Rain Fade Slope Analysis for Terrestrial Microwave Link in Malaysia

Md. Rafiqul Islam, Hassan Dao and Al Hareth Zayuod Wireless Communication and Signal Processing Research Group

Department of Electrical and Computer Engineering, Kulliyyah of Engineering International Islamic University Malaysia, Jalan Gombak, KL 53100

E-mail: rafiq@iiu.edu.my

Abstract— This paper analyzes fade slope during raining events based on measured data on terrestrial microwave link in Kuala Lumpur operating at 15GHz and 23 GHz. Fade slope is examined on a statistical basis for the two sites. The annual statistics of fade slope are presented and compared to Van de Kamp model which is recommended by ITU-R. Keywords— Fade Dynamics, Fade slope, Attenuation

I. INTRODUCTION In an equatorial region like Malaysia, microwave

communication is suffered from high attenuation due to rain that has rainfall rate at 0.01% of time is 126 mm/h (R0.01 = 126 mm/h) [1]. Since heavy rain occurs for a small percentage of the time, links in high rain rate regions could operate on lower margins when it is not raining. The transmitter power will be adaptively increased to operate at a higher level to counter fading, when it is required [2].

For development of adaptive fade countermeasure algorithms required knowledge of the dynamic behaviour of rain attenuation to assess the required speed which the system can track attenuation changes [3]. Fig. 1 shows definition of fade dynamics which are fade slope, fade duration, interfade duration at a certain fade threshold [4].

Fig.1. Features characterising the dynamics of fade events [4].

Fade slope is an interested aspect of dynamic behaviour

that defined as the rate change of attenuation time series. A system designer can utilise this information to develop fade compensation techniques such as Adaptive power control (APC) and forward error-correction (FEC) [2].

This paper presents an analysis of rain fade Slope between two terrestrial Line-of-Sight (LOS) links operating at 15 GHz

and 23 GHz in Kuala Lumpur, Malaysia. Comparison of rain fade slope is determined some second-order statistics of fade slope for both frequencies.

The next section briefs on the research experimental setup. Section III presents statistical analysis while section IV shows our results and discussion and conclusion is given in section V.

II. EXPERIMENTAL SETUP Measured data was collected from Digi telecommunication located at Indah Villa (IV) Condominium nearby Sunway Lagoon and Kolej Damansara Utama (KDU). The received power was determined from the measured Automatic Gain Control (AGC) level and sampling every 1 second interval during rain events. The parameters of these links are shown and summarised in table I [5]. In order to define the zero dB attenuation, the instantaneous attenuation of any microwave link can be determined from the RF signal level [6].

Table I Parameters for the Links

Link Frequency (GHz) Pol. Length

(Km) Measurement

Periods IV 15 V 3.96 Jan 1999 –

June 2000 KDU 23 V 0.91 Jan 1999 –

June 2000

III. STATISTICAL ANALYSIS Received signals have two components such as rain

attenuation and scintillation during rain events. Therefore, fast fluctuation due to scintillation cause is necessarily to filter before analysis. Since the slope was obtained from raw attenuation Ar(t), the filter was used to remove scintillation caused by moving average as shown in equation 1. In this experiment, ta 10 second moving average was applied.

( ) ( )tAt

tAa

a

t

ttr

a∑−

= 1 (1)

Once the rapid component has been removed, fade slopes

are calculated for each attenuation threshold. Fade slope ζ(t) is defined as the rate of change of rain attenuation A(t) as equation 2 [7],

Proceedings of the 2009 IEEE 9th Malaysia International Conference on Communications 15 -17 December 2009 Kuala Lumpur Malaysia

978-1-4244-5532-4/09/$26.00 ©2009 IEEE 250

( ) ( ) ( ) ( )sdBt

tAttAtΔ

−Δ+=ς (2)

Where A is attenuation (dB) and Δt is sampling time

interval (second) which is 1 second. Recently, ITU-R [4] proposed a new draft model for

prediction of a fade slope. For conditional probability (probability density function) that the fade slope is equal to ζ for a given attenuation value, A is calculated by,

sdBAp 22 ))/(2)|(

ζζ σζπσζ = (3)

Where ξσ is the standard deviation of the conditional fade slope at a given attenuation level A given by, dB/s),( AtfFs B Δ=ζσ (4)

Where F function which gives the dependence on the time interval length Δt and the 3 dB cut-off frequency of the low pass filter fB. This filter is replaced by the average of sample in the block of length ta around it. The effective bandwidth as a function of block length ta is given by [8] as

10445.0=Bf (5)

While

( ) bbbB

Btf

tfF /1

2

)2(/1

2),(Δ+

π=Δ (6)

With b = 2.3, and s parameter depends on climate and elevation angle which is 0.0452 obtained from [9].

IV. RESULT AND DISCUSSION Rain Attenuation A and fade slope ζ based on Indah Villa

(IV) link is shown in Fig.1, on 5th January 2000. High attenuation occurred during heavy rain fall between 14:45pm - 17:51pm. The highest attenuation occurring during the time 33.17dB and -3.84dB/s and 1.351 dB/s for fade slope in negative and positive respectively.

The conditional probability density of fade slope determined from the measured data for different attenuation levels (1dB, 3dB, 5dB, 8dB, 12dB, 15dB and 20dB) are presented in Fig. 3 at 23 GHz. Fade slope distribution becomes broader when attenuation levels increases. These density functions are generally symmetrical around zero for positive and negative slopes. The time symmetry is also confirmed by calculating the absolute values of the median fade slope, which are very close to zero for attenuation levels of 1, 3, 5, 8, 12, 15, and 20 dB.

(a)

(b)

Fig. 2 (a) Rain Attenuation (dB) and (b) Fade slope (dB/s) based on Indah Villa link, 5th January 2000

Figures 4-8 present the conditional probability density

function of fade slope determined for the measured data for different attenuation levels compared with the curves determined from the Van de Kamp model. Fig.5 and Fig.7 are for IV link and it is clear that the curves are not fit to each other. The same result for KDU link is presented in Fig. 4 and Fig. 8. In other hand, Fig. 6 shows that Van de Kamp model fits to our data only in the case of 5 dB attenuation.

Fig. 3 Conditional Probability Density of Fade slope for attenuation levels

on KDU link.

Fig. 4 The Conditional Probability Density of Fade slope for 1 dB

attenuation level on link KDU with Van de Kamp model.

251

Fig. 5 The Conditional Probability Density of Fade slope for 3 dB

attenuation level on link IV with Van de Kamp model.

Fig. 6. The Conditional Probability Density of Fade slope for 5 dB

attenuation level on link KDU with Van de Kamp model.

Fig. 7 The Conditional Probability Density of Fade slope for 15 dB

attenuation level on link IV with Van de Kamp model.

Fig. 8. The Conditional Probability Density of Fade slope for 20 dB

attenuation level on link KDU with Van de Kamp model.

Fig. 9 Standard Deviation versus Attenuation Comparision

among IV link, KDU link and Van De Kamp model

Standard deviation of fade slope was measured in both links and compares to the model in dB/s which varying on attenuation (dB). To find the relation between these the value of σ was calculated from the conditional distribution of ζ. Figure 9 presents standard deviation of fade slope as a function of attenuation A. The standard deviation of both links has got similar shapes while from model is not fit with our data.

The mean value of fade slope is measured and tends to increase with attenuation in positive side as similar to the standard deviation. Fig.10 shows statistics of mean in dB/s in terms of attenuation from 1-20 dB and linear fitted with measurement for both links. The mean tend to increase proportional, in KDU link mean range between 0.005619 dB/s and 0.09025 dB/s and from -0.021 dB/s till 0.01387 dB/s for IV link.

Fig. 10 Mean statistic varying with Attenuations

V. CONCLUSION Rain fade data from two terrestrial LOS microwave links operating at frequencies 15 GHz and 23 GHz have been analyzed for fade slopes in Kuala Lumpur, Malaysia. Correlation between the mean values and standard deviation of fade slope and attenuation has been noticed. Standard deviation and mean of fade slope occurrence increases with attenuation. Fade slope at high attenuations also were analysed. Positive and negative fade slopes are symmetrically

252

distributed and the distribution becomes flatter at higher attenuation levels.

The Van De Kamp model is not fit to our measured data especially at lower attenuation threshold. This model should be modified in order to use in Malaysian tropical climate specially for terrestrial microwave links.

ACKNOWLEDGMENT The authors are very grateful to Wireless Communication Centre, Universiti Teknologi Malaysia for providing rain attenuation data and also to Research Management Centre, International Islamic University Malaysia for supporting the project by research grant Ref: IIUM/504/RES/ G/14/3/07/ EDW B 0803-96.

REFERENCES

[1] J.S. Mandeep, “Equatorial Rainfall Measurement on Ku-Band Satellite Communication Downlink”, Progress in Electromagnetics Research, PIER 76, 2007.

[2] G. Liu, J.T. Ong, E. Choo and C.L. Law, “Fade slope for four LOS links in Singapore analyses and prediction”, Electronics Letters, Vol. 38, P. 425-426, April 2002.

[3] H. Dao, MD. Rafiqul, K. Al-Khateeb, “Fade Dynamics review of Microwave Signals on Earth-Space Paths at Ku-Band”, Proceeding ICCCE08, May 2008.

[4] ITU-R P.1623, “Prediction method of fade dynamics on Earth-space paths”, ITU, Geneva, Switzerland, 2003.

[5] Mutiara, “Links Locations over Peninsular Malaysia”, Technical Note from Mutiara Planning Division, Kuala Lumpur, 1997.

[6] R. Singliar, B. Heder, and J. Bito, “Rain Fade Slope Analysis”, BROADWAN Published Report, Paper W03A.01, December 2005.

[7] L. Castanet and M.V. de Kamp, “Modelling the Dynamic properties of the propagation channel”, 5th Management Committee Meeting of the COST 280 Action, May 2003

[8] M.J.L. Van de Kamp, “Statistical Analysis of Rain Fade Slope”, IEEE Transac. On Antenna and Propagation, Vol.51, No.8, August 2003.

[9] F.F. Franklin, K. Fujisaki, and M. Tateiba, “Fade Dynamics on Earth-Space Paths at Ku-Band in Fukuoka, Japan Fade-Slope Evaluation, Comparison, and Model”, IEEE Antenna and Wireless Propagation Letters, Vol. 5, 2006.

253