Impact of Sectorization/Vehicular Traffic on Minimum Cell Size for Information Capacity Increase in Cellular Systems

Transcription

1 ELECTOICS, VOL. 7, O., JUE 03 Ipact of Sectorization/Vehicular Traffic on Miniu Cell Size for Inforation Capacity Increase in Cellular Systes Kwashie A. Anang, Predrag B. apajic, and uiheng Wu Abstract In this paper results of atheatical analysis supported by siulation are used to study the ipact of sectorization/vehicular traffic on the theoretical liit for cell size radius reduction in cellular wireless counication systes. Inforation capacity approach is used for the analysis. Attention is given to the active co-channel interfering cells. Because at carrier frequencies greater than GHz, co-channel interfering cells beyond the first tier becoes doinant as the cell size radius reduces. esults show that for sectorized cellular wireless counication syste operating at carrier frequency greater than GHz and having saller cell size radius in a traffic environent the second tier co-channel interference still becoes active. This causes a decrease in the inforation capacity of the cellular wireless syste. For exaple for a heavy vehicular traffic environent, at a carrier frequency f C = 5.75 GHz, basic path loss exponent α = and cell radius = 00, 300 and 500 for a six sector cellular the decrease in inforation capacity, because of interference fro the second tier was 5.47, 3.36 and.78%. Index Ters Breakpoint distance, land obile radio cellularsyste, radio propagation, spectru efficiency, sectorization. Original esearch Paper DOI: 0.75/ELS370A T I. ITODUCTIO HE radio frequency spectru is an iportant paraeter in the design and ipleentation of a wireless counication syste. Because it is liited and regulated by international agreeents [], []. Cellular wireless systes are therefore partly used to achieve spectral efficiency and have been in operation since the late 970 s. A high overall spectral efficiency is achieved at the frequency planning level by reducing the cell size radius [3]. educing cell size radius has caused cell sites to be installed in Manuscript received 0 April 03. Accepted for publication 7 May 03. This is an expanded version of work presented at the IX International Syposiu Industrial Electronics IDEL 0, Banja Luka, Bosnia and Herzegovina, oveber 0 and at the 5th IEEE European Syposiu on Coputer Modeling and Siulation EMS 0, Madrid, Spain, oveber 0. K. A. Anang, P. B. apajic and. Wu are with the Wireless and Mobile Counication esearch Centre, School of Engineering, University of Greenwich, Chatha Maritie, ME4 4TB, UK, Fax : +44-(0) The corresponding author, K. A. Anang can be reached on Tel: +44-(0) , and E-ail: k.anang@gre.ac.uk. ever increasing densities [4]. However, Zhou et al. reported that there ay be a liit to cell size radius reduction [5], because of an increase in co-channel interference. Since co-channel interference is one of the ultiate factors which deterines the bit error rates (BEs) available to a user. The rapid developent of high-speed data rate wireless counication syste by service providers and the need for high-bit-rate services at obile terinals have spurred the use of broadband channels in wireless counication systes. Thus the UHF bands (900 and 900 MHz) norally used for cellular wireless counication are not suitable for wireless broadband application. For broadband channels carrier frequency needs to be increased [6]. Therefore, future and eerging cellular wireless counication systes beyond the third generation (B3G) will be accoodated at carrier frequencies greater than GHz [6] [8]. Increasing carrier frequency leads to an increase in free space path loss and diffraction loss. An increase in the path loss eans cell size radius needs to be reduced to saller radius. For saller cell size radius cellular syste co-channel interference becoes severe and ore difficult to control [6]. uerous studies on cellular wireless counication systes have given ranges of axiu and iniu cell size radius for inforation capacity increase [5], [9] []. Most of these studies to proceed analytically took into account co-channel interference fro the first tier, assuing interference outside the first tier to be negligible. Because of the assuption of large path loss exponent [5]. esults fro a study by Anang et al. shows that at higher icrowave carrier frequencies greater than GHz co-channel interference outside the first tier, (second tier) becoes active and it was reported that there is a theoretical liit to cell size radius reduction []. However, the study was for a non-sectorized cellular wireless syste. A study on the ipact of vehicular traffic on the inforation capacity of a non-sectorized cellular wireless syste, operating at carrier frequencies greater than 3 GHz was presented in [3]. The ipact of cell sectorization on the inforation capacity of cellular wireless networks was presented in [4], without the inclusion of the effect of vehicular traffic. However, at higher carrier frequencies and saller cell size radius; vehicles, pedestrians and other objects on the road affect the

2 ELECTOICS, VOL. 7, O., JUE 03 inforation capacity perforance of cellular wireless systes [7]. Therefore, the ain contribution of this paper is as follows: We study the ipact of cell sectorization on the inforation capacity perforance of future and eerging cellular wireless systes, which will be operating at higher icrowave carrier frequency greater than GHz and saller cell size radius, where first and second tier co-channel interference are doinant. We study the ipact of cell sectorization/vehicular traffic on the inforation capacity of future and eerging cellular wireless systes, which will be operating at higher icrowave carrier frequency greater than GHz and saller cell size radius, where first and second tier co-channel interference are active. The rest of the paper is organised as follows. Section II describes the syste odels for propagation, secotrized co-channel interference, user distribution and outlines the basic assuptions used in the odeling. Section III focus on the spectral efficiency of the cellular wireless syste used for our inforation capacity analysis. Section IV presents theoretical analysis and siulation results for the ipact of cell sectorization/vehicular traffic on the inforation capacity. Finally, we conclude this paper in Section V. II. POPAGATIO AD SYSTEM MODELS A two-diensional hexagonal saller cell size radius network is assued, where the BSs are uniforly distributed. Cells for clusters (co-channel cell) around reference cells. BSs located at the center of each cell receive signals fro all users in the syste which is attenuated according to the power-law path loss. A. Users Distribution The cell shape is approxiated by a circle of radius, for atheatical convenience. It is assued that all obiles, (desired and interfering users) are uniforly and independent distributed in their cells. Mobile stations (Ms) are also assued to be located in the far field region. The probability distribution function (PDF) of a MS location relative to a BS in polar co-ordinate is given by ( r 0 ) r, θ ( r, θ ) = ; 0 r,0 θ π, () π ( ) 0 where 0 corresponds to the iniu distance a obile can be fro the BS antenna (to be in the far field region), which defines a sall circular area around the MS to be kept free fro interferes. A reasonable value around 0 is recoended for saller cell size radius systes. B. Propagation Path Loss The radio environent of a cellular syste is described by: () path loss, () shadowing and (3) ultipath fading. For the purposes of this study, we ake the siplifying assuption that shadowing and ultipath fading is negligible. That is leaving only the variation of averaged received power with distance. The analysis and siulation uses the two-slope path loss odel [0], to obtain the average received power as function of distance. Fro this odel the average received signal power P r [W] is given by: K Pr = P t, () α r ( + r / g) where K is the constant path loss factor, and it is the free space path loss at a reference distance r 0 =, r [] is the distance between BS and MS. α is the basic path loss exponent (roughly ), is the additional path loss exponent (between -8). P t [W] is the transitted signal power. The breakpoint distance, g = 4 h b h /λ c, where λ c is the carrier wavelength. BS antenna height h b = 5, and MS antenna height h =.5. In this work the exact value of K and P t is not required for the analysis. Therefore we assue K =, P t = and focus on the attenuation factor α P t = r ( + r / g). (3) C. Sectorized Two Tier Co-Channel Interference The first and second tiers of co-channel interference are considered for interference generation. The desired obile is located in the central cell and the interfering obiles are in cells in the first and second tiers as shown in Fig.. To siplify the analysis the following assuptions have been ade in the co-channel interference odel. First the syste is considered to be interference-liited, with theral noise power negligible relative to the co-channel interference power [5]. Therefore, the ratio of carrier to noise C reduces to the carrier-to-interference power ratio CI. All inter-channel interference is considered to be negligible [5]. All BSs are assued to transit the sae power, and for siplicity we assue each cell to be circular shape. Fro [], for an onidirectional antenna cell site layout pattern the nuber of co-channel interfering cells in a given tier n is given by n = I n; ( n =,,3,4,...), (4) Fig.. Onidirectional, secotorized cellular wireless counication systes showing first and second tier co-channel interferers.

3 ELECTOICS, VOL. 7, O., JUE 03 3 where I is the nuber of interfering cells in the first tier and n is the nth tier nuber and it is always an integer. ow for sectorized cells (direction antennas), (4) is odified as follows: I n n = ; ( n =,,3,4,...), (5) S where S is the nuber of sectors in the cell. For onidirectional cellular syste, S =, for 0 o and 60 o sectorized cellular syste S = 3 and 6. eference [6] stated that the uplink interference at a served BS is the non-coherent su of interference signals fro the user served by the BS and the users served by other BSs. Likewise the desired user CI, γ, is defined as the ratio of averaged received signal power fro a MS at a distance r [] fro the desired BS to the su of interfering received signal power. Therefore, the desired user CI, γ, can be written as follows: Pd Pd ( r) γ = =, (6) I / S I / S PI P ( r ) + P ( r ) i= i i i= i i where P d [W], is the received power level of desired MS and P I [W] is the power su of individual interferers in tiers and. I and I is the nuber of co-channel interfering cell in tiers and of an onidirectional cellular syste. For hexagonal cell site layout with cluster size c = 7, I = 6 and I =. P i and P i [W] is the average power level received fro the ith interfering MSs at distances r i and r i [] fro the desired BS. III. AEA SPECTAL EFFICIECY The ultiate capacity of a land obile radio syste is directly related to its spectral efficiency [7]. The spectral efficiency of a cellular wireless syste can be expressed in a nuber of ways such as nuber of channels per cell, Erlangs/k, the nuber of users/k, etc. However in this paper, we adopted the definition suggested by [8]. This definition gives a ore coplete picture of the spectru efficiency by expressing it in ters of capacity, bandwidth, and area. The area spectral efficiency (ASE) is defined as the achievable su rate [bits/sec] (of all users in a cell) per unit bandwidth per unit area which is given by [8] as: A e s k = = C k, (7) πw (D / ) where W is the total bandwidth allocated to each cell, D is the reuse distance, s is the total nuber of active serviced channels per cell. The achievable su rate C k is the Shannon capacity of the kth user, which depends on γ, the received carrier to interference power ratio CI of that user and W k the bandwidth allocated to the user. The Shannon capacity forula assues the interference has Gaussian characteristics. Because both the interference and signal power of the kth user vary with obiles locations and propagation conditions, γ varies with tie, therefore the average channel capacity of the kth user is given by [8] as + 0 C k = Wk log ( + γ ) pγ ( γ ) dγ, (8) where p γ (γ), is the probability distribution function (PDF) of the average ean CI(γ) of the kth user. The transission rate is assued to be continuously adapted relative to the CI in such a anner that the BE goes to zero asyptotically. In (8) assuing that all users are assigned the sae bandwidth, C k = ( C ) becoes the sae for all users, therefore A e can be written as 4 s C 4 s C Ae = =, (9) πwd πw u where u is defined as the noralized reuse distance and is given by the ratio of reuse distance and cell radius (D/). For a TDMA syste, the total bandwidth is allocated to only one active user per tie slot, (that is =, W k = W). Substituting this into (9) yields + 4 log ( + γ ) ( γ ) u 0 Ae = p γ dγ. (0) π IV. SECTOIZED IMPACT AALYSIS In this section, we analyse the ipact of sectorization on the inforation capacity perforance of saller cell size radius cellular syste operating at carrier frequency greater than GHz, in the presence of first and second tier co-channel interference. The analysis applies to a TDMA, (tie-division ultiple access) based cellular wireless syste. Because, it is the ost representative of cellular wireless syste. The analysis is based on fully loaded systes with fixed cluster size c = 7. Though there is an excessive deand to broadcast, (downlink) high speed data in eerging counication services, because of space we confine our study on the uplink between a MS-to-BS. A. Analysis ecall fro section II-A; user are randoly located in their respective BSs, therefore γ is a rando variable which depends on the rando position of the user and the sus of interference fro tier and. Without power control the average-case interference configuration corresponds to the case, where all the I and I co-channel interferes are at the center of their respectively BSs, at a distances r i = D [] and r i = D [] Fig.. Sectorized cellular syste geoetry of the desired and interfering obile in two co-channel cells.

4 4 ELECTOICS, VOL. 7, O., JUE 03 fro the desired MS s as shown in Fig.. ote power control is essential for direct sequence CDMA systes; therefore we did not consider it in our analysis. Assuing that the transitted power of all users is the sae and substituting (3) into (6) yields Pd ( r) γ ( r, I, I ) = I / S I / S P ( r ) + P ( r ) = i= i i α ( + r / g) I / S α I / S τ ( + τ / g) + i= i= r i= i i (τ ) α ( + (τ ) / g) () α τ g S I = r g + r α g g + I I g + τ g + τ where τ is the product of u and, u is the noralized reuse distance and is the cell size radius. S is the nuber of sectors in the cell. Because γ is a function of r, the desired user capacity is given by C r,, ) = W log ( + γ ( r,, )). () ( I I 0 I I Substituting () in (9) yields the ASE conditioned on the de-sired obile position r, for a fully-loaded syste. Integrating () over the desired user s position PDF () yields the average ASE for the average interference configuration as: 4 Ae ( r, I, I ) = log ( + γ ) π r u 0 ( r) dr. (3) It is clear fro (3) that the average ASE ainly depends on the ean CI, which is a function of rando locations of the MS. This akes the ASE atheatically intractable to solve. A coputer siulation is therefore used to solve it. B. Siulations Monte Carlo siulation is used to estiate A e, because it appears to be atheatically intractable to explicitly solve analytically. Fig. 3 shows a siulation flowchart for the ASE of a sectorized cellular syste, and the basic paraeters used for the siulation are presented in Table I. In the siulation the desired user is randoly located, and uniforly distributed as described in subsection II-A of section II. When the desired Paraeter TABLE I SIMULATIO PAAMETES Value Type of syste oni, 3-sector and 6-sector Cell radius, 00 to 000 Effective road height, h 0.3 and.3 [7], [9] Path loss exponent,(α), Additional path loss exponent, () Cluster size, C 7 BS antenna height, h B 5 [9] MS antenna height, h M.5 [0] Mobile Distribution Unifor/ando uber of co-channel tiers Co-channel interferences ando and first and second tiers Frequency reuse factor, U 4 [8] Frequencies, f C 0.9,, 3.35, 8.45 and 5.75 GHz Fig. 3. Flowchart for sectorized cellular syste area spectru efficiency siulation. user position is located the siulation algoriths is coposed of the following steps: ) The polar coordinates (x i, θ i ) and (x i, θ i ) of the I and I co-channel interferes are randoly picked according to (). ) Fro Fig., (geoetry for analysis) the distance r i for each co-channel interferer fro tier to the desired BS is calculated as r i i i i = D + x Dx cos( θ ). (4) 3) The distance r i for each co-channel interferer fro the second tier to the desired BS is calculated as r i ( i i i = D) + x 4Dx cos( θ ). (5) 4) The two-slope path loss odel (), is used to calculate the average received signal power of the desired user and interfering obiles in the first and second tier of co-channel cells (P d, P i s and P i s), therefore the CI is calculated as γ =. α I / S + I / S r ( g r) + i= α i= α ri ( g + ri ) ri ( g + ri ) (6) 5) The ASE, A e is calculated as

5 ELECTOICS, VOL. 7, O., JUE Ae = log ( + γ ). (7) π u epeating the proceed above (fro steps -5) after locating the desired user position. A e, is estiated by taking the average of all the observations of A e as given by (7). C. uerical and Siulations esults Figs. 4, 5 and 6 show plot of ASE as a function of cell size radius for oni-directional, three sectors and six sector cellular systes. The figures quantified the fact that sectorization reduces co-channel interference, thus iproves CI, which causes an increase in inforation capacity of the cellular wireless systes. The curves in Fig. 4 show the plot for an oni-directional cellular syste for different carrier frequency f c, using the interference odel presented in [8], and the odel presented in this work (). The curves show that when f c = 900 GHz and = 0. k, the decrease in inforation capacity was 6%. ow for f c =, 3.35, 8.45 and 5.75 GHz, the decrease in inforation capacity was 8.55, 0.5, 3.73, and 5.3%. At = 0.3 k for f c = 0.9,, 3.35, 8.45 and 5.75 GHz, the decrease in inforation capacity was 3.7, 4.9, 6.5, 9.8, and.39%. In the case of = 0.5 k for f c = 0.9,, 3.35, 8.45and 5.75 GHz, the decrease in inforation capacity was 3.7, 4.0, 4.88, 7.4 and 9.4%. The curves in Fig. 5 show the case of a three sector cellular wireless counication syste. The curves show that for f c = 0.9,, 3.35, 8.45 and 5.75 GHz at cell radius =0. k, Fig. 5. Average uplink Area Spectral Efficiency (ASE) versus cell radius for three sector cellular syste at different carrier frequencies f c. (Fully-loaded syste with 6 and co-channel interfering cells in first and second tier I = 6 and I = ; basic and extra path loss exponent: α = and = ; MS and BS antenna heights: h =.5 and h b = 5 ). - single tier interfering odel (f c = 900 MHz), - two tier interfering odel (f c = 900 MHz), 3 - single tier interfering odel (f c = GHz), 4 - two tier interfering odel (f c = GHz), 5 - single tier interfering odel (f c = 8.45 GHz), 9 - single tier interfering odel (f c = 5.75 GHz), 0 - two tier interfering odel (f c = 5.75 GHz). Fig. 4. Average uplink Area Spectral Efficiency (ASE) versus cell radius for onidirectional cellular syste at different carrier frequencies f c. (Fully-loaded syste with 6 and co-channel interfering cells in first and second tier I = 6 and I = ; basic and extra path loss exponent: α = and = ; MS and BS antenna heights: h =.5 and h b = 5 ). - single tier interfering odel (f c = 900 MHz), - two tier interfering odel (f c = 900 MHz), 3 - single tier interfering odel (f c = GHz), 4 - two tier interfering odel (f c = GHz), 5 - single tier interfering odel (f c = 8.45 GHz), 9 - single tier interfering odel (f c = 5.75 GHz), 0 - two tier interfering odel (f c = 5.75 GHz). Fig. 6. Average uplink Area Spectral Efficiency (ASE) versus cell radius for six sector cellular syste at different carrier frequencies f c. (Fully-loaded syste with 6 and co-channel interfering cells in first and second tier I = 6 and I = ; basic and extra path loss exponent: α = and = ; MS and BS antenna heights: h =.5 and h b = 5 ). - single tier interfering odel (f c = 900 MHz), - two tier interfering odel (f c = 900 MHz), 3 - single tier interfering odel (f c = GHz), 4 - two tier interfering odel (f c = GHz), 5 - single tier interfering odel (f c = 8.45 GHz), 9 - single tier interfering odel (f c = 5.75 GHz), 0 - two tier interfering odel (f c = 5.75 GHz).

6 6 ELECTOICS, VOL. 7, O., JUE 03 the decrease in the inforation capacity between the two interference odel was 4.6, 6.5, 7.94, 0.36 and.56%. For 0.3 k at carrier frequencies f c = 0.9,, 3.35, 8.45 and 5.75 GHz, the decrease in ASE was 3, 3.9, 4.8, 7.09 and 8.7%. For 0.5, the decrease was.6, 3., 3.85, 5.6 and 7.8%. We can therefore conclude that for a three sector cellular wireless counication syste as the carrier frequency increases and cell size radius reduces, second tier co-channel interference becoes severe. The curves in Fig. 6 show the case of a six sector cellular wireless counication syste. The curves show that for f c = 0.9,, 3.35, 8.45 and 5.75 GHz, at cell radius = 0. k, the decrease in the inforation capacity between the two interference odel was 4.0, 5.57, 6.77, 8.8 and 0%. For 0.3 k at carrier frequencies f c = 0.9,, 3.35, 8.45 and 5.75 GHz, the decrease in ASE was.64, 3.4, 4.7, 6.05 and 7.43%. For 0.5, the decrease was.3,.83, 3.35, 4.89 and 6.4%. We can conclude that for a sectorized cellular wireless counication syste as the carrier frequency increases and cell size radius reduces, second tier co-channel interference becoes doinant. Tables II, III and IV show the results of percentage decrease in ASE between the two interference odels; for different cellular network sectorization; carrier frequency f c and cell size radii : 0., 0.3 and 0.5 k. D. Cobined Effect of Sectorization and Vehicular Traffic In this section, we consider the ASE of a fully loaded sectorized cellular syste in a light/heavy vehicular traffic environent, by including effective road height in the two-slope path loss odel. This is the scenario for an urban line-of-sight (LOS) environent, when carrier frequencies are greater than GHz. ) Analyses-Modified Breakpoint Distance: The odified breakpoint distance proposed by Masui et al. [7], is incorporated into the syste odel, (that is the path loss odel), for the study of ipact of vehicular traffic on the inforation capacity perforance of a sectorized cellular wireless syste. The odified breakpoint distance g is given by [7], as TABLE II DECEASE I ASE BETWEE THE TWO ITEFEECE MODEL: OMIDIECTIOAL CELLULA SYSTEM h =.5, h b = 5 and α = 900 MHz 6 GHz GHz GHz GHz MHz 3.7 GHz GHz GHz GHz MHz 3.7 GHz GHz GHz GHz 9.4 TABLE III DECEASE I ASE BETWEE THE TWO ITEFEECE MODEL: THEE-SECTO CELLULA SYSTEM h =.5, h b = 5 and α = 900 MHz 4.6 GHz GHz GHz GHz MHz 3 GHz GHz GHz GHz MHz.6 GHz GHz GHz GHz 7.8 TABLE IV DECEASE I ASE BETWEE THE TWO ITEFEECE MODEL: SIX-SECTO CELLULA SYSTEM h =.5, h b = 5 and α = 900 MHz 4.0 GHz GHz GHz GHz MHz.64 GHz GHz GHz GHz MHz.3 GHz GHz GHz GHz 6.4 4( h h) ( h h) b g =, h < λc h, (8) where h is the effective road height, which is due to vehicles, pedestrians and other objects on the road. ow, h depends on the average height of traffic on the road, which is the average height of vehicles and pedestrians height on the road [7]. For light vehicular traffic the value of h is between 0.3 and 0.74, and for heavy vehicular traffic, it is between.9 and.64 [9]. For the siulation of the cobine effect of sectorization and vehicular traffic on the inforation capacity perforance of the cellular wireless syste, step 4) of the algorith described in Section IV-B is changed as follows to incorporate the odified breakpoint. γ = α I / S + I / S r ( g r) + i= α i= α ri ( g + ri ) ri ( g + ri ) E. Siulations esults Sectorized/Vehicular Traffic (9) The cobined effect of sectorization and vehicular traffic

7 ELECTOICS, VOL. 7, O., JUE 03 7 on the ASE of a saller cell size radius cellular syste, operating at different carrier frequencies f c is shown in Figs The results show that the ASE curves conserve the sae relative shape as Figs. 5 and 6. Coparing the figures, it can be seen that both secotrized cellular systes operating in traffic environent have higher area spectru efficiency than the secotrized cellular systes when there is no traffic. The ASE of a three sector cellular syste for light and heavy vehicular traffic is shown in Figs. 7 and 8. The curves show that for light vehicular traffic, and carrier frequencies f c = 0.9,, 3.35, 8.45 and 5.75 GHz, at cell size radius = 0. k, the decrease in the inforation capacity between the two interference odels was 4.9, 6.0, 7.4, 9.94 and.6%. For 0.3 k, the decrease in the inforation capacity between the two interference odels was.8, 3.63, 4.44, 6.56 and 8.%, and for 0.5 k, it was.5, 3.00, 3.59, 5.9 and 6.8%. Fro the three sector and heavy vehicular traffic result, the decrease in inforation capacity between the two interference odels for carrier frequencies f c = 0.9,, 3.35, 8.45 and 5.75 GHz at cell size radius = 0. k, is.55,.97, 3.45, 4.9 and 6.40%. For 0.3 k, it is.5,.30,.49, 3.0 and 3.8%. For 0.5 k, the decrease was.9,.7,.7,.69 and 3.6%. These results are tabulated in Table V and VI. The ASE of a six sector cellular syste in a light and heavy vehicular traffic environent is shown in Figs. 9 and 0. The curves show that for light vehicular traffic, and carrier frequencies f c = 0.9,, 3.35, 8.45 and 5.75 GHz, at cell size Fig. 8. Average uplink Area Spectral Efficiency (ASE) versus cell radius for three sector cellular syste and heavy vehicular traffic at different carrier frequencies f c. (Fully-loaded syste with 6 and co-channel interfering cells in first and second tier I = 6 and I= ; basic and extra path loss exponent: α = and = ; MS and BS antenna heights: h =.5 and h b = 5 ; effective road height h =.3 ). - single tier interfering odel (f c = 900 MHz), - two tier interfering odel (f c = 900 MHz), 3 - single tier interfering odel (f c = GHz), 4 - two tier interfering odel (f c = GHz), 5 - single tier interfering odel (f c = 8.45 GHz), 9 - single tier interfering odel (f c = 5.75 GHz), 0 - two tier interfering odel (f c = 5.75 GHz). Fig. 7. Average uplink Area Spectral Efficiency (ASE) versus cell radius for three sector cellular syste and light vehicular traffic at different carrier frequencies f c. (Fully-loaded syste with 6 and co-channel interfering cells in first and second tier I = 6 and I= ; basic and extra path loss exponent: α = and = ; MS and BS antenna heights: h =.5 and h b = 5 ; effective road height h = 0.3 ). - single tier interfering odel (f c = 900 MHz), - two tier interfering odel (f c = 900 MHz), 3 - single tier interfering odel (f c = GHz), 4 - two tier interfering odel (f c = GHz), 5 - single tier interfering odel (f c = 8.45 GHz), 9 - single tier interfering odel (f c = 5.75 GHz), 0 - two tier interfering odel (f c = 5.75 GHz). Fig. 9. Average uplink Area Spectral Efficiency (ASE) versus cell radius for six sector cellular syste at different carrier frequencies f c. (Fully-loaded syste with 6 and co-channel interfering cells in first and second tier I = 6 and I = ; basic and extra path loss exponent: α = and = ; MS and BS antenna heights: h =.5 and h b = 5 ; effective road height h = 0.3 ). - single tier interfering odel (f c = 900 MHz), - two tier interfering odel (f c = 900 MHz), 3 - single tier interfering odel (f c = GHz), 4 - two tier interfering odel (f c = GHz), 5 - single tier interfering odel (f c = 8.45 GHz), 9 - single tier interfering odel (f c = 5.75 GHz), 0 - two tier interfering odel (f c = 5.75 GHz).

8 8 ELECTOICS, VOL. 7, O., JUE 03 TABLE VI DECEASE I ASE BETWEE THE TWO ITEFEECE MODEL: THEE-SECTO CELLULA SYSTEM/HEAVY VEHICULA TAFFIC h =.5, h b = 5, h =.3 and α = 900 MHz.55 GHz GHz GHz GHz MHz.5 GHz GHz GHz GHz MHz.9 GHz GHz GHz GHz 3.6 Fig. 0. Average uplink Area Spectral Efficiency (ASE) versus cell radius for six sector cellular syste at different carrier frequencies f c. (Fully-loaded syste with 6 and co-channel interfering cells in first and second tier I = 6 and I = ; basic and extra path loss exponent: α = and = ; MS and BS antenna heights: h =.5 and h b = 5 ; effective road height h =.3 ). - single tier interfering odel (f c = 900 MHz), - two tier interfering odel (f c = 900 MHz), 3 - single tier interfering odel (f c = GHz), 4 - two tier interfering odel (f c = GHz), 5 - single tier interfering odel (f c = 8.45 GHz), 9 - single tier interfering odel (f c = 5.75 GHz), 0 - two tier interfering odel (f c = 5.75 GHz). radius = 0. k, the decrease in the inforation capacity between the two interference odels was 3.73, 5.7, 6.33, 8.45 and 9.53%. For 0.3 k, the decrease in the inforation capacity between the two interference odel was.5, 3.9, 3.8, 5.66 and 7.05%, and for 0.5 k, it was.5,.69, 3.4, 4.54 and 5.87%. For the six sector and heavy vehicular traffic the decrease in inforation capacity between the two interference odels for carrier frequencies f c = 0.9,, 3.35, 8.45 and 5.75 GHz, at cell size radius = 0. k, is.9,.65, 3.03, 4.4 and 5.47%. For 0.3 k, the decrease in the inforation capacity TABLE V DECEASE I ASE BETWEE THE TWO ITEFEECE MODEL: THEE-SECTO CELLULA SYSTEM/LIGHT VEHICULA TAFFIC h =.5, h b = 5, h = 0.3 and α = 900 MHz 4.3 GHz GHz GHz GHz MHz.84 GHz GHz GHz GHz MHz.48 GHz GHz GHz GHz 6.76 TABLE VII DECEASE I ASE BETWEE THE TWO ITEFEECE MODEL: SIX-SECTO CELLULA SYSTEM/LIGHT VEHICULA TAFFIC h =.5, h b = 5, h = 0.3 and α = 900 MHz 3.73 GHz GHz GHz GHz MHz.5 GHz GHz GHz GHz MHz.5 GHz GHz GHz GHz 5.87 between the two interference odel was.0,.08,.3,.78 and 3.36%, and at 0.5 k, it was.87,.0,.07,.39 and.78%. These results are tabulated in Table VII and VIII. Fro these results, we can conclude that, for a sectorized cellular syste, operating at higher carrier frequency and having saller cell size radius in a traffic environent, the second tier co-channel interference still becoes doinant. The result confirs the need to include second tier co-channel interference in the inforation capacity perforance analysis of future and eerging cellular wireless counication systes. We can also conclude that the presence of vehicles and pedestrians in the environent tends to itigate the severity of the co-channel interference. V. COCLUSIO In this paper, because of the iportance of co-channel interference on the inforation capacity perforance of cellular syste, we have shown that even for sectorized cellular wireless syste operating at carrier frequency greater than GHz and saller cell size radius, in a traffic environent second tier co-channel interference still becoes doinant. Therefore there is a need to include second tier co-channel interference in the design and syste odel of

9 ELECTOICS, VOL. 7, O., JUE 03 9 TABLE VIII DECEASE I ASE BETWEE THE TWO ITEFEECE MODEL: THEE-SECTO CELLULA SYSTEM/HEAVY VEHICULA TAFFIC h =.5, h b = 5, h =.3 and α = 900 MHz.9 GHz GHz GHz GHz MHz.0 GHz GHz GHz GHz MHz.87 GHz GHz GHz GHz.78 eerging and future cellular wireless counication systes. Future work will focus on including ultiple tiers of co-channel interfering cells, correlation coefficient, shadowing and ultipath fading. In future we will also use ore realistic propagation and syste odel's scenario in ters of user's distribution and radio environent. EFEECES [] W. C. Y. Lee, Spectru efficiency in cellular, IEEE Trans. Veh.Technol., vol. 38, pp , May 989. [] K. Pahlavan and A. H. Levesque, Wireless data counication, proc.ieee, vol. 8, pp , Sept [3] Y. Yao and A. U. H. Sheikh, Investigations into co-channel interference in icrocellular obile radio systes, IEEE Trans. Veh. Technol., vol. 4, no., pp. 4, May 99. [4] T. K. Sarkar, Z. Ji, K. Ki, A. Medouri, and M. Salazar-pala, A survey of various propagation odels for obile counication, IEEETrans. Antennas Propagat., vol. 45, pp. 5 74, Jun [5] S. Zhou, M. Zhao, X. Xu, J. Wang, and Y. Yao, Distributed wireless counication syste: a new architecture for future public wireless access, IEEE Coun. Mag., vol. 4, pp. 08 3, 003. [6] J. Takada, J. Fu, H. Zhu, and T. Kobayashi, Spatio-teporal channel characterization in a suburban non line-of-sight icrocellular environent, IEEE J. Select. Areas Coun., vol. 0, no. 3, pp , April 00. [7] H. Masui, T. Kobayashi, and M. Akaike, Microwave path-loss odeling in urban line-of-sight environents, IEEE J. Select. Areas Coun., vol. 0, no. 6, pp. 5 55, Aug 00. [8] G. Hern andez-valdez, F. A. Cruz-p erez, and D. Lara-rodr ıguez, Sensitivity of the syste perforance to the propagation paraeters in los icrocellular environents, IEEE Trans. Veh. Technol., vol. 57, no. 6, pp , ov [9] V. H. Macdonald, The cellular concept, Bell Systes TechnologyJounarl, vol. 58, no., pp. 5 4, Jan [0] P. Harley, Short distance attenuation easureents a t 900 hz and.8 ghz using low antenna heights for icrocells, IEEE J. Select. AreasCoun., vol. 7, no., pp. 5, Jan [] Y. Liang, A. Goldsith, G. Foschini,. Valenzuela, and D. Chizhik, Evolution of base station in cellular networks: denser deployent versus coordination, ieee international conference on counications, pp , May 008. [] K. A. Anang, P. B. apajic, T. I. Eneh, and Y. ijsure, Miniu cell size for inforation capacity increase in cellular wireless network, in Proc. 73rd IEEE Vehicular Technology Conference (VTC'0), Budapest, Hungary, May 0, pp [3] K. A. Anang, P. B. apajic, T. I. Eneh, L. Bello, and G. Oletu, Ipact of vehicular traffic on inforation capacity of cellular wireless network at carrier frequencies greater than 3 ghz, in Proc. 5th IEEE EuropeanModelling Syposiu on Matheatical odelling and Coputer Siulation (EMS' 0), Madrid, Spain, ov. 0, pp [4] K. A. Anang, P. B. apajic, and. Wu, Ipact of sectorization on the iniu cell size for inforation capacity increase in cellular wireless network, in IX International Syposiu Industrial Electronics (IDEL'0'), Banja Luka, Bosnia and Herzegovina, ov. 0, pp. 0 5 [5] W. C. Y. Lee, Mobile counication design fundaentals. ew York, Y: John Wiley & Sons, 993, p. 4. [6] S. Singh,. B. Mehta, A. F. Molisch, and A. Mukhopadhyay, Moent atched lognoral odeling of uplink interference with power control and cell selection, IEEE Trans. Wireless Coun., vol. 9, no. 3, pp , Mar. 00. [7] D.. Hatfield, Measures of spectral efficiency in land obile radio, IEEE Trans. Electroagn. Copat., vol. ec-9, no. 3, pp , Aug [8] M. Alouini and A. J. Goldsith, Area spectral efficienc y of cellular obile radio systes, IEEE Trans. Veh. Technol., vol. 48, no. 4, pp , Jul [9] ITU, Propagation data and prediction ethods for planning of short range outdoor radio counication systes and radio local area networks in the frequency range 300 hz to 00 ghz, ecoendation ITU- P.4 -, itu adio counication Assebly. [0] G. T5.996, 3gpp sc channel odels, 3GPP T5.996, vol. v6..0, Sept. 003.