IEEE International Symposium on Industrial Electronics (ISlE 2009)

Seoul Olympic Parktel, Seoul, Korea July 5-8, 2009

Electric Bikes Energy Management-

Game-theoretic Synthesis and Implementation
Boe-Shong Hong* Tsu-Yu Lin and Wen-lui Su
Department of Mechanical Engineering
National Chung Cheng University, Chia-Yi 621, R.O.C (Taiwan)
E-mail: [email protected]

Abstract- Specially for electric bikes, this work develops LPV performance-in-specification and reveals more interaction
H 00-game synthesis to regulate the trade-off between energy among all players.
consumption per distance and propulsion capability in transience. Furthermore, the performance of energy consumption being
Programmed into the embedded controller, the synthe~ize~ specified on per distance renders the generalized plant served
regulator functions as the transmission in transient state, which IS
for control synthesis to be nonlinear. To accommodate this ad
analog to the gear transmission in steady state. Computer
simulations and road tests are provided to verify the proposed hoc vehicular nature, we incorporate the linear parameter
synthesis and implementation of such a new transmission varying control [9-12] into the game-theoretic synthesizer to
technology. obtain the nonlinear feedback. In the implementation of the
Keywords: Electric vehicles, Game theory, Linear Parameter-
nonlinear compensation, we use the gain-scheduled technique
varying systems, H control.
00
by taking the bike speed as the gain-scheduled parameter on-
output from a Luenberger observer. Similar scenarios can be
I. INTRODUCTION found in many engineering practices, such as in turbocharged
diesel engine [13], in wind energy conversion system [14], and
In order to carry out the policy of energy conservation and
in diesel auxiliary power unit [15].
create new industry, Taiwan government draws up the related
Most of power sources available in the industry of electric
law to gives fresh impetus to electric bike industry [1]. The
vehicles suffer low ratio of energy either to cost, such as
industry and academic institutions continues to explore new
lithium-ion battery, fuel cell, high-capacity ultracapacitor, and
ways to produce electric bike that markets accept.
ultrahigh-speed flywheel, or to weight, such as lead-acid
Traditionally, the controller design of electric bikes relied on
battery. To manage batteries to reduce consumption of electric
classical control in frequency domain. A common strategy of
energy, Meissnerm and Richter [16], Plett [17], Rand [18] and
protecting the lead-acid battery from overcharging and the DC
Saakes et. al. [19] on-line measured the state-of-charge and/or
motor overheating is to regulate the current overshoot by
state-of-health of considered batteries, and then regulated
phase-lag compensator [2-4]. For the energy conservation,
current outputs for conserving energy and increasing battery's
customers begin to concern the efficiency of their electric bikes.
lifespan. Karden et. al. [20] and Faggioli et. al. [21] installed
Therefore, control engineers have to satisfy the multi-objective
ultracapacitors between batteries and power converters to
performance not only economy-in-energy but also quickness-
increase the energy transformation in braking phases from
in-motion. Because user can formulate practical needs as linear
mechanical kinetics to electricity.
matrix inequalities freely and numerical solve linear
Restructuring power converters can lead to vital reduction of
programming conveniently, model-based control has been give
electricity-to-heat. For examples, Martinez and Ray [22]
a lot of attention in recent years [5-7].
designed bi-directional DC/DC power converters to "soften"
In this paper, we investigate the interaction between energy
switching of currents, which was later improved with four
utilization and motion responses of electric bike dynamics.
power MOSFETs; Hofsajer et. al. [23] used RCD snubber to
Here, the economy-in-energy (EE) and quickness-in-motion
reduce energy consumption from high-frequency on-offs of
(QM) act two players entering into negotiations with a
power transistors; and He and Nelems [24] embedded
negotiation parameter being assigned for quantification of the
microprocessors into power converters for optimal timing of
EE-QM trade-off of the controlled dynamics. The negotiation
switching to suppress electricity loss.
parameter plays like gear ratio of transmission in transient state,
As for the controller circuits and programs, we are taking the
which is needed for energy economy especially for low-speed
above arguments. As for the structure of the article,
electric vehicles that are used in jammed traffic. The control
achievements of the work are presented as follow. Section 2
synthesis developed here is based on multiple Hamiltonians ~8]:
performs modeling parameterization for EB dynamics with
every player is assigned with a Hamiltonian. Compared with
specific simplifications. Section 3 develops the EE-QM
the single Hamiltonian approach that use one Hamiltonian for
regulator through constructing a generalized plant. Section 4
all players, this synthesis involves less conservatism in

978-1-4244-4349-9/09/$25.00 ©2009 IEEE

2131
presents the implementation of feedback dynamics produced
p
by the EE-QM regulator into PICI6X microcontroller. Also
presented is the procedure to improve controller designs for EB
industry. Section 5 discusses findings of computer simulations
and road tests. In the section 6, we conclude the work. ,~ . ...

II. MATHEMATICAL MODELING ;

Battery EM
The electric bikes (EBs) dynamics served for control
analyses and syntheses is shown in Figure I. The realization of ·············;;·····1f········;································;·············1
the EB dynamics is of two state variables, designated as the
motor current i and the tire speed OJ. The equivalent circuit
consists of the equivalent inertia of moment J , inductance L ,
resistance R and damping B , which are identified in advance
of control /observer syntheses. Chou et. al. [2] devised an .. , - '"1
inverse dynamics equipment to identify those parameters. ,- - - - - - - - - - - - - - - - - - - -, !._-- -- - -- - -- -- - - ------- -- -- -- --- - - - -- -- -- ------ - -- -- -- --- - -- -,

Kirchhoff theorems upon the equivalent circuit in Figure I Fig. I. Equivalent representation for dynamics of electric motorcycles
realizes the state space of the EB dynamics by
di . III. SYNTHESIS OF EE-QM REGULATION
L- == - P KOJ - RI + U
dt (I) We employ the following symbols to construct a generalized
dOJ
J dt"== . B OJ plant served for the synthesis of H co -negotiation-based EE-
pKl-
QM regulator. There are four state variables:
The time t, state x and input u in Eq. (1) are further made I) X l : the motor current i related to the rate of energy
dimensionless by dissipated from the resistor;
t~-,
t OJ~-=,
OJ .
/~-=,
i u
u~-=,
(2) 2) x2 : the tire speed OJ related to the rate of energy dissipated
"m OJ i u from the damper;
where "m stands for the time constant of mechanical parts 3) X3 : the integration of tracking error of motor current
J / B ; Ii for the maximum of average voltage across the el = r - i , which indicates the quickness of acceleration under
motor; and 05 and l for the steady-state tire speed and motor
servocontrol ; and
current, respectively, as the average voltage delivered by the
4) x 4 : the integration of tracking error of tire speed e2 = r - OJ ,
power converter is kept at Ii . The motor voltage Ii is identical
to the voltage across the battery if the power converter of the which indicates the quickness of speed under servocontrol,
controller adopts buck choppers. where r E [0,1] is the referenced commands of motor current
With the dimensionless setting of Eq. (2), the steady state of and also tire speed.
Eq. (I) signifies that pAi = B05 and pK05 + Rl = ii . Therefore, The former two state variables, Xl and x2 , quantify the
the non-dimensional state-space realization of the EB dynamics current rate of energy consumption, and the latter two state
can be rewritten as variables, X3 and X 4 , memorize the past responses of motions.
di = -f3i - rpw+ (13 + rp)u Also defined are two performance variables and two
dt corresponding performance indexes:
(3)
dai .
-=/-w 5) z, = [X3 x4 uY ,the performance variable pointed to QM;
dt 6) z2 = [Xl X2 uY ,the performance variable pointed to EE;
where 13 indicates the mechatronic ratio of AC characteristics,
7) rl: the performance index of QM, being the bound of the
13 == "m/"e, where "eis the time constant of electric parts,
L 2 -gain from the reference r to the QM variable =1 ,
"e= L / R , and tp indicates the mechatronic ratio of power
integration by distance dB == OJ dt ,
ratings , rp == 105 2 / L? These two dimensionless parameters
define the mechatronic property of the considered EB, which fllzlCt)112axtt~ fdCOJ)lPct)12axtt , r(t)E[O,I] , VtE[O,T] ,
should be determined first during the EB design process. It is VT>0(4)
inferred from Eq. (3) that the non-dimensional motor current i and
and tire speed OJ are equal in steady state , valued within [0, I]. 8) r2 : the performance index of EE, being the bound of the
H cc -norm from the reference r to the EE variable =2 ,
integration by distance, dB = adt ,

2132
(11=2 (t)11 2adt ~ fr1 2
(W)lr(t)1 ad: , r(t) E [0,1] , \it E [0, T] ,
Substituting Eq. (10) into Eq. (9) yields the dynamics of
feedback system:
tiT> 0 (5) x =(A - B2K (m))x + BI (m)r
A smaller QM index YI implies that the time for reaching Zl =(CI(m)-DI2(m)K(m))x+Dllr . (11)
referenced speed or acceleration is shorter, that is, the motion z2 = (C2(m)-D22(m)K(m))x+D2Ir
of EB is quicker under servocontrol. A smaller EE index Y2 Based on the Bounded Real Lemma [5], Eq. (4) and Eq. (5)
means that heat dissipated from mechanical and electric hold if there exists a positive-defmite, symmetric matrix X
resistances are less before the motions arrive at steady state. It simultaneously satisfying the following two matrix inequalities:
is to be noted that the QM variable Zl and the EE variable Z2 ax
±[l-+X(A-B2K)T +(A-B2K)X X(Cl - D12K)T Bl
include the motor voltage u to protect it from becoming am (12)
unreasonably large for regulating the EE and QM. In order to (Cl -~2K)X -YII
formulate energy-economic performance being on per distance B[ ~
into the generalized plant, we define new variable for the
generalized performance, Zl =.J{;jZI , Z2 =.J{;jz 2 , and the
(13)
generalized disturbance r = .J{;jr .
With the introduction of new symbols above, the EB
dynamics of Eq. (3) can be extended to the following
generalized plant that incorporates the EE and QM where fJ is the rate bound of the tire speed, known in the
performances: phase of propulsion design. The introduction of a new matrix
0 0 0
Y by Y = KX transforms Eq. (12) and Eq. (13) into two linear
Xl -fJ -ffJ Xl fJ+ffJ
matrix inequalities (LMls):
x2 1 -1 0 0 x2 1 0 0
+- r+ u (6)
x3 -1 0 -£1 0 x3 .J{;j 1 0 ±p ax + AX + (AX)T (CIX)T B1
am [ B2Y + (B2Y)T (D 12y )T
x4 0 -1 0 -£2 x4 1 0
c1x 0 D ll < D12Y rl]
T
B1 DI~ 0 0 0 r:I ]
Zj =.J{;j[~ ~ ~ ~]:: +[~}+,~} x4
(7)

±pax +AX + (AX)T (C 2 X )T B1 (D 22 y )T

(14)

am [B2Y +(B 2Yl

c,» 0 D2 1 < D22Y

.J{;j[~ ~
r2]

Z2 =
o 0
:
00
:]::
3
x4
+
[:]r + .J{;j[:]u
0 1
(8)
T
B1 Drl 0 0

(15)
Any quadruple (X, Y, YI' Y2) satisfying Eqs. (14)-(15)
0 r:J

where we choose small values £1 and £2 to replace zeros so defmes a feasible feedback K = YX- l guaranteeing the
that H -control syntheses escape from numerical singularity.
00
performances of feedback dynamics to have QM index YI and
Let us briefly write Eqs. (6)-(8) as EE index Y2. It is easy to check that the set of feasible pairs of
x = Ax+ BI (m)r + B2 u indexes (Yu Y2) is convex.
Zl = CI (m)x + Dllr + D l2(m)u (9) Based on the convexity mentioned above, we provides a
numerical procedure in the following to represent the set of
Z2 =C2(m)x+D2Ir+D22(m)u
feasible performances (YI' Y2) in a 2D plane coordinated by
and employ the state-feedback
YI and Y2 :
u=-K(m)x (10)
with the feedback gain K(m) being to-be-determined. In the [Step 1] Assign a negotiation parameter a and then set
proposed feedback system, a current senor is installed on the Y2 = aYI in Eqs. (14)-(15). Involve the mincx.m routine in the
board of power converter with its output y being sent to a Matlab LMI-Iab toolbox to solve the optimal solution
Luenberger observer for reconstructing the tire speed and (X,Y,YI,Y2) with a minimized rl' Mark the obtained point
filtering noises out of the measured signals of motor current. (YI' Y2) in the YI - Y2 2D plane, as shown in Figure 2, and the
The Luenberger observer is synthesized based on the EB
dynamics in Eq. (3). Since the time constant of electric parts
corresponding optimal feedback K = YX- l for the a -
negotiation.
t e is much smaller in practice than that of mechanical parts r m ,
the observed tire speed is taken as a really measured one.

2133
[Step 2] Choose a set of negotiation parameters a and repeat Check the resulted EE and QM performances to see the need of
Step 1 to obtain the corresponding set of optimal points repeating Step 2 with a new negotiation parameter a .
(Yl ' Y2) in the Yl - Y2 20 plane. Connect those points into a [Step 3] Implement the feedback dynamics into a PIC
curve of optimality. This curve divides the Yl - Y2 20 plane microcontroller
We implement the gain-scheduled feedback dynamics into a
into two regions: feasible performances and infeasibility.
PIC 16F87X chip. Peripheral Interface Controller (PIC) has
For standard mechatronic ratios of AC characteristics
suitable cost and complexity for EB control, and has embedded
f3 =150 and power ratings rp =325, the overall solutions of Eqs.
analogue-to-digital (AID) and pulse width modulation (PWM)
(14)-(15) obtained through Steps 1 and 2 are demonstrated in
modules, allowing for inputting analogue signals from sensors
Figure 2. Figure 2 represents in a non-dimensional fashion the
and outputting PWM signals to switching power converters,
personality of electric motorcycles: the trade-off between EE
respectively. Due to its low memory and data-bits, Finite
and QM. This vehicular personality has been commonly
Impulse Recursive (FIR) is not considered to program the
experienced , and can be parameterised with the H co -
feedback dynamics. Moreover, PIC16 microcontroller has
negotiation -based EE-QM regulator newly developed in this
comparatively simpler structure of internal timers than DSP
paper.
3.45 r -- r----r - -,---r-- --,- , -----,- - ..,.-,- ---.- - -,
microprocessor and embeds AID and PWM modules, so we did
not apply Infinite Impulse Recursive (IIR) for DSP
3.40 implementation , either. Instead, the so-called matrix iteration
! 00 :
time-fixed (MIT) [25] method is developed to implement
3.35 ! control/observer dynamics, which is briefly described as
f Feas ible region
J follows.
EE 3.30 I
'ir---if--------+ a = I.S - - - - f - - - - ----+1 The MIT interprets a linear dynamics as an iterated operation
Index
of matrices:
(y,) 3.25
Xk +l = el>xk + IUk
3.20 Yk = CXk + DUk '

the matrices (eI>, I) of which are calculated by the continuous-

ais
to-digital mfile (c2d.m) with the holder of inputs being zero-
3.10 order and the sampling time of outputs being T, . Note that the
sampling time used in computer simulations is also set equal to
2.2 2.4 2.6 2.8 T, . The time intervals from the instant the input Uk enters the
QM index (y,) AID to the moment the output Yk +l leaves the PWM are held
Fig. 2. Trade-off between EE and QM at a constant speed fixed and equal to Ts ' too. This time-fixed approach renders
IV. CONTROLLER CIRCUIT AND PRO GRAM the real-timed operation to match the computer-timed operation,
therefore achieves robust implementation of feedback
In this section, we present an implementation of control dynamics into microcontrollers even with AID and PWM
rules for transient transmission. Specifically, we can program a modules. The microprocessor between the AID and PWM can
regulator of energy consumption into the microcontroller of the adopt IEEE754 24-bits standard of floating number, which is
EB controller to boost energetic economy. The circuitry and enough for the purpose of EB control.
programming of EB controllers install the EE-QM regulations. [Step 4] Include the microcontroller into an EB controller
It can be used in EB industry, and is prepared for experiments circuitry
of this work, too. The procedure to make the underlying EB Figure 3 shows main components of an industrial EB
controller is described as follows: controller. On the circuit board are (1) a current sensor; (2) a
[Step I] Identify equivalent parameters of EB PIC chip embedded with AIDs, PWMs and the MIT programs;
With supported materials such as [2-3], identify the (3) voltage regulators; (4) switching buck choppers ; (5) sensor
equivalent inductance L , resistance R and torque coefficient and transducer circuits; (6) the amplification circuits driving
K of electric parts as well as the equivalent damping Band the gates of power MOSFETs in the choppers; and (7) starting
inertia of moment J of mechanical parts. and braking relays.
[Step 2] Parameterize EE-QM regulation [Step 5] Perform road tests
Choose a proper H co -negotiation parameter a in the EE- Firstly, the circuit schema for the system diagram is plotted
QM regulator, and then calculate the corresponding feedback in a circuit-analysis application package for numerical tests.
controller K attached with the pair of performance indexes Then, a corresponding circuit board of the controller is made in
(Yl' Y2) . Plug the controller K into the EB dynamics in Eq. (1) Lab for instrumental tests. Following the successful schematic,
that has been parametrically identified in Step 1, and then a printed circuit board (PCB) can also be made in Lab for
perform computer simulations with the sampling time T, . performing road tests. Finally, the responses of road tests are
collected and compared with that of computer simulations in

2134
Step 2. Iteration of Step I to Step 5 is generally needed for current sensor is of wide bandwidth and, as of such, lures high-
arriving at a satisfied design of EB controller [2]. frequency noises into the loop. Though noises contaminate
measurement, we can discern the coincidence between
Command measured responses and simulated responses , which validates
si nal the accuracy of computer simulations upon the reduced EB
Lead-Ac id dynamics with identified parameters for studying the
Batte
personality of the considered EB.
It is computationally convenient to change the independent
variable of the EB dynamics in Eq. (3) from the time to the
angular displacement B (dB / dt = OJ ):
di i u
- = -fJ--rp+(fJ+rp)-
de 0) 0)
(16)
dO) =~-1
de 0)

We simulate based on Eq. (16) the responses of energy

consumption, motor current and tire speed versus the angular
displacement B of the EB. Figure 5 shows situations in a
jammed traffic. Their responses of energy consumption
indicate that the economic type (a = 1.1 ) saves about 25%
Con tro ller
circuit energy with respect to the quick type (a = 2.0). Energy is
saved because of reduced electric dissipation resulted from the
Signa l depression of current overshoots during the acceleration and
deceleration . This also protects the lead-acid battery from
Low power overcharging and the DC motor from overheating, and thus
High power saves energy. That is out of the scope of this paper.
In Figure 6, we plot the relationship between the amount of
Fig. 3. Means plus functions of the EB controller
energy being saved and the negotiation parameter chosen in the
V. RES ULTS OF COMPUTER SIM ULATIONS AND ROAD T ESTS EE-QM regulator. It is found that the percentage of energy
saved with respect to the quick type (a = 2.0 ),77, is almost
We purchase an electric bike for experiment, keeping its DC
proportional to the reciprocal of negotiation parameter 1/ a .
motor and lead-acid battery ( Ii = 24 V; I = 20 A;
This linear relationship verifies the practicability of taking
m= 3200 rpm) and taking away its controller. An improved negotiation parameters to parameterise the trade-off between
controller is designed and manufactured based on the Steps 1-5
EE and QM in EB dynamics.
described in Section 4, which installs an ACS754SCB-200
Hall-effect current sensor/transducer (single bias; 0-1OOA input; VI. CONCL USION
0-5V output). The dimensions and equivalent parameters of the This work contributes an example to the academic
experimental EB are identified. A toy DC-motor is pivoted on realization of our common experience that quicker motions
the axial of the EB motor for sensing tire speeds, and a Laptop lead to larger consumption of energy. To the industry of
Dspace is carried on the running EB for the real-timed signal electric bikes, this work provides a motion-energy regulator to
acquisition of motor currents and tire speeds. The signal achieve the transient transmission of electric bikes.
conditioning inside the PICI6F87X chip is as follows: (I) the
command signal OV-5V corresponds to the motor current OA-
20A; (2) the output of current sensor OV-5V corresponds to the R EFERENCES
motor current OA-IOOA; (3) the frequency of the PWM is 20k [1] "Cabinet submits guidelines ," Taipei Times, Feb. OS, pp. 3, 2009.
Hz with its duty cycle 0%-100% being proportional to motor
[2] M.-F. Chou, B.-S. Hong, and T.-Y. Lin, "Inverse Dynamic Equipment and
voltage OV-24V; and (4) the sampling time in the MIT Controller Design for Electric Bikes," Journal ofScience and Technology
program is T, = 0.01 s. (Mandarin) , vol. 15, No.3 , pp. 167-175,2006.
Firstly, we choose two types of feedback dynamics to check [3] M.-F. Chou, A Driver Design f or High Lifespan Electric Vehicles, Master
the ability of the developed EE-QM regulator to quantify the Thesis, National Chung Cheng University, Taiwan, 2005.
trade-off between QM and EE. One is an economic type with [4] B.-S. Hong, Cr-C, Huang and C.-II. Chiu, "Linear Quadrati c Energy-
the negotiation parameter being a = 1.1 and corresponding Motion Regulators of Electric Motorcycles," 17th IFAC World Congress,
Seoul, Korea, July, 2008, pp.6396-6401.
EE-QM indexes (Yl, Y2) being (2.80,3.08), and the other is a
[5] P. Gahinet and P. Apkarian, "A Linear Matrix Inequality Approach to }f cc
quick type with the negotiation parameter being a = 2.0 and control," International Journal of Robust and Nonlinear Control, Vol. 4,
corresponding indexes (Yl,Y2) being (1.71,3.42). Figure 4 pp. 421-448,1994.
records the responses of road tests of these two types of [6] P. Gahinet and A. Nemirovskii, The LMI Lab, LMI-Lab Toolbox for Usc
controllers. We find that Allegro ACS754SCB-200 Hall-effect with Matlab, Chap. 8, The MathWorks inc., Natick, Massachusetts , 1994.

2135
[7] A. Sciarretta and L. Guzzella, "Control of Hybrid Electric Vehicles ' , IEEE
Control System Magazine," pp. 60-70, April 2007. 400 ,--,----,-----,-----,-------,---,-----,-----,,------,

[8] B.-S. Hong, "Robust Management of Advertising Game Equilibrium,"

Proceedings of the American Control Conference, Minneapolis,
Minnesota, 2006, pp. 2250-2253 .
[9] B.-S. Hong and A. Ray, "Observer-embedded L2-gain Control ," Applied
Mathematics Letters, Vol. 14, pp. 563-569,2001.
[10] B.-S. Hong, A. Ray, and V. Yang, "Wide-range Robust Control of 80 100 120 140 160
Combustion Instability," Combustion and Flame, Vol. 128, pp. 242-258,
2002. 80 , -- ,----- ---,-- ----,-- ---,- - ,---- ---,--- ---,-- ---.----.
[l l ] B.-S. Hong, "Observer-based Parameterized LPV L2-gain Control
Synthesis," Proceedings of American Control Conference, Anchorage, 60 Qui ck type
Alaska, May 2002, pp. 4427-4432 .
112] B.-S. Hong, "Output Feedback Linear Parameter Varying (LPV) L2-gain
Control," Journal of Dynamical Systems, Measurement, and Control, Vol.
125, pp. 485-489 , 2003.
[13] M. Jung and K. Glover, "Calibratablc Linear Parameter-varying Control
of a Turboeharg ed Diesel Engine Control," IEEE Transactions on
Systems Technology, Vol. 14, pp. 45- 62, 2006 . Time (sec)

[14] F.D. Bianchi, R. J. Mantz, and C. F. Christiansen, "Gain Scheduling Fig. 4. Measured responses of economic type and quick type
Control of Variable-speed Wind Energy Conversion Systems using
Quasi-LPV models," Control Engineering Practice, Vol. 13, pp. 247-255 , 10 ~ _. ' ' ~
2005. Quick type Economic ~------------- -.-- _
Energy 5 ....~.-.-_~ ~-----••---.-.---------
[15] B. He and M. Yang, "Robust LPV Control of Diesel Auxiliary Power
Unit for Series Hybrid Electric Vehicles," IEEE Transactions on Power
Electronics, Vol. 2 1, pp. 791- 798, 2006. 2 3 4 5 6 7 8 9 10

fl6] E. Meissnerm and G. Richter, "Battery Monitoring and Electrical energy

Management Precondition for Future Vehicle Electric Power Systems,"
Journal of Power Sources, Vol. 116, pp. 79-98, 2003 .
[17] G. L. Plett, "E xtended Kalman Filtering for Battery Management Systems
00 3 4 5 6 7 8 9 10
of LiPB-based HEV Battery Packs: Part I. Background," Journal of
Power Sources, Vol. 134, pp. 252-261, 2004. 2

[18] D. A. J. Rand, "The Lead/Acid Battery-a Key Teehnology for Global Current
1 '~'-\l
',...~ ~
~--'~'
'\ '~... ~
~---'0 I~J--\1
Y'.....,l !l \', I
Energy Management," Journal of Power Source, Vol. 64, pp. 157-174, o l /Economic typ - "ick type --.)./ V
1997.
-10 2 3 4 5 6 7 8 9 10
fl9] M. Saakes, R. Woortmeijer and D. Schmal, "Bipolar Lead-Acid Battery
for Hybrid Vehicles," Journal of Power Sources, Vol. 144, pp. 536-545, Range
2005.
Fig. 5. Motion responses and energy consumption in a jammed traffic
[20] E. Karden, S. Ploumen, B. Fricke, T. Miller and K. Snyder, "Energy
Storage Devices for Future Hybrid Electric Vehicles," Journal of Power
Sources, Vol. 168, pp. 2-11, 2007 .
[21] E. Faggioli, P. Rena, V. Danel, X. Andrieu, R. Mallant, H. Kahlen,
"S upercapacitors for the Energy Management of Electric Vehicle,"
Journal ofPower Sources, Vol. 84, pp. 261-269, 1999.
[22] Z. R. Martinez, and B. Ray, "Bi-directional DC/DC Power Conversion
using Constant Frequency Multi-resonant Topology," Applied Power Percentage o f
Electronics, Vol. 2, pp. 991-997 , 1994. ene rgy saved
[23] I. W. Hofsajer, J. A. Ferreira, J. D. van Wyk, and M. F. K. Holm, "A 'I
Planar Integrated RCD SnubberNoltage Clamp," IEEE Industry
Application Magazine, Vol. I, pp. 24-29,1995.
[24] D. He and R. M. Nclems, "Peak Current-mode Control for a Boost
Converter using an 8-bit Microcontroller,' IEEE 34th Annual Conference
on Power Electronics Specialist, Vol. 2, pp. 938-943 , 2003 .
[25] H.-J. Luo, A Digital Signal Processor System for PICI6F87X Chip , 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2
Master Thesis, National Chung Cheng University, Taiwan, 2005.
Negotiation parameter a

Fig. 6. Energy saved versus negotiation parameter

ACKNOWLEDGMENT

We thank National Science Council of Taiwan for its

financial support for one year under the project number NSC
97-2221-E-194-037 .

2136