Energy-Efficient Cellular Counications Powered by Sart Grid Technology Itiaz Nasi, Mostafa Zaan Chowdhury, and Md. Syadus Sefat Departent of Electrical and Electronic Engineering Khulna University of Engineering & Technology, Khulna-9203, Bangladesh E-ail: itiaz_nasi@hotail.co, zceee@yahoo.co, sefat_2357@yahoo.co bstract The energy efficiency aspect of cellular networks is a vital topic of research over the recent days. s energy consuption is on the rise and the price of electricity is increasing very rapidly, necessity to reduce electricity usage in every aspect is becoing uch ore significant. The power grid infrastructure, fro which the cellular networks attain the required electricity for operation is considering a significant change fro the traditional electricity grid to the sart grid. The base stations, which are the ain candidates for energy consuption in cellular networks, reain in operation even when there are a very few users or no user at all. In this paper, we propose an effective application scenario of the fetocell technology for power saving purpose of cellular networks. Effective deployent of feto-access-points in the required places can reduce the power usage of cellular networks by turning off the redundant base stations. Moreover, the signal to noise plus interference ratio is also enhanced in coparison to the traditional schee. The base station turn off probability and the energy consuption in the overall network are analyzed. The coparison results show that the proposed energy efficient odel can significantly reduce energy consuption and increase the service quality of the cellular users. Keywords Sart grid, cellular network, energy consuption, QoS, and fetocell I. INTRODUCTION The sart grid (SG), also called sart power grid, intelligent grid, intelligrid, future grid or intragrid, is an enhanceent of the 20th century power grid. The traditional power grids are norally used to carry power in one way direction only. These power grids generally carry power fro a few central generators to a large nuber of users or consuers. On the other hand, SG creates an autoated and distributed advanced energy delivery network by using the two-way flow of electricity. The SG uses odern inforation technologies and is capable of distributing power in a ore efficient way. It also ephasizes on energy saving of the different sart appliances and coodities []. Recently, rapidly rising energy price has led to a trend of studying the energy efficiency aspect of counication networks. The cellular operators pay a handsoe aount of electricity bill which has becoe a significant portion of their total expenditure. The CO 2 eissions produced by wireless cellular networks all over the world are equivalent to those fro ore than 8 illion cars. In a typical wireless cellular network, base stations (BSs) consues 60 80% energy of the overall cellular network s energy consuption. lost 20,000 new base stations are deployed throughout the world every year to provide necessary support to the cellular users [2]. Our survey works suggest that a base station consues ore than 85% of its peak energy even when there is little or no traffic. Therefore, a dynaic operation of the cellular base stations based on the traffic is very uch essential for the energy efficient approach of cellular networks. The proposed odel in this paper is expected to reduce the energy consuption of cellular networks to a significant level. However, this odel is uch ore applicable for soe reote areas where there are a very few cellular users and a base station reains in operation continuously to serve those few users. Using the siulation results, we show that the proposed schee can save energy of cellular networks and iprove the quality of service (QoS) to a certain level. The rest of this paper is organized as follows. Section II shows the proposed energy efficient odel. Section III shows the perforance analysis of the proposed schee. Finally, conclusions are drawn in the last section. II. PROPOSED ENERGY EFFICIENT MODEL Suppose a cellular network is powered by sart grid where there is a flow of electricity and counication siultaneously as shown in Fig.. Figure. cellular network powered by the sart grid [2].
Figure 2. Coverage extension and energy saving with coordinated ultipoint (CoMP) [2]. In [2], the authors proposed a odel in which the redundant base station is turned off and the users of that network are served by coverage extension of the neighboring base stations using the coordinated ultipoint ethod for energy saving purpose. This odel is showed in Fig. 2. If the redundant base station is turned off as shown in Fig. 2, a decent aount of energy can be saved if the existing user is served by the neighboring base stations. But we propose another odel for saving energy of cellular networks using fetocell network which ay turn out to be a very proising solution for saving ore energy of cellular networks especially in reote areas. In our proposed odel, feto-access-points are installed in soe places which have the ost probability of carrying a obile user in the entire cellular network served by a base station. The feto-access-points (FPs) are low-power, sallsize hoe-placed base stations that enhance the service quality for the indoor obile users [3]. Fig. 3 shows an exaple of the proposed energy efficient odel. When all the users reain inside the fetocellular coverage region, the acrocellular BS is allowed to turn off. But if any user stays outside the fetocellular coverage, the acrocellular BS ust reain in operation to serve that particular user. Periodic checking of user s location and the QoS is tested by the dynaic checking application of the SG. Our purpose is not only saving energy of cellular networks but also to increase the service quality of the existing users. Feto-access-points are ideal for this case as these devices run with very low operating power and at the sae tie can enhance the QoS to a significant level. Figure 3. Proposed energy efficient odel with fetocell/macrocell integrated network.
checking of the position of each user inside the network. By adopting this process, the acrocellular BS will soeties reain turned on and soeties reain turned off. III. PERFORMNCE NLYSIS This section provides the probability analysis perforance and the signal quality perforance of our proposed odel.. Probability nalysis When a redundant BS is shut down and the users of that area are to be served using FPs, then it is required to analyze the probability of the existence of a user under the fetocell coverage area. If all users are within the fetocell coverage area then the BS ay reain shut down and service is provided to the user by the fetocell network, but if any user reains outside the fetocell coverage area, then the BS ust be turned on in order to provide service to that user. For siplicity in calculation, we have divided the total considered area into 4 sub-circles each with a fixed distance fro the center point of the ain circle. The coverage area of a acrocellular BS is assued to be 500. We perfored our analysis assuing 5 FPs in the total coverage area. ll the stations considered in the coverage area have a specific factor according to the iportance of that station for carrying users. The factors assued for each station is given in Table I. The coverage area of a FP is deterined by a=4 r 2 where r is the radius of each FP coverage. Figure 4. Flow chart showing the condition for the BS to be turned off in the proposed schee. Fetocells are sall base stations with low transission power, and with alost all cellular functionalities, but a liited coverage to end users in indoor environents. fetocell allows service coverage extended to indoor environents, especially where wireless access is liited or unavailable, without expensive deployent cost. flow chart is shown in fig. 4 which shows the algoriths of the proposed schee. The network will first check the existence of a user. If no user exists in the network, both the acrocellular BS and the FPs will reain off. If any users exist, then the network will search for user s location. When any of the users reain outside the FP coverage area, the BS will not be allowed to turn off. But, if all users exist inside the fetocellular network, then the network will check the signal strength available for the users if the BS is shut down. The BS will shut down if the signal strength is found within acceptable liit, otherwise; it will reain on in order to provide network support to the existing users. This odel requires a periodic If there exist stations with n nuber of factors then Let, is the area under user existence probability factor f, 2 is the area under user existence probability factor f 2, 3 is the area under user existence probability factor f 3, and n is the area under user existence probability factor f n. Let the reaining area where no FPs are installed is M with a user existence probability factor of f M The probability of a user to exist in the FP coverage area is given by P = f N f N i i + f + f N2 2 N2 2 2i 2i + f + f N3 3 N3 3 3i 3i +... + f +... + f Nn n Nn n ni ni + f where N is the nuber of places under user existence probability factor f, N 2 is the nuber of places under user existence probability factor f 2, N 3 is the nuber of places under user existence probability factor f 3, N n is the nuber of places under user existence probability factor f n. The probability of the BS to be turned off with the nuber of user can be expressed as ( ) n f p = (2) where f p is the average iportance factor considered for the entire region and n is the nuber of user. p ()
The analysis is perfored assuing only 5 FPs in the considered total area. The perforance would be better if ore nuber of FPs are used. However, FPs should be installed in the places where there is a big possibility of users to exist. Unplanned deployent of FPs in the network ay result in no significant advantage for saving energy. Table II suarizes the values of the paraeters that we used in our analysis. TBLE I: SUMMRY OF NUMBER OF FP ND DIFFERENT FCTORS USED IN NLYSIS Station Nae No. of FP Factor (f) Office 3 (=f ) Super shop 2 0.8 (=f 2 ) Counity center 3 0.8 (=f 2 ) Residence 0.7 (=f 3 ) Residence 2 0.7 (=f 3 ) The probability of a acrocellular BS to be turned off increases rapidly with the increase of FPs installed in the considered network. s FPs are very low power consuing devices, installent of a large nuber of FP in the network will still result in less energy consuption copared to that by the BS alone. Fig. 5 shows the coparison of the probability for 3 different nuber of users existing in the network. s the nuber of user increases, the probability of the BS to be turned off decreases. So, the BS needs to be turned on when the nuber of user is ore in the network as there will be a probability of at least one user to stay outside the fetocell network region. Fig. 6 shows the coparison between 3 different cases for BS turn off probability. The energy saving possibility is enhanced if ore FPs are used. If only 5 FPs are used in the network, the probability of the BS to be turned off is nearly 5% when 5 users exist in the network. But if 25 FPs are used for the sae network, the probability of the BS to be turned off is ore than 50% for the sae nuber of user. Residence 3 0.7 (=f 3 ) Residence 4 0.7 (=f 3 ) Bank (=f ) Hotel 0.7 (=f 3 ) Shop 0.7 (=f 3 ) Shop 2 0 0.3 (=f 3 ) Reaining free space 0 0.0 (=f 3 ) TBLE II: SUMMRY OF THE PRMETER VLUES USED IN NLYSIS Paraeter Carrier frequency Transit signal power by the acrocellular BS Transitted signal power by a FP Power required for operation of a acrocellular BS Power required for operation of a acrocellular BS Height of a acrocellular BS Height of a FP Distance of the MS fro reference FP Lpen Lsh Range of a acrocellular BS Value 800 [MHz].5 [kw] 5 [W] 2 [kw] 8 [W] 00 [] 2 [] 7 [] 20 [db] 8 [db] 500 [] Figure 5. Probability of acrocellular BS to be turned off with increasing FP for different nuber of users. Range of a FP verage iportance factor, f p 0.7 Total received noise 5 [] 7 * 0-7 [W] Figure 6. Coparison of acrocellular BS turn off probability with nuber of user for different cases.
Figure 8. Coparison of energy consuption between the proposed odel and traditional odel Figure 7. Coparison of SNIR levels between the proposed odel and the traditional odel B. SNIR nalysis The propagation odel for a acrocell case can be expressed as: 0 [ ] ( ) [ ] L = 36.55 + 26.6log f 3.82log h a h + 44.9 6.55log h log d + L db sh 0 c 0 b 0 a ( h ).[ log f 0.7] h (.56 log f 0.8) (3) = 0 c 0 c where fc is the center frequency in MHZ, h b is the height if the acrocellular BS in eter, h is the height of the MS in eter, d is the distance between the acrocellular BS and the MS in kiloeter, Lsh is the shadowing standard deviation. The propagation odel for a fetocell case can be expressed as [3]: 2 f = 20log0 c + log0 + 4 28 [ ] (4) L f N d n db where fc is the center frequency in MHZ of the fetocell, n is the nuber of walls between the MS and the FP, and d is the distance between the FP and the MS in eter. The received SNIR level of a fetocell user in a acrocell/fetocell integrated network can be expressed as: SNIR = I S 0 (5) + I N f + Siulation results show that the proposed odel provides better signal-to-noise plus interference level. Fig. 7 deonstrates the coparison of SNIR level between traditional schee and proposed schee. The effect of the acrocellular BS and the nuber of FPs has been considered here. s the BS will not reain turned on for all the tie in the proposed odel, there will be a decrease in the average noise level. Thus, the SNIR level is increased in the proposed odel according to equation (5). C. Coparison of Energy Consuption In our proposed odel, the acrocellular BS will only reain shut down if the existing FP can provide proper service to all the active users. In general, the nuber of active user becoes low in the iddle of the night. So the BS ay reain shut down for ost of the tie during late night. Thus, a large aount of energy can be saved in that period of night. However, the BS ay reain turned off even in different ties of the day if the FPs can provide proper service to the existing users in that period of tie. The FPs consue very low energy copared to the acrocellular BS. So using any FPs can still result in significant energy saving. If we copare the energy consuption of the cellular network for the traditional odel and the proposed odel, there is a large aount of energy saved in the proposed odel as the ain energy consuing eleent for cellular network i.e., the acrocellular BS reains off for a certain period of tie in this odel. Fig. 8 shows the coparison of energy consued between the traditional odel and the proposed odel. The saving of energy can be increased further by installation of FPs in required places where there is a great possibility of carrying obile users. For vehicular environent, the proposed odel in [4] is good enough to serve the obile users. IV. CONCLUSIONS Saving of energy in cellular network counication is an iportant issue as energy prices are increasing day by day. Energy efficiency of the cellular network can be enhanced by the use of feto-access-points. Proper utilization of FPs can result in iproved service quality and reduction in the consued energy of acrocellular base station. If the base station is allowed to turn off even for a sall aount of tie, a significant aount of energy can be saved as ost of the energy required in the cellular counication is consued by the base station. Our proposed odel can be used extensively for energy saving purpose in cellular counications. Fetocells are a novel wireless networking technology that has several advantages including lower cost, better signal quality, and reduced infrastructure cost. In this paper, we
discussed the advantages of fetocellular network for saving energy of cellular counication networks. The various perforance analysis and the coparison results are presented. The siulation results deonstrate that the deployent of fetocells will significantly reduce the consuption of energy of cellular networks. REFERENCES [] X. Fang, S. Misra, X. Guoliang, D. Yang, "Sart Grid The New and Iproved Power Grid: Survey," IEEE Counications Surveys & Tutorials, vol. 4, no. 4, pp. 944,980, Oct. 202. [2] B. Shengrong, F. Richard Yu, Y. Cai,and X. P. Liu, When the Sart Grid Meets Energy-Efficient Counications: Green Wireless Cellular Networks Powered by the Sart Grid, IEEE Transactions on Wireless Counications, vol., no. 8, pp. 304-307, ug. 202. [3] M.Z. Chowdhury, S. H. Chae andy. M. Jang, "Group Handover Manageent in Mobile Fetocellular Network Deployent," International Conference on Ubiquitous and Future Networks (ICUFN), July 202. [4] M.Z. Chowdhury, S. Q. Lee, B. Han Ru, N. Park, Y. M. Jang, "Service Quality Iproveent of Mobile Users in Vehicular Environent by Mobile Fetocell Network Deployent," International Conference on ICT Convergence (ICTC), Sept. 20.