COST 231 HATA MODEL FOR LTE

TECHNOLOGIES
Barrera Morales Ángel Rodrigo Amaguaya Uvidia Edison Fernando Torres Clavijo Diego Andrés
ESPOCH ESPOCH ESPOCH
Telecommunications School Telecommunications School Telecommunications School
[email protected] [email protected] [email protected]

Abstract—This article will study the different propagation

models for LTE technologies, given the wide evolution and The Okumura Hata Propagation Model, which is the
global implementation of LTE it is necessary to have some object of study of this work, is a model for predicting
propagation models that allow us to predict propagation losses
according to the characteristics in which it will be implemented, propagation losses in urban and suburban areas. Predicting
aspects such as power and frequency of transmission and the the power of a receiver is characterized by its use for mobile
following research will focus on the analysis of the Okumura communications, working at a frequency ranging from 150
Hata Propagation Model for the long-term evolution network [MHz] to 2000[MHz] in fact different from the Okumura
(4G LTE) at a frequency of 800 to 1800 [MHz]. Model working at a slightly longer frequency. The height
of the base in this model works from 30 [m] to 200 [m],
Keywords - Propagation models, Okumura-Hata, Wireless
of course it depends on the requirements of use. The height
channel, 4G LTE.
of the mobile ranges from 1 [m] to 10 [m] and depends on
where the mobile is located. This model has a correction
I. INTRODUCTION
factor by area, the correction factor for the effective height
LTE (Long Term Evolution) is a wireless broadband data of the mobile antenna depends on the type of city which can
transmission technology that is primarily designed to be be: small cities, medium cities and large cities.[7][8][9][10]
able to support the constant access of mobile phones and
portable devices to the internet.[1] Developed by 3GPP, LTE Regarding LTE and LTE-A (later evolution), cellular
is a new generation compared to UMTS (third generation networks are detailed that are not limited to providing
or 3G) and GSM (second generation or 2G). It allows high services to mobile terminals, but also allow connection to
bit rates with low latency, is cheap and easy to deploy by fixed wireless networks with a good transmission speed,
operators, and avoids fragmentation by type of duplication. capable of handling the requirements of the real world. The
LTE uses OFDMA (Orthogonal Frequency Division Multiple 4G technology so far are long term evolution (LTE) and
Access) links to minimize interference and increase spectral IEEE’s Wimax. The technology is based on multicarrier
efficiency, using channels of variable size between 1.25 and modulation like orthogonal frequency division multiplexing
20 MHz.[2][3] and Multiple Input Multiple Output (MIMO). The system
Evolution of LTE technologies: First generation 1G (voice has 100Mbps at high mobility and around 1Gbps at low
channels with analog technology, the link quality is very mobility.[10]
low, handover is imprecise and there are no security), second
generation 2G (offers portability due to its SIM card), third
generation 3G (higher transmission speed).[4] II. COST-231 HATA PROPAGATION MODEL
It was born from an extensive campaign of measures
Propagation models are a mathematical expression whose carried out in Tokyo, in the range of 100 MHz to 1920 MHz,
purpose is to characterize the propagation of radio waves by the Model Okumura.[11] The main result of Okumura’s
based on a series of parameters, such as frequency, distance work was a set of curves that provide the average level of
and other factors that will depend on each model. Among attenuation relative to free space, depending on the frequency,
some of the propagation models for LTE technologies we find: distance between transmitter and receiver, the height of the
Xia-Bertoni Model (describes the propagation of the signal base station antennas and the mobile station, as well as
in the UHF band), Walfisch-Bertoni Model (valid when there several correction factors specific to different types of path.
is no line of sight between the base station and the mobile ),
Deterministic models (the same inputs always give rise to the Hata established a series of numerical relationships that
same outputs, without the uncertainty principle taking place), describe the graphical method proposed by Okumura. The
Semi-empirical models (they incorporate empirical correction main result provided by the Okumura Hata model is the
factors).[4][5][6] median value of the basic propagation loss, depending on the
frequency (f), distance (d), and antenna heights of the base frequencies. With this method, the expression remains:
station hb and mobile hm .[12]
Lp (dB) = P L0 − 13.82log10 hBST − a(hM S ) + [44.9 −
The Okumura Hata model is restricted to the following 6.55log10 hM S ]log10 d + Ck
limits: Where
P L0 : 46.3 + 33.9log10 f
f : 150 to 1500 [M Hz] Ck : Area correction factor
hb : 30 to 200 [m] Lp : Path loss in(dB)
hm : 1 to 10 [m] f : The carrier frequency(in M Hz)
d : 1 to 20 [km] hBST : Base station antenna height(m)
hM S : Mobile antenna height(m)
The Okumura Hata model expresses the basic loss of a(hM S ) : Correction factor for mobile station
propagation, Lb, as follows[13]: d : base station to mobile separation distance(km)
Ck = 0 dB for medium sized city and suburban areas and
Lb = 69.55 + 26.16 log f - 13.82 log hb - a(hm ) + 3dB for urban centers
(44.9 - 6.55 log hb ) log dm
Now, the valid frequency range for the above equation
where a hm is a correction factor that depends on the is between 1500 and 2000 MHz, base atenna height of
height of the mobile and is calculated as follows: 30m − 200m, mobile antenna height is between 1m-10m,
and separation distance 1km.[7]
1) For urban areas:
a) Small or medium-sized cities:
III. ARTICLES
a(hm ) = (1.1 log f - 0.7) hm - (1.56 log f - 0.8)
where 1 ≤ hm ≤ 10 [m] A. Scenario I

b) For large(cities: The first scenario corresponds to an article in ife Journal of

8.29(log1.54hm )2 − 1.1 f ≤ 200[M Hz] Science approved in April 2019. Whose theme is ”Research
a(hm ) =
3.2(log11.75hm )2 − 4.97 f ≥ 400[M Hz] and Modelling of fourth generation pathloss long-term
evolution network along major highways in lakes (Nigeria)”.
2) For suburban areas:
Lb = Lb (urban) − 2[log(f /28)]2 − 5.4
Propagation measurements were carried out with an
estimated population of more than 22 million inhabitants with
3) For rural areas:
a tropical climate with an average annual rainfall of 1693 mm.
Lb = Lb (urban) − 4.78log(f )2 + 18.33logf − 40.94
Propagation measurements were taken at 1800MHz, from
three base stations of evolved nodes (eNodeBs) located along
In general, this calculation method provides good results in three different routes in Lagos, and the received power of the
urban and suburban environments, but not in rural areas, as it reference signal (RSRP) was measured for the three eNodeBs,
does not take into account the ripple of the land or the effects using the drive test equipment of Huawei technologies.[14]
derived from the degree of urbanization along the way.

Because this model was only valid for frequencies between

150 and 1500 MHz, the Cost-Hata method was developed,
which provides an expression for calculating propagation
losses at higher frequencies. With this method, the expression
remains:

Lb (dB) = 46.3 + 33.9log10 (f ) − 13.82log10 (hrx ) −

a(hm ) + (44.9 − 6.55log10 hrx )log10 d

Now, the valid frequency range for the above equation Table I: Modeling measurements and parameters.
is between 1500 and 2000 MHz, leaving the range for the
other parameters of the equation equal to the previous general Comparison of measured and predicted route loss. The loss of
case.[13] route measured in what is investigated. Routes are compared
to the COST-231 Hata. This model has been selected for
COST-231 Hata method was developed, which provides inline comparison with the frequency range for which it is
an expression for calculating propagation losses at higher most valid.
Figure 1: Comparison of measured pathloss against free Figure 4: Comparison of measured and predicted pathloss
space loss. along route 3.[14]

Figure 2: Comparison of measured and predicted pathloss

along route 1.[14]

Table II: Basic statistics of measured path loss compared

with predicted pathloss.[14]

Figure 3: Comparison of measured and predicted pathloss

along route 2.[14]
textbfTable III: Standard deviation and percentage errors of
Figure 6: Measurement points in the West.
path loss models.[14]

The COST-231 Hata model showed the least ingrained mean The coordinates of the measurement points (Red Dots)
square errors of 7.42dB, 7.63dB, and 9.64dB along routes 1, were found through a free mobile application called GSM
2, and 3, respectively. This model was selected and modified, Field Test. BTS (Green Points) data to identify CID and LAC
using the least squared algorithm. The model was seen to pre- was performed using cell phone and GSM Field Test app.
dict the loss of path of data measured with enhanced RMSEs
of 5.20dB, 4.89dB, and 8.78dB, for investigated routes 1, 2,
and 3, respectively. Nearer-zero RMSEs show improvements
in model prediction accuracy, and the modified model could be
very useful for renewing network planning, system design, and
the deployment of wireless networks on and around selected
roads in Lagos, Nigeria.
B. Scenario II
It is a Master’s thesis whose theme is ”Statistical Adjustment
to Propagation Models for Mobile Networks”
Figure 7: GSM Field Test mobile mobile app.
The radio spectrum measurement campaign was carried
out in the city of Bogota D.C, it took a time of 72h in the
South and West [15] area. In total, the measurement campaign
was conducted over thirteen days. Spectrum occupancy was
measured in the range of 824 MHz to 849 MHz and also at
1700 MHz to 1900 MHz.[16]

Table VI: Parameters of the South and West Zones.[16]

Figure 5: Measurement points in the South zone.

The transmission power was taken according to Nokia’s
tool (36 to 45 dBm), for the calculation of the received
power. The receiving power is the one that is finally
compared to the measurements obtained in each zone, the
gain of the transmission antenna is 3dBi and the gain of the
receiving antenna of 1dBi; these values add to the difference
in transmission power and propagation losses found through
the model.
The difference in the propagation losses of the model
between the south and the western zone is minimal, when
implementing measurements of a single zone is sufficient to
make the adjustment, this because the difference in the height
Figure 8: Model COST-231 Hata – South Zone [824
of the antennas is minimal (10 meters), the transmission
MHz].[16]
power is the same with an equal coverage of 300 meters and
the transmission power equal to being the same equipment in
the BTS.

Table VI: Propagation Losses and PRx – COST-231 Hata

West Zone (1-20 m).[16]
Table V: Propagation losses PRx COST-231 Hata South
Zone (1-20 m).[16]
The third article corresponds to an American Journal
of Engineering Research article November 2015 that is
entitled ”Comparative study of path loss models for wireless
communication in the urban and suburban environment for
Port Harcourt, Nigeria”. Authors: Akujobi Ekeocha, Nosiri
Onyebuchi and Achumba Ifenyinwa.

The study was based on the comparative analysis of

propagation models for the Global Mobile Communications
System at 900MHz. Drive test analyses were carried out
from two selected lands in Rivers State, namely GRA Phase
 Figure 12: Probabilistic Spectrum Occupancy Pattern –
COST-231 Hata.
Figure 9: Model Okumura-Hata – West Zone [824
MHz].[16]

Figure 10: COST-231 Hata model and measurement Figure 13: View of drive test routes.
regression line.

Figure 11: Study area. Figure 15: Comparison of Predicted and Measured Path loss
in suburban environment.
II and Aggrey Road classified as urban and suburban areas
The COST-231 Hata Model was shown to show considerable
respectively, to evaluate the best propagation model for the
performance and can be tuned for better performance. It can
study area.[17]
be said that the performance of the propagation model depends
on the environment. Therefore, the influence of environmental
factors should be emphasized and properly accounted for
when considering base station sites. Therefore, one of the
best possible ways to get better channel characteristics is to
optimize existing path loss models based on measured data
collected from a specific area.[17]
 Table VII:Empirical and Measured Path loss values for
Urban Environment.

Figure 14: Empirical and Measured Path loss values for

Suburban Environment.
Table VIII: Comparison between predicted and measured
Path loss in an urban environment.

IV. RESULTS ANALYSIS

For the calculation of propagation losses using The COST-
231 Hata Model in 4G LTE technologies, the deployment
scenario, the available mapping information, the transmitter
height and the terrain profile of the scenario considered are
taken into account. In the case of the first scenario carried out
in Lagos Nigeria Table II shows the measured path losses and
expected road losses that compared to scenario II (Bogotá,
Colombia), both studied in urban area propagation losses are Table IX: Comparison of Predicted and Measured Path loss
greater in scenario I and this is because in Lagos Nigeria in suburban environment.
the distance between the transmitting and receiving antenna
is greater with a value of 20 [Km] than in Bogota , which
V. CONCLUSIONS
implies that the greater distance greater interference which
makes the signal weaker. 1) The COST-231 Hata Propagation Model for a 4G LTE
Table v (South Zone) and Table VI (West Zone) show the mobile network at a frequency of 1800 [MHz] for
losses from propagation and reception power of Scenario lagos (Nigeria) roads using Huawei Technologies drive
II which compared to scenario III carried out in Harcourt test equipment, showed the least ingrained mean square
(Nigeria) the losses are greater in Harcourt and this is because errors of 7.42 [dB], 7.63 [dB] and 9.64 [dB] along routes
the distance between the transmitting and receiving antenna is 1, 2 and 3, this model is highly chosen when it is in
greater than the proposal in Bogota , both studied in urban search of predicting the path losses of data measured
areas. with RMSES in this case in this scenario resulted in
Among the 3 scenarios the one that obtained the greatest losses of 5.20 [dB] for Route 1, 4.89 [dB] for Route 2
propagation losses was harcourt first because the distance and 8.78 [dB] for Route 3. Undoubtedly this propagation
between the transmitting and receiving antenna is greater than model is ideal when the purpose of its use is to improve
in the other two scenarios, second because the height of the the accuracy of signal prediction.
transmitting antenna was greater whose value is 40 [m] than 2) Scenario II in the City of Bogota D.C in the South
in Lagos (30 m) and Bogota (35 m), since when we worked and West areas at a frequency of 824 [MHz] resulted
with the Model COST-231 Hata depends to a large extent on in a loss of 65.37 [dB] in the South Zone, 66.21 [dB]
the height of the transmitter , when the transmitter is high, this and 67.01 [dB] corresponding to distances of 18, 19
propagation model does not model signal fading which implies and 20 [m], while in the West Zone it resulted in
that predictions are more reliable if diffraction is modeled from losses of 65.02 [dB] , 65.85 [dB] and 66.62 [dB] at
land and construction information; However, when this type of distances of 18, 19 and 20 [m]. The difference in the
information is not available, the COST-231 Hata offers a good propagation losses of the model between the south and
approximation to the average value of the received signal. the western zone is minimal, because if we realize in the
 loss result table there is little variation, which implies [3] J. E. A. Peña, “Sistema mu – mimo en una red 4g – lte,” 2014.
that to make measurements is sufficient a single zone [4] J.-S. S. Sung-Woong, JoJun Hyuk, “An analysis of lte-maritime’s
path loss models for the mobile communications system in maritime
because the height of the antenna is minimum of 10 environments,” 2019.
[m], the transmission power is the same with an equal [5] S. N. K. Marwat, “Lte channel modeling for the level simulations
coverage of 300 [m] and the transmission power equals system,” 2016.
[6] J. Walfisch and H. Bertoni, Theoretical Model of UHF Propagation in
being the same equipment in the BTS. Urban Environments, 1988.
3) Scenario III in Harcourt, Nigeria, was based on the [7] J. M. Huidobro, “4g. lte systems.’ under ’mobile communications. gsm,
analysis of the Global Mobile Communications System umts and lte systems,” 2013.
[8] I. A. O. Zachaeus K. Adeyemo, Owolabi K. Ogunremi, “Optimization
at 900MHz with the use of model COST-231 Hata, drive of okumura-hata model for long term evolution network deployment in
test analysis is of two selected terrains in Rivers State, lagos, nigeria,” 2016.
namely GRA Phase II and Aggrey Road classified as [9] E. C. in the Field of Scientific and T. Research, Digital Mobile Radio:
COST 231 View on the Evolution towards 3rd Generation Systems, 1996.
urban and suburban areas respectively, resulting in av- [10] C. Dalela, “Radio chapropagation and nnel modeling comparison study
erage path losses of 115.16 [dB] for urban environment for wireless systems,” 2013.
at a distance of 0.5 [Km] while for the suburban area [11] C. A. S. Morales, “Recommendations uit r p.525, p.526, p.838, p.833,”
Radio propagation - 2017/04 -SAIA B - Heddy Lu Gimenez, 2018.
resulted in 114.32 [dB] at the same distance as the urban [12] M. C. E. Section, “Empirical formula for propagation loss in land mobile
area. It is therefore concluded that the performance of radio services,” IEEE Transactions on Vehicular Technology ( Volume:
the COST-231 Hata Model depends on the environment 29, Issue: 3, Aug. 1980), 1980.
[13] F. J. D. Parsons DSc (Eng), FREng, The Mobile Radio Propagation
so that the influence of it is considered. Channel, Second Edition, 2001.
4) The fundamental factor of the COST-231 Hata Model [14] A. L. Imoize and A. I. Oseni, “Investigation and pathloss modeling of
is the frequency since when conducting the study of fourth generation long term evolution network along major highways in
lagos nigeria,” Ife Journal of Science vol. 21, no. 1, 2019.
the scenarios shown it is concluded that the frequency [15] ITU-R, “Spectrum occupancy measurement,” International Telecommu-
change more predominantly affects the results obtained nications Union, 2011.
in each of the simulated environments. It should be noted [16] R. E. C. Sánchez, “Statistical adjustment to propagation models for mo-
bile networks,” Francisco José de Caldas District University, COLOM-
that the LTE system depends not only on the base radios BIA, 2015.
for the coverage offered, but also on the mobile terminals [17] N. O. Akujobi Ekeocha and A. Ifenyinwa, “Comparative study of path
that support LTE. loss models for wireless communication in urban and sub-urban envi-
ronment for port harcourt, nigeria.” American Journal of Engineering
5) It is concluded that the loss path decreases with the Research (AJER), 2015.
increase in distance due to a corresponding decrease
in field strength and to culminate the COST-231 Hata
model for the urban region is the closest of all propaga-
tion models considered to classify the environment into
consideration as it gives a better and closer prediction
to the measured path loss.

VI. RECOMMENDATIONS
1) For a good prediction of propagation losses it is
recommended first to have the information of the
territorial distribution so that the model can be
characterized and therefore decide whether or not it
is applicable to the specific conditions of the city in
which you want to work.

2) Another factor to consider is the influence of the

environment should be emphasized and duly accounted
for when considering base station sites. Therefore,
one of the best possible ways to get better channel
characteristics is to optimize existing path loss models
based on measured data collected from a specific area.

R EFERENCES
[1] E. A. D.-R. y. J. L. L.-B. Sergio Vidal-Beltrán, “Simplified propagation
model for lte on the 2.1ghz frequency,” 2017.
[2] K.-E. D. Suneetha Rani, Subrahmanyam VVRK Behara, “Comparison of
standard propagate cntion model (spm) and stanford university interim
radio propagation models (suis) for long-term evolution (lte),” 2018.