International Conference on Computer and Communication Engineering (ICCCE 2012), 3-5 July 2012, Kuala Lumpur, Malaysia

978-1-4673-0479-5/12/$31.00 ©2012 IEEE

Is Hata Path Loss model a Macro-Cell Model Only?

A Question of Debate

Mahdi A. Nisirat, Mahamod Ismail, Liyth Nissirat Department of Electrical, Electronics and System

EngineeringUniversiti Kebangsaan Malaysia, 43600 UKM Bangi, Selangor Darul Ehsan, Malaysia

mamnisirat@gmail.com, mahamod@eng.ukm.my liyth@hotmail.com

Salim AlKhawaldeh Electrical Engineering Department, Faculty of Engineering

Technology, Al-Balqa' Applied University, Jordan skhawaldeh@yahoo.com

Abstract—Since the first publication of Hata Path loss model in 1980, it was imposed by the author that the model is valid for distances more than 1 km. Therefore, Researchers have categorized the model to suit Macro cells of distances from 1 km up to 20 km. Many recent and older researches have argued that Hata model can be used even in micro cell distances down to tens of meters from the transmitter under certain conditions. Accurate results are achieved in such distances which supports the claim that this model is precise enough down to tens of meters. The argument is to evaluate between simplicity, and cost efficacy of applying Hata model verses optimum accuracy required in view of the fact that real measurements are to be collected after all. In this brief article, arguments are highlighted to realize if this model is a matter of choice to be used in micro cells, and though the imposed constrained as the range of validity to be greater than 1 km, would be relaxed. Examples are given to support the claim.

Keywords-Hata path loss models; terrain parametrs; optimization; micro – cell systems

I. INTRODUCTION Many early propagation path loss models have gone

under intense optimization and correction procedures to make them more realizable, efficient, accurate and prevalent. The main target is to come up with the most highly accurate propagation path loss model that would fit a broad category of different landscape criteria’s, geographical variation and weather situations. Corrections, optimizations and new parameters have been added or imposed to get the optimum design. Hata path loss model was among one of the most widely used models for calculating outdoor propagation path losses. Since the introduction of the model by Hata on 1980 [1], it was imposed by the author that validity of the model is for distances over 1 km. Originally Hata work is an empirical model based on the former known path loss model proposed by Okumura in 1968. Okumura model was proposed for large area coverage, or what is called macro – cells, and based completely on graphical curves to estimate different model parameters [2].

Many old and recent studies have claimed that Hata path loss model can be used for micro – cell coverage areas, for distances less than 1 km, under some existing assumptions. In this paper the validity of such claims are under analysis to evaluate such statement.

II. HATA PATH LOSS MODEL

This model was originally proposed for urban areas and with correction factors proposed to suit losses calculated for suburban and rural areas. The formulas given are reformulation of Okumura’s model and were obtained by curve prediction methods [1]. For urban areas the path loss equation is given as

dloghlog55.69.44

hahlog82.13

flog16.2655.69dBL

t

rr

curban �

(1)

For open area category given as

94.40flog33.18

flog78.4LdBL

c

2curbanopen �

(2)

For suburban area, it is of the form

4.528flog2LdBL

2c

urbansuburban ���

���

��

�� (3)

rha , given in the urban relation, is Hata correction factor, for small and medium size areas, it is of the form [1]

dB8.0flog56.1h7.0flog1.1ha

c

rcr � (4)

274

for large areas it is as

MHz300ffor c �

dB1.1h54.1log29.8ha 2rr � (5)

MHz300ffor c �

dB97.4h75.11log2.3ha 2rr � (6)

where d is the transmitter – receiver antenna separation distance and it is proposed to be valid for 1 km - 20 km, fc is the operating frequency band for the range 150 MHz - 1500 MHz. The ranges of the transmit antenna ht and the receive antenna hr heights are 30 m - 200 m and 1 m - 10 m respectively [1].

Generally, the main equation of Hata path loss model would be assumed of the form

)d(logBAL� (7)

where A is referred to as the DC part or the offset part, and would represent the factor

rr

c

hahlog82.13flog16.2655.69dBA �

(8)

B refers to the rate of change (slope) of the path loss formula and represents the factor

thlog55.69.44kmdBB � (9)

III. REVIEW OF EXISTING LITRATURE USED HATA MODEL

A. Hata path loss model in Macro – Cell range

Hata model is assumed as one of the most simple path loss models. This is due to the fact that the model requires the pre – existence of only four parameters for operation, distance, operating frequency, transmit and receive antenna heights. Corrections are available for different landscape categories in order to statistically account for different losses found in different scenarios. Though found indefinite by many [3], the proposed parameters were accurate enough in putting the model among one of the most prevailed path loss models ever.

In [3] it is concluded that Hata model originally was proposed for large area coverage and would predict path loss with different accuracies, though the variation is still within accepted level of deviation. A 6 dB average standard

deviation was reported; even practically higher values were often encountered. Authors suggest that the model would be used in micro – cell range of distances down to 200 m with acceptable level of accuracy.

More deviation is reported as concluded by [4]. In the mentioned study, Hata model were used as an optimization model in many different terrain regions in India, with frequency bands of 150, 320, and 440 MHz. It was noted that a deviation of about 12 dB was reported from the real measured data. More over, more deviation was expected in more rough terrain areas.

In the same sequence, authors in [5] found that measured path loss data collected in Oman area, which considered open and arid area, deviated from Hata path loss model. The deviation was reported to be in the range of 15.3 dB. This deviation has urged the need to correct for the model, which was advised by a constant DC shift of about 15 dB. The deviation was attributed to the differences in landscape and rain fall levels between Oman and Tokyo.

On the other hand, authors in [6] have proposed a new terrain based model to suit data measurements in different locations of Denmark. Data gathered of frequency bands of 143.9 MHz, 435MHz, 970 MHz, and 1900 MHz, for distances up to 11 km. Error standard deviation of the range 6 – 12 dB from Hata loss model was calculated. The same trend of deviation with less value of about 2 – 3 dB was calculated for the new proposed model. This systematic error was attributed to terrain roughness deviations of different areas.

A study conducted in rural areas in New Zealand [7], for frequency ranges of 2.4 GHz and 5.8 GHz has analyzed the accuracy of about 28 different proposed models found in literature. Results have shown that even in rural areas and with tuning of all the required different parameters of different models, the best RMSE deviation from real measured path loss was not less than 12 dB. Hata model deviates about 20 dB in the given study. They concluded that most models available in literature need to be tuned for specific areas.

Contrary to the last conclusion, authors of [8] have compared different models such as Hata, Okumura, and Egli with real measured data obtained in K. L. area, Malaysia. Analysis has shown that Hata model has the smallest mean relative error of about 15.46 %. Data collected was of CDMA range systems.

B. Hata path loss model in Micro – Cell range

Many literatures have applied Hata for microcell areas and with transmit antenna heights below 30 meter.

As concluded in [9], a correction terrain parameter based on RMSE analysis and regression fitting to optimize both of

275

the DC part and the rate of change of Hata urban relation for micro cell areas, for the frequency band of 2,100 MHz. The deviation of the measured data from the expected Hata path loss model is reported to have an average of 15 dB. This deviation was assumed by the authors due to high slope terrain effect and, as a result, a correction was proposed for hilly areas with high degree of slope.

The study performed by [10] has shown that Hata model would be accurate within the range of 8.5 dB for the frequency of 400 MHz, and within 16.5 dB for the 8 GHz. Data were collected in different regions in Japan in a distance range of less than 1 km down to 100 meter. In addition to that, regression analysis concluded that terrain factors such as street slope and street width would be highly correlated to path loss in the given regions.

Both studies performed in [11, 12] have predicted losses in micro cell areas in different locations in Jordan for the 900 MHz band. In those papers the argument stated that Hata model would be valid in microcell areas of less than 1 km distance given that antenna heights are above roof tops. In [11] the study proposed a correction parameter based on the roughness of the route between the transmitter - receiver and was used to correct for the DC part of Hata urban formula. A correction of 3 dB is achieved by this method in suburban area category [11], and up to 2.75 dB and 2.5 dB for open and urban areas [12].

In [13] a study conducted for the frequency band of 1800 MHz, for different suburban areas in Jordan. Measuring route distance ranges were below 1 km which is assumed a micro – cell range. In the study, high correlation factor is reported between the measured path loss and the standard deviation of the path of the measurements. Linear regressing parameters are used to correct for both of the DC part and the slope of the Hata urban path loss relation. A deviation of up to 17 dB from Hata model was reported.

With the same approach, the study conducted in [14] was also performed in Jordan for different area categories mainly urban, suburban, and open areas. Measured path losses in micro – cell regions for the frequency band of 1800 MHz were predicted utilizing a terrain roughness parameter. Reported results claimed RMSE average corrections of 21 dB, 20 dB, and 17 dB for open, suburban, and urban areas respectively.

In addition to that, the study executed in [14] has applied Hata model for distance ranges below 600 meter with different transmit antenna heights in the vicinity of roof tops, for 1800 MHz. It was concluded that Hata suburban model predicts the data very well in the given range. Optimized correction factor is proposed to minimize overall prediction error in the given area.

In the same manner, authors in [15] has applied Hata model for micro and/or small cells, at 900 and 1800 MHz. Results concluded that Hata model predicts path loss very closely down to tenths of meters. In addition to that it was reported much of the behavior of the 900 MHz bands would be reflected to the 1800 MHz band with a shift up loss of 10

dB. More over, it is found that more loss is reported in lower roof top antennas than the over roof top antenna heights.

Authors in [16] have compared Hata suburban, Egli, and COST 231 Walfisch – Ikegami model to the measured path loss data. It was concluded that Hata suburban model is the best fitting model compared to others. Both the offset dc value and the slope of Hata suburban model where optimized to suit different areas at different frequencies in Malaysia, the study was conducted for distance ranges of less than 100 meter length. This distance range is assumed not valid for Hata model, though results showed that Hata model is accurate within this range. Optimized values by the least square method (LSM) are proposed to minimize relative errors in the given areas.

Finally as reported by the authors of [17], Hata model is based on the work of Okumura’s model. Thus Hata imposed the distance validity range to be greater than 1 Km as to resemble as accurate as possible the initially based model proposed by Okumura.

In most of the literature mentioned above, not only the distance range of validity of Hata model was relaxed, but also the operating frequency range and the transmit antenna height.

IV. COMMENTS AND DISCUSSION Hata – Okumura model is well known of being very

simple and least demanding model with high level of accuracy in the macro – cell range of validity. Such an advantage led to the accreditation of this model by the International telecommunication union (ITU). This has encouraged many researchers to accept the adaptation of Hata model in the microcell region.

As mentioned above different level of accuracies was reported in different locations and situations. RMSE deviation in the macro – cell range of up to 20 dB was reported. In the same manner, RMSE deviation in the micro – cell range was within the same level. This deviation as being accepted in the macro – cell range would validate the application of Hata in the micro – cell range.

To optimize Hata model in the micro – cell range, a general accepted rule that good accurate micro-cell models should account for the topological landscape variation that exists between the transmitter and the receiver. But if too many factors are taken in to account this may render the simplicity and overwhelm the main advantage of the model. Landscape parameters and factors can be divided in general, but not limited, to

� Building distribution, utilization factor, height, width, knife edges, and gaps between buildings, type of structured materials.

� Streets width, length, and location of cross roads, street signs, advertisement signs, trees, and parked vehicles.

� Effective transmitter – receiver height, angle of incidence, clearance angle, and field polarization.

276

� Existence of hilly areas with effective height difference and terrain parameters of the area between transmitter and receiver.

Many models found in literature have adopted one or more of the last landscape criteria’s, such as in [18-19]. On the other hand, such models require either an exhaustive amount of data or immense deal of calculation and analysis. Hata model has implicitly corrected his model by introducing correction parameters for different area categories, accounting for landscape variation. Therefore many are currently trying to introduce, simple and basic, landscape and terrain parameters to Hata model in order to optimize the prediction accuracy.

Looking for the simplest applied model is essential especially for the preliminary stage of planning. After all real data collection, optimization, and correction of system parameters is a standard procedure for the long lasting quality of service (QOS). This part would include changing the power, the angle, the height, and the direction of the transmitter for better coverage. Though expensive and time consuming, it would be a must to sustain the required QOS.

V. CONCLUSION As a conclusion Hata model is still considered as one of

the simplest and most used path loss prediction models. As reported the level of deviation from Hata model in the Micro – cell level lies within the measured level of deviation for the Macro – cell range. This result would initiate the base of accepting Hata model as a general applicable model as a micro – macro cell path loss model. The adaptation of the model requires that the transmit antenna height be above roof tops. By the author’s point of view, good micro - cell models should also include different terrain and landscape parameters to achieve optimum accuracies. On the other hand, heavily detailed terrain and landscape data may render the desire of simple prediction models. At the end, real optimization and road test for good lasting QOS, no matter the pre-used path loss model, is indispensable.

VI. ACKNOWLEDGMENTS The authors would express their appreciation to the

support of this work under the governmental grant number Oup-2012-182.

REFRENCES [1] M. Hata, “Empirical formula for propagation loss in land mobile radio

services,” IEEE Trans. Vehicular Technology, pp. 317–325. 1980. [2] R. Mardeni, L. Y. Pey, “The Optimization of Okumura’s Model for

Code Division Multiple Access (CDMA) System in Malaysia,” European Journal of Scientific Research, pp.508-528, 2010.

[3] M. J. Mehler, “The Microcellular Propagation Challenge,” IEE Colloquium on Micro-Cellular Propagation Modelling, pp.1 – 4, 1992.

[4] M. V. S. N. Prasad, K. Ratnamala, M. Chaitanya and P. K. Dalela, “Terrestrial communication experiments over various regions of Indian subcontinent and tuning of Hata’s model,”Annals of Telecomunications, SpringrLink, pp. 223 – 235, 2008.

[5] N. Elfadhil, M.A. salam, A. Al-Lawati, O. Al-Qasmi, M. Al-Gheithi and Z. Nadir, “Modification of an Open Area Okumura-Hata

Propagation Model Suitable for Oman.” IEEE TENCON, pp.1 – 4, 2005.

[6] J. T. Hviid, J. B. Andersen, J. Toftgkd, and J. Bldjer, “Terrain-Based Propagation Model for Rural Area-An Integral Equation Approach,” IEEE Transactions on Antennas and Propagation, pp. 41 – 46, 1995.

[7] C. Phillips, S. Raynel, J. Curtis, S. Bartels, D. Sicker, D. Grunwald, and T. McGregor, “The Efficacy of Path Loss Models for Fixed Rural Wireless Links,” Passive and Active Measurments, SpringerLink, pp. 42 – 51, 2011.

[8] R. Mardeni, L. Y. Pey, “The Optimization of Okumura’s Model for Code Division Multiple Access (CDMA) System in Malaysia,” European Journal of Scientific Research, pp. 508 – 528, 2010.

[9] T. Fujitani, S. Tomisato, and M. Hata, “Experimental Study of Mobile Propagation Loss Correction Formula for a Slope Terrain Area,” IEEE, Vehicular Technology conference, pp. 1-5, 2010.

[10] K. kitao, and S. Ichitsubo, “Path loss prediction formula for microcell in 400 MHz to 8 GHz band,” IEEE, Electronics Letters, pp. 685 – 687, 2004.

[11] M. A. Nisirat, M. Ismail, L. Nissirat, S. Al-Khawaldeh, “A Terrain Roughness Correction Factor for Hata Path Loss Model at 900 MHz,” Progress In Electromagnetics Research C, pp. 11-22, 2011.

[12] M. A. Nisirat, M. Ismail, L. Nissirat, S. Al-Khawaldeh, “A Hata Based Model Utilizing Terrain Roughness Correction Formula,” IEEE, Telecommunication Systems, Services, and Applications (TSSA), pp. 284 – 287, 2011.

[13] M. A. Nisirat, M. Ismail, L. Nissirat, S. Al-Khawaldeh, “Linear regrission route roughness parametr to correct Hata path loss prediction formula for 1800 Mhz,” Progress In Electromagnetics Research Symposium Proceedings, pp. 1429 -1431, 2012.

[14] M. A. Nisirat, S. Al-Khawaldeh, M. Ismail, L. Nissirat, “Hata based propagation loss formula using terrain criterion for 1800 MHz,” DOI: 10.1016, Elsevier, AEUE journal, 03.001. 2012.

[15] N. A. Mansour, “Rf Predictions and Modeling for Microcells and Pcs Cell Design,” IEEE, Vehicular Technology conference, pp. 1745 – 1749, 1994.

[16] P. E. Magensen, P. Eggers, C. Jensen, and J. B. Andersen, “Urban Area Radio Propagation Measurements at 955 and 1845 MHz for Small and Micro Cells,” IEEE, GLOBECOM 91, pp. 1297 – 1302, 1991.

[17] R. Mardeni, and K. F. Kwan, “Optimization of Hata Propagation Prediction Model In Suburban Area in Malaysia,” Progress In Electromagnetics Research C, pp. 91 - 106, 2010.

[18] T. S. Rappaport, Wireless Communications Principles and Practice, 2nd ed., Prentice Hall: New Jersey, 2002, pp. 105-177.

[19] Siwiak K., Radiowave propagation and Antennas for Personal Communications, 2nd ed., Boston Artech House Publishers, 1998.

[20] A. Neskovic, N. Neskovic, and G. Paunovic, “Modern Approaches In Modeling of Mobile Radio Systems Propagation Environment,” IEEE Communications Surveys, pp. 2 - 12, 2000.

277