 A ðk! ÞQðk! Þ 1
12. Mojiri M, and Bakhshai A. An adaptive notch filter for fre- 0 L L
Appendix 2 L0 q s
1 s2 e 2& q s
þ þ1
!q2 !q
A 0 0 ¼ hss 0
1A eL q s
A0 ðsÞ 1 QðsÞ ¼ hss 0
0 s 2 2& q s
þ þ1
1 A ðsÞQðsÞ !q2 !q
0 s¼j!L
2& j! ! ! 2 þ ! 2 ! 2 eL0 q j!L
To simplify the calculation, omit the ðsÞ. q L q L q q
) jA0 j ¼ hss 0
Lq j!L !L 2 þ 2& q j!q !L þ !q2 !q2 A0 eL0 q j!L
1 e
t q j!L þ1 0
) A 0 Lq j!L ¼ hss
1 A0 e
t q j!L þ1
2& j! ! ! 2 þ ! 2
q L q L q
eLq j!L eLq j!L !q 2 ½cosð2& q !L =!q Þ þ jsinð2& q !L =!q Þ
) jA0 j 1 ¼ h 0
1 A
) jA0 j ¼h 0
t q j!L þ 1 ss
t q j!L þ 1
0
2 2 ss
!L þ 2& q j!q !L þ !q
) jA0 jjt q j!L þ 1 eLq j!L j ¼ hss 0 jt q j!L þ 1 A0 eLq j!L j
!q2 ½ReðA0 Þ þ jImðA0 Þ
) jA0 jjt q j!L þ 1 cosðLq !L Þ jsinðLq !L Þj ½cosð2& q !L =!q Þ þ jsinð2& q !L =!q Þ
¼ hss 0 jt q j!L þ 1 A 0 ½cosðLq !L Þ þ jsinðLq !L Þj
) jA0 j2 f½2& q !L !q !q 2 sinð2& q !L =!q Þ2
) jA0 jjt q j!L þ 1 cosðLq !L Þ jsinðLq !L Þj þ½!q 2 !L 2 !q 2 cosð2& q !L =!q Þ2 g
¼ hss 0 jt q j!L þ 1 ½ReðA 0 Þ þ jImðA0 Þ½cosðLq !L Þ ¼ hss 0 f½!q2 !L 2 !q2 ReðA0 Þcosð2& q !L =!q Þ
þ jsinðLq !L Þj þ !q2 ImðA 0 Þsinð2& q !L =!q Þ 2 þ ½2& q !q !L
) jA0 jjt q j!L þ 1 cosðLq !L Þ jsinðLq !L Þj !q2 ImðA 0 Þcosð2& q !L =!q Þ !q2 ReðA0 Þsinð2& q !L =!q Þ2 g
¼ hss 0 jj½t q !L ImðA0 ÞcosðLq !L Þ ReðA 0 ÞsinðLq !L Þ
) !q 2 4& q 2 !L 2 jA 0 j 2 4& q !L !q 3 jA0 j 2 sinð2& q !L =!q Þ
þ 1 ReðA0 ÞcosðLq !L Þ þ ImðA0 ÞsinðLq !L Þj
þ !q 4 jA0 j 2 sin 2 ð2& q !L =!q Þ þ !q 4 jA 0 j 2 þ !L 4 jA0 j 2
2 2 2
) A 0 f½t q !L sinðLq !L Þ þ ½1 cosðLq !L Þ g þ !q 4 jA0 j 2 cos 2 ð2& q !L =!q Þ 2!q 2 !L 2 jA 0 j2
0 2 2
¼ ðhss Þ f½t q !L ImðA 0 ÞcosðLq !L Þ ReðA0 ÞsinðLq !L Þ 2!q 4 jA 0 j 2 cosð2& q !L =!q Þ þ 2!L 2 !q 2 jA0 j 2 cosð2& q !L =!q Þ
þ ½1 ReðA 0 ÞcosðLq !L Þ þ ImðA 0 ÞsinðLq !L Þ 2 g ¼ hss 0 !q4 þ hss 0 !L 4 !q3 4& q !L hss 0 ImðA 0 Þcosð2& q !L =!q Þ
) jA0 j 2 fðt q !L Þ 2 2t q !L sinðLq !L Þ þ 2 2cosðLq !L Þg þ 2!q4 hss 0 ImðA 0 Þsinð2& q !L =!q Þ
¼ ðhss 0 Þ 2 fðt q !L Þ 2 þ jA0 j 2 2t q !L ImðA0 ÞcosðLq !L Þ !q2 2!L 2 hss 0 ImðA0 Þsinð2& q !L =!q Þ
2t q !L ReðA0 ÞsinðLq !L Þ þ 1 2ReðA0 ÞcosðLq !L Þ þ !q2 hss 0 2!L 2 ReðA0 Þcosð2& q !L =!q Þ þ !q 2 4& q 2 !L 2 hss 0
þ 2ImðA0 ÞsinðLq !L Þg þ 2!q4 hss 0 ReðA 0 Þcosð2& q !L =!q Þ þ !q4 jA0 j 2 hss 0
þ 2!q2 !L 2 hss 0 !q3 4& q !L hss 0 ReðA0 Þsinð2& q !L =!q Þ
) t q 2 !L 2 ðjA 0 j 2 X 2 Þ þ 2!L X 2 ImðA0 Þt q cosð!L pt q =4Þ
þ ½2!L X 2 ReðA0 Þ 2!L jA0 j 2 t q sinð!L pt q =4Þ ) !q 4 ½2jA0 j 2 hss 0 jA0 j2 hss 0 ½2!q 4 jA0 j 2
¼ ðhss 0 Þ 2 jA0 j 2 2jA0 j 2 þ ½2jA0 j 2 þ 2!q4 hss 0 ReðA 0 Þcosð2& q !L =!q Þ
2X 2 ReðA 0 Þcosð!L pt q =4Þ 2!q4 hss 0 ImðA 0 Þsinð2& q !L =!q Þ
2 2
þ 2X ImðA0 Þsinð!L pt q =4Þ þ X þ ½!q3 4& q !L hss 0 ReðA0 Þ 4& q !L !q 3 jA0 j 2 sinð2& q !L =!q Þ
þ !q3 4& q !L hss 0 ImðA 0 Þcosð2& q !L =!q Þ
Appendix 3 þ !q 2 ½4& q 2 !L 2 jA 0 j 2 4& q 2 !L 2 hss 0
2!L 2 jA0 j 2 2!L 2 hss 0 þ ½2!L 2 !q 2 jA0 j 2
1 QðsÞ !q2 hss 0 2!L 2 ReðA 0 Þcosð2& q !L =!q Þ
A 0 ðsÞ ¼ hss 0
1 A0 ðsÞQðsÞs¼j!L þ !q2 2!L 2 hss 0 ImðA 0 Þsinð2& q !L =!q Þ