MEV


 

 

J. Mechatron. Electr. Power Veh. Technol. 06 (2015) 113-122 

 

Journal of Mechatronics, Electrical Power, 

and Vehicular Technology 
 

e-ISSN: 2088-6985 

p-ISSN: 2087-3379 
 

 

 

www.mevjournal.com 
 

© 2015 RCEPM - LIPI All rights reserved. Open access under CC BY-NC-SA license. Accreditation Number: 633/AU/P2MI-LIPI/03/2015. 

doi: 10.14203/j.mev.2015.v6.113-122 

COMPARATIVE STUDY BETWEEN INTERNAL OHMIC RESISTANCE AND 

CAPACITY FOR BATTERY STATE OF HEALTH ESTIMATION 
 

M. Nisvo Ramadan, Bhisma Adji Pramana, Sigit Agung Widayat,  

Lora Khaula Amifia, Adha Cahyadi *, Oyas Wahyunggoro 
Department of Electrical Engineering and Information Technology, Universitas Gadjah Mada 

Jalan Grafika no 2 Universitas Gadjah Mada Yogyakarta 55581 Indonesia 
 

Received 26 October 2015; received in revised form 27 November 2015; accepted 01 December 2015  

Published online 30 December 2015 

 

Abstract 
In order to avoid battery failure, a battery management system (BMS) is necessary. Battery state of charge 

(SOC) and state of health (SOH) are part of information provided by a BMS. This research analyzes methods to 

estimate SOH based lithium polymer battery on change of its internal resistance and its capacity. Recursive least 

square (RLS) algorithm was used to estimate internal ohmic resistance while coloumb counting was used to predict 

the change in the battery capacity. For the estimation algorithm, the battery terminal voltage and current are set as 

the input variables. Some tests including static capacity test, pulse test, pulse variation test and before charge-

discharge test have been conducted to obtain the required data. After comparing the two methods, the obtained 

results show that SOH estimation based on coloumb counting provides better accuracy than SOH estimation based 

on internal ohmic resistance. However, the SOH estimation based on internal ohmic resistance is faster and more 

reliable for real application. 

 

Keywords: battery management system; state of health; lithium polymer; recursive least square; coulomb counting. 

 

I. INTRODUCTION 
Global warming is one among hot issues a lot 

of people talk about. In 2013, the average of earth 

temperature has reached 14.6°C, it was up to 

0.6°C compared to the mid-20th century [1]. One 

of the main causes of global warming is the 

excessive use of fossil fuel. As it is known 

number of cars, which are using fossil fuels, 

continues to grow significantly. This case makes 

the demand of fossil fuels increases and is known 

as one major cause of global warming. 

Electric car is one way to reduce fossil fuel 

consumption in order to hold global warming. 

The use of electric car is estimated to increase 

rapidly by 2020 [2]. For this car, lithium battery 

is widely used as the main energy source. Thus 

battery is key to the success or failure of electric 

cars. To avoid battery failure, battery 

management system (BMS) is usually required. 

BMS is a system aimed for regulating the battery 

work in its prime area of operation and providing 

information to the user to perform the necessary 

actions such as stopping the use of battery or 

charging the battery. In this case, BMS optimizes 

the operation of the electric vehicle by knowing 

the capacity of the battery that has been used as 

well as ensuring extended battery life. It also 

controls the cell charging, protects the battery, 

sets the battery condition, and maintains the 

balance voltage among battery cells. 

It is known that state of charge (SOC) and 

state of health (SOH) are part of the information 

provided by BMS. As SOC and SOH can not be 

measured directly, an algorithm based on battery 

model is needed to estimate them. There are 

various battery models that have been developed. 

In general, battery model is divided into four 

classes: physical model (electrochemical), 

statistical model, analytical model, and electrical 

equivalent circuit model [3]. Electrical equivalent 

circuit model is widely used. This model allows 

battery parameters analysis using mathematical 

calculations of circuit components [3-5]. Battery 

in most cases is characterized using parameters of 

lumped circuit model. With various approaches, 

*Corresponding author. Telp : +62-89673487276 

 E-mail : adha.imam@ugm.ac.id 

http://dx.doi.org/10.14203/j.mev.2015.v6.113-122
mailto:adha.imam@ugm.ac.id


M.N. Ramadan et al. / J. Mechatron. Electr. Power Veh. Technol. 06 (2015) 113-122 

 
114 

battery parameters can be found using many 

algorithms, such as I-ARX [3], and recursive 

least square [4]. 

SOH describes the general condition and 

performance of the battery compared to that of 

battery when it is still new. Moreover SOH 

estimation allows us to estimate the life of battery 

to avoid battery failure due to improper usage. 

On electric cars, SOH informs the user that the 

battery replacement is needed when it reaches a 

certain degradation threshold. 

The study about SOH increases recently. A 

common approach is considering SOH estimation 

as a black box that is solved using artificial 

intelligence (AI) and machine learning (ML) with 

various battery condition without considering 

battery aging mechanism. The studies of AI and 

ML algorithms include neural network (NN) [5], 

[6], fuzzy logic [7], and support vector machine 

(SVM) [8]. In this study, valid and sufficient data 

are required to provide accurate estimation. 

Others methods, which are based on numerical 

analysis, describe battery aging mechanism such 

as mathematic equation [9], equivalent circuit 

model [10–12], and electro-chemical model [13]. 

In this paper, a battery SOH estimation 

method is proposed based on internal ohmic 

resistance and capacity of the battery where 

performances are compared. As the battery is 

used, its internal ohmic resistance will change so 

that it can be used for SOH estimation. The 

capacity will degrade due to battery aging that 

informs the degradation of SOH. The battery 

capacity can be obtained using coulomb counting 

method that computes current which flows of the 

battery. 

 

II. BATTERY MODEL  
 

A. Equivalent Circuit Model 
The first-order Resistor-Capacitor (RC) model 

is one among the best options available to be 

used in this work. The reason is the trade off 

between complexity, accuracy, and robustness as 

stated by X Hu [14]. The capacitance Ccap in 

Figure 1 represents the SOC of the battery. Using 

coulomb counting, the SOC can be defined as 

follows:  

𝑆𝑂𝐶 = 𝑆𝑂𝐶𝑜 −
1

𝐶𝑁
∫ 𝐼 𝑑𝑡

𝑡

𝑡𝑜
 (1) 

where SOC0 is SOC at initial time to, CN is 

capacity value in standard condition of the 

battery, η is coulombic efficiency that equals 1 

while discharge and is smaller than 1 in charge 

and I represents current which is negative at 

charge and positive at discharge [15]. 

In Figure 1, Voc is cell open circuit voltage 

(OCV) of battery cell, R0 is internal ohmic 

resistance, Rp is diffusion resistance, Cp is 

diffusion capacitance, Ibatt, Vt are the 

corresponding current and terminal voltage. Here 

Vt is set as output variable. Current source, Ibatt, 

represents the current flowing out of the battery 

cell. Ibatt acts as input variable. R0 describes 

internal resistance. The mathematical equations 

for the equivalent circuit are as follows: 

𝑢�̇�  = −
𝑢𝑝

𝐶𝑝𝑅𝑝
+

𝐼𝑏𝑎𝑡𝑡

𝐶𝑝
, (2) 

�̇�𝑜𝑐 =  
𝐼𝑏𝑎𝑡𝑡

𝐶𝑐𝑎𝑝
,  (3) 

𝑉𝑡   =  𝑉𝑜𝑐 − 𝑢𝑝 − 𝐼𝑏𝑎𝑡𝑡 𝑅0 (4) 

where 𝑢𝑝  is the voltage at the parallel RC 
network. 

 

B. Battery Testing and Schedule System 
When a car is moving irregularly, at a certain 

moment the batteries undergo a large load, but in 

other time encounter low constant load according 

to the usage. Changing road conditions can make 

the battery load varies unpredictably. Thus, this 

study included a few additional tests to see the 

battery life as well as to ensure that the battery 

model is valid. Schematic design of the 

experiments is shown in Figure 2. 

 

Figure 1. Battery model structure [10] 

 

Figure 2. Experimental schematic design 



M.N. Ramadan et al. / J. Mechatron. Electr. Power Veh. Technol. 06 (2015) 113-122 115 

The objective of this experimental design is to 

obtain the battery parameter in the form of 

current and voltage. In this work, the 

experimental devices consist of DC power source 

using imax B6 LiPro Balance Charger, dummy 

load using GW Instek PEL-2004, microcontroller 

ARDUINO UNO 32, and Matlab® R2013a 

software for computation and analysis purposes. 

Moreover, Turnigy Lithium Polymer Battery 

with a nominal capacity of 2.2 Ah was used in 

this study. The battery specifications are shown 

in Table 1 [16]. 

The experimental procedures are begun with 

static capacity test at 1C discharge current as 

shown in Figure 3. This test is conducted to 

determine the battery capacity and to obtain the 

corresponding battery SOC. The next test is pulse 

test. In this test, the battery is discharged for 30 

seconds and then rested for 30 seconds. After that, 

connect to load again and repeat the proses until 

the cut off voltage is reached. Cut off voltage is 

defined as terminal voltage at 20% SOC. 

Discharge pulse is 30 seconds for better accuracy 

of OCV curve because more data points OCV. 

Rest 30 seconds because at that time the battery 

terminal voltage has reached steady state. 

Enlarged rest time will have no effect on the 

condition of the battery terminal voltage. SOC 

20% is used as discharge limits to avoid over 

discharge and try to treat the battery as safely as 

possible. 

Pulse test is conducted to identify the battery 

model parameters of equivalent circuit and to 

obtain the OCV-SOC relationship. The next test 

is a test with varying input. The test is used for 

parameters of equivalent circuit model validation. 

The last test is aging cycle test where each 

cycle is done with constant current charging or 

discharging mode until the voltage reaches a 

specified value. Battery cycle life is usually 

determined by the number of cycles of charge – 

discharge. A battery can work well before its 

nominal capacity falls below 80% of the initial 

capacity of the battery [3][9][17]. Therefore, a 

number of cycles are needed to see the effect that 

occurs in the battery. 

 

III. BATTERY MODEL PARAMETER 
IDENTIFICATION 

 

A. Identification Method 
Based on the battery model shown in Figure 1, 

there are four series of parameters that must be 

obtained, those are Ccap, Ro, Rp, and Cp. The 

identification details of these parameters are 

illustrated as follows:  

a) The value of Ccap is obtained from static 
capacity test as shown in Figure 4. 

𝐶𝑐𝑎𝑝 =
𝑖 ∆𝑡

∆𝑉
. (5) 

b) In the pulse test, Voc is obtained from steady 
state voltage of each pulse as shown Figure 5. 

c) The resistance Ro is proportional to the 
reduction of voltage drop connected to the 

load. Resistance Rp and capacitance Cp are 

Start

Static Capacity Test

Pulse Test

Pulse Variation test

CC – Charge & CC – Discharge 

End
 

Figure 3. Battery test schedule 

Table 1.  

Battery specifications 

Parameter Value 

Capacity 2.2 Ah 

Max Discharge Current 44 A 

Max Charge Rate Current 4.4 A 

Charge Limit Voltage 4.2 V 

Discharge Limit Voltage 2.7 V 

 



M.N. Ramadan et al. / J. Mechatron. Electr. Power Veh. Technol. 06 (2015) 113-122 

 
116 

associated with voltage changes as shown in 

Figure 5. RLS Algorithm is used to obtain 

the value of Ro, Rp, and Cp. Linear 

relationship of input and output is obtained 

through a transfer function model of the 

battery. 

𝑉 = 𝑉𝑡 − 𝑉𝑜𝑐  (6) 

𝑉𝑡 − 𝑉𝑜𝑐 = [
𝑅𝑝

1+𝑠𝑅𝑝𝐶𝑝
+ 𝑅0] 𝐼 (7) 

𝑠 =  
2

𝑇

𝑘−1

𝑘+1
 (8) 

where T is the sampling period. Then 

equation (7) can be written as: 

𝑉(𝑘) + 𝑎1𝑉(𝑘 − 1) = 𝑏0𝐼(𝑘) + 𝑏1(𝑘 − 1) (9) 

where  

𝑎1 =
𝑇−2𝑅𝑝𝐶𝑝

𝑇+2𝑅𝑝𝐶𝑝
  (10) 

𝑏0 =
(𝑅𝑝+𝑅0)𝑇+2𝑅0𝑅𝑝𝐶𝑝

𝑇+2𝑅𝑝𝐶𝑝
 (11) 

𝑏1 =
(𝑅𝑝+𝑅0)𝑇−2𝑅0𝑅𝑝𝐶𝑝

𝑇+2𝑅𝑝𝐶𝑝
 (12) 

a1, b0, b1 are parameters that need to be solved. In 

this work, these parameters will be defined using 

RLS algorithm. 

The process in obtaining the parameters using 

RLS Algorithm is described as follows: 

𝐺(𝑘) =
𝑃(𝑘−1)𝜑(𝑘)

1+𝜑𝑇(𝑘)𝑃(𝑘−1)𝜑(𝑘)
  (13) 

𝜃(𝑘) = 𝜃(𝑘 − 1) + 𝐺(𝑘) [ 𝑉(𝑘) − 𝑉(𝑘 − 1) −
𝜑𝑇 (𝑘)𝜃(𝑘 − 1) ] (14) 

𝑃(𝑘) = 𝑃(𝑘 − 1) − 𝐺(𝑘)𝜑𝑇 (𝑘)𝑃(𝑘 − 1) (15) 

where 

𝜑(𝑘) = [ 𝑉(𝑘 − 1), 𝐼(𝑘), 𝐼(𝑘 − 1) ]𝑇 ,  (16)

𝜃(𝑘) = [ −𝑎1, 𝑏0, 𝑏1 ]
𝑇 . (17) 

The initial estimation value of parameter θ(0) 

and covariance matrix P(0) are first determined. 

Ro, Rp, and Cp parameter can be obtained by 

rewriting equation (10), (11), and (12) as: 

𝑅𝑝 =
2(𝑎1𝑏0+𝑏1)

1−𝑎1
2

,  (18) 

𝐶𝑝 =
𝑇(1+𝑎1)

2

4(𝑎1𝑏0+𝑏1)
, (19) 

𝑅0 =
𝑏0−𝑏1

1+𝑎1
. (20) 

B. Battery Parameter Estimation Result 
Figure 6 represents the relationship between 

SOC and OCV. From this relationship, the SOC 

can be predicted if value of Voc is known. Figure 

7, 8, and 9 respectively are R0-SOC, Rp–SOC, 

and Cp-SOC curves which are approximated by a 

second order polynomial.  

MATLAB® / Simulink™ was used to simulate 

the behavior of the battery. By trial and error, the 

parameters value are selected at 90% SOC. This 

choice will give good estimation errors. It seems 

that the model is able to follow the changes in the 

terminal voltage Vt as shown in the pulse 

variations test 1 and 2 in Figure 10(a) and 11(a). 

In addition, the model is able to predict the 

battery Vt when the current variation is added. It 

is also indicated by the relative error. In the pulse 

variation test 1, the maximum relative error is 

less than 3%, while in the pulse variation test 2, 

the maximum relative error is less than 1.5% as 

shown in Figure 10(b) and 11(b). 

This research was conducted at room 

temperature and the battery was not maintained at 

a certain temperature. However, the battery 

model can work optimally with the mean relative 

error for pulse variation of 0.25% for test 1 and 

the mean relative error pulse variation of 0.19% 

for test 2.  

 

  
Figure 4. Static capacity test Figure 5. Zoom-in of discharge pulse test 

 



M.N. Ramadan et al. / J. Mechatron. Electr. Power Veh. Technol. 06 (2015) 113-122 117 

Based on this error value, this model is 

acceptable. As shown in Figure 4, the battery 

discharge curve indicates that the battery is a 

nonlinear system as the voltage drops drastically 

when it is nearly depleted. When the battery 

capacity is approaching its lowest discharged 

limit, the battery voltage drops quickly and the 

condition is difficult to be modeled. Hence, the 

model well describes the battery on the linear 

part where the SOC is between 80% and 20%. 

 

IV. STATE OF HEALTH (SOH) 
As the lithium batteries start degrading once 

manufactured, due to the chemical degradation of 

the active material and other electrochemical 

phenomena, the internal ohmic resistance will 

increase and the capacity will decrease with age 

[18]. The process of degradation will change the 

battery performance. By increasing the number of 

cycles, the battery performance will decrease. 

This condition is monitored by SOH. SOH 

can be described as the battery performance at 

the present time compared to the performance at 

ideal condition and the battery’s fresh state [19]. 

There are many ways to determine the value of 

SOH, which are based on internal ohmic 

resistance, the battery capacity, the slope of 

charge or discharge curve, and the curve area of 

charge or discharge. However, the determination 

of the SOH based on internal resistance is more 

frequently used. 

 

A. SOH Based on Internal Ohmic Resistance 
One of the battery parameters that changes as 

a result of the degradation process is an internal 

ohmic resistance. By increasing number of cycles, 

the battery internal resistance will also increase.  

Therefore, the internal resistance can be 

selected to determine the value of SOH. 0% SOH 

means that the battery reaches its end of life. A 

battery reaches its end of life, when the internal 

resistance rises twice its initial value [17]. In 

other words: 

 SOH= 100% <=>   R=RIV 

 SOH= 0%  <=>   R=REOL=2xRIV 

with RIV is the initial value of internal resistance, 

and REOL is the value of internal resistance at the 

end of life. Therefore, SOH can be formulated 

into: 

𝑆𝑂𝐻 = (2 −
𝑅

𝑅𝐼𝑉
) 𝑥100%. (21) 

  
Figure 6. SOC-OCV plot 

 

Figure 8. Rp-SOC plot 

 

  
Figure 7. Ro-SOC plot Figure 9. Cp-SOC plot 

 



M.N. Ramadan et al. / J. Mechatron. Electr. Power Veh. Technol. 06 (2015) 113-122 

 
118 

Other forms of SOH based on internal resistance 

is formulated as: 

𝑆𝑂𝐻 = (1 +
𝑅𝐼𝑉−𝑅

𝑅𝐼𝑉
) 𝑥100% (22) 

where, R is the actual internal resistance and RIV 

is internal resistance of new battery [20]. 
The flowchart of SOH estimation on the basis 

of internal resistance is shown in Figure 12. Data 

from the experiment is shown in Figure 13. It is 

seen that the resistance of the battery increases 

with the increasing number of cycles. SOH curve 

in Figure 14(a) was SOH based on internal 

resistance which was obtained from experiments. 

Figure 14(b) shows SOH curve from the battery 

model, the internal resistance value was obtained 

from the estimation using RLS Algorithms. 

 
(a) 

 
(b) 

Figure 10. Validation result in pulse variation test 1; (a) Terminal voltage in pulse; (b) Terminal voltage error in pulse 

 

 
(a) 

 
(b) 

Figure 11. Validation result in pulse variation test 2; (a) Terminal voltage in pulse; (b) Terminal voltage error in pulse 

 



M.N. Ramadan et al. / J. Mechatron. Electr. Power Veh. Technol. 06 (2015) 113-122 119 

The result of SOH from the model has similar 

trend of that from the experimental. When the 

cycle is more than 20, the estimated SOH also 

shows 0% which is the same as the experimental 

results. In addition, the shape of SOH estimation 

curve is similar to the SOH experimental curve. 

However, this estimation algorithm does not give 

good estimation accuracy. As shown on the 

relative error of SOH in Figure 15, it can be seen 

that the greater error occurs when the cycle is 

more than 20.  

The SOH estimation based internal ohmic 

resistance can be determined easily. By obtaining 

the internal resistance only, the SOH can be 

estimated immediately. Therefore, this method is 

suitable to be applied on electric vehicle. 

 

B. SOH Based on Battery Capacity  
In addition to using the internal ohmic 

resistance, SOH of battery can also be 

determined using its capacity. The battery 

Start

Pulse test

All data ?

Recursive least square algorithm

Ro

SOH=2-(R/RIV)

End

no

yes
Ro=RIVPulse 1

yes

no

 

Figure 12. Flowchart of SOH estimation based on internal ohmic resistance 

 

Figure 13. Change of internal ohmic resistance 



M.N. Ramadan et al. / J. Mechatron. Electr. Power Veh. Technol. 06 (2015) 113-122 

 
120 

capacity will decrease with increasing number of 

charge-discharge cycles. The battery reaches its 

end of life which means 0% SOH when the 

capacity falls to 80% of its initial value [17] 

 SOH= 100%  <=>   C=CIV 

 SOH= 0%  <=>   C=CEOL=0.8xCIV 

with CIV is the initial value of the capacity, and 

CEOL is the value of the capacity at the end of life. 

SOH equation based on the capacity [17] is: 

𝑆𝑂𝐻 = (

𝐶

𝐶𝐼𝑉
−0.8

0.2
) 𝑥100% (23) 

and 

𝐶 =  ∆𝑆𝑂𝐶 ∗ 𝑄 (24) 

with C is the value of computed capacity,  ∆𝑆𝑂𝐶 
is the difference between initial SOC and SOC at 

the current time, and Q is the new battery 

capacity. C was computed with two methods, 

coulomb counting method and open circuit 

voltage method respectively. The flowchart of 

SOH estimation based on the battery capacity is 

shown in Figure 16. The result of the C value is 

shown in Figure 17. As shown in Figure 17, the 

capacity of a new battery is no more than 2.2 Ah 

but more than 1.78 Ah. That happens because the 

used value is battery capacity at 80% SOC. 

Figure 17 also shows the SOH change for both 

methods.  

Both methods show similar trends of SOH. 

The SOH which is based on coulomb counting 

method is smaller than the SOH which is based 

on open circuit voltage method so that it reaches 

the end of life threshold first. Both methods are 

easy to use in estimating SOH, but usually need 

longer computing time until all data is computed 

 

(a) 

 

 

(b) 

Figure 14. (a) SOH from experiment; (b) SOH from model 

 

 

 

Figure 15. Relative error of SOH  

 



M.N. Ramadan et al. / J. Mechatron. Electr. Power Veh. Technol. 06 (2015) 113-122 121 

V. CONCLUSION 
Based on the simulation and experimental 

results, it can be concluded that the model can be 

used to estimate the SOH as proven. The SOH 

estimation method based on battery capacity is 

more accurate than that based on internal ohmic 

resistance. However, the SOH estimation method 

based on internal ohmic resistance is faster. The 

methods also confirm that the SOH of the battery 

decreases according to the increasing number of 

charge-discharge cycles. 

 

 

Start

Finish

Initial SOC

Compute C

SOH = ( C/CIV - 0.8 ) / 0.2 * 100%

Compute SOC

CIV = C

All data ?

Cycle 1 ?

All cycle ?

yes

no

no

no

yes

yes

 

Figure 16. Flowchart of SOH estimation based on capacity 

 
(a) 

 
(b) 

Figure 17. Change of battery capacity; (a) Capacity - Cycle; (b) SOH - Cycle 



M.N. Ramadan et al. / J. Mechatron. Electr. Power Veh. Technol. 06 (2015) 113-122 

 
122 

ACKNOWLEDGEMENT 
This research program is fully supported by 

Universitas Gadjah Mada Faculty of Engineering. 

Therefore, the authors would like to express their 

sincere appreciation to the Faculty for providing 

the grant. 

 

REFERENCES 

[1] S. Cole and M. Leslie, “Long-term global 

warming trend sustained in 2013,” 2014. 

[Online]. Available: 

http://climate.nasa.gov/news/1029/. 

[2] A. Dinger et al., “Focus Batteries for 

Electric Cars,” 2010. 

[3] T. Sar et al., “A Hybrid Battery Model and 

State of Health Estimation Method for 

Lithium-Ion Batteries,” pp. 1349–1356, 

2014. 

[4] S. Li et al., “Study of Battery Modeling 

using Mathematical and Circuit Oriented 

Approaches,” pp. 1–8, 2011. 

[5] H. Lin et al., “Estimation of Battery State of 

Health Using Probabilistic Neural Network,” 

vol. 9, no. 2, pp. 679–685, 2013. 

[6] D. Andre et al., “Comparative study of a 

structured neural network and an extended 

Kalman filter for state of health 

determination of lithium-ion batteries in 

hybrid electric vehicles,” Eng. Appl. Artif. 

Intell., vol. 26, no. 3, pp. 951–961, 2013. 

[7] P. Singh and D. Reisner, “Fuzzy logic-based 

state-of-health determination of lead acid 

batteries,” 24th Annu. Int. Telecommun. 

Energy Conf., pp. 583–590. 

[8] V. Klass and M. Behm, “A support vector 

machine-based state-of-health estimation 

method for lithium-ion batteries under 

electric vehicle operation,” vol. 270, pp. 

262–272, 2014. 

[9] Z. Guo et al., “State of health estimation for 

lithium ion batteries based on charging 

curves,” J. Power Sources, vol. 249, pp. 

457–462, 2014. 

[10] Y. Zou et al., “Combined State of Charge 

and State of Health estimation over lithium-

ion battery cell cycle lifespan for electric 

vehicles,” J. Power Sources, vol. 273, pp. 

793–803, 2015. 

[11] C. R. Gould et al., “New Battery Model and 

State-of-Health Determination Through 

Subspace Parameter Estimation and State-

Observer Techniques,” vol. 58, no. 8, pp. 

3905–3916, 2009. 

[12] A. Hentunen et al., “Electrical battery 

model for dynamic simulations of hybrid 

electric vehicles,” 2011 IEEE Veh. Power 

Propuls. Conf., pp. 1–6, Sep. 2011. 

[13] X. Li and S. Choe, “State-of-charge (SOC) 

estimation based on a reduced order 

electrochemical thermal model and 

extended Kalman filter,” 2013 Am. Control 

Conf., pp. 1100–1105, Jun. 2013. 

[14] X. Hu et al., “A comparative study of 

equivalent circuit models for Li-ion 

batteries,” J. Power Sources, vol. 198, pp. 

359–367, 2012. 

[15] L. Lu et al., “A review on the key issues for 

lithium-ion battery management in electric 

vehicles,” J. Power Sources, vol. 226, pp. 

272–288, Mar. 2013. 

[16] A. Hand, “No Title.” [Online]. Available: 

http://www.hobbyking.com/hobbyking/store

/__317__85__Batteries_Accessories-

Turnigy_Lipoly.html. [Accessed: 20-Nov-

2015]. 

[17] A. Zenati et al., “A Methodology to Assess 

the State Of Health of Lithium-ion Batteries 

Based on the Battery ’ s Parameters and a 

Fuzzy Logic System,” IECON 2010, 2010. 

[18] E. Davis et al., “Internal ohmic 

measurements and their relationship to 

battery capacity – epri’s ongoing technology 

evaluation,” pp. 1–10. 

[19] M. U. Cuma and T. Koroglu, “A 

comprehensive review on estimation 

strategies used in hybrid and battery electric 

vehicles,” Renew. Sustain. Energy Rev., vol. 

42, pp. 517–531, 2015. 

[20] J. Remmlinger et al., “On-board state-of-

health monitoring of lithium-ion batteries 

using linear parameter-varying models q,” J. 

Power Sources, vol. 239, pp. 689–695, 2013.