See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/339976572 LoRa-based IoT Network Planning for Advanced Metering Infrastructure in Urban, Suburban and Rural Scenario Conference Paper · December 2019 DOI: 10.1109/ISRITI48646.2019.9034583 CITATIONS READS 15 559 3 authors, including: Muhammad Imam Nashiruddin Telkom University 77 PUBLICATIONS 205 CITATIONS SEE PROFILE All content following this page was uploaded by Muhammad Imam Nashiruddin on 30 June 2020. The user has requested enhancement of the downloaded file. 2019 International Seminar on Research of Information Technology and Intelligent Systems (ISRITI) LoRa-based IoT Network Planning for Advanced Metering Infrastructure in Urban, Suburban and Rural Scenario Yosia Bagariang Center for Regulation & Telecommunication Management School of Electrical Engineering, Telkom University Bandung, Indonesia yosiasibagariang@student.telkomunive rsity.ac.id Muhammad Imam Nashiruddin Center for Regulation & Telecommunication Management School of Electrical Engineering, Telkom University Bandung, Indonesia imamnashir@telkomuniversity.ac.id Abstract—Internet of Things is predicted to be the future of digital business in Indonesia. The trend of IoT implements on a large scale as more sensors and devices are connected. LoRa is one of the promising applications of the Internet of Things with long-range wireless communication in broad areas. LoRa network deployment in Advanced Metering Infrastructure (AMI) is one of the most significant components in smart metering which measure and collect data from smart meters to be analyzed utilities distribution and consumption. In the IoT, the scenario must achieve two main goals: efficient communication like massive connectivity and defense against environmental conditions. In this paper, we propose the LoRa for Advanced Metering Infrastructure in three utilities (electric, gas, and water meter) to implement in urban, suburban, and rural areas in case of Medan and surroundings by calculating capacity and coverage planning to find out the optimum number of LoRa gateways. Use the method of forecasting IoT connected devices then predict the amount of LoRa gateways need with two scenarios. The first scenario requires the capacity planning by volume demand of the customer's supply of the system. Meanwhile, the second scenario needs coverage planning mostly depends on the link budget and geographical. The simulation results in Forsk Atoll 3.3.2 indicate that the whole determined areas can serve by the number of gateways that have to obtain within acceptable the mean of best signal levels is -76,28 dBm, 75,7 dBm, -81,67 dBm for the urban, suburban and rural scenario, respectively. Keywords— Internet of Things, LPWAN, LoRa, Advanced Metering Infrastructure I. INTRODUCTION Several communication technologies and standards have emerged to fulfill the communication requirements of the Internet of Things (IoT) propose for Low-Power Wide Area Network (LPWANs). Examples of such LPWAN are LoRa (Long Range). LoRa is one such LPWA protocol and to be targets deployments where end-devices have limited energy, where end-devices have not to need to transmit more than a few bytes at the time [1]. However, the management of LoRa networks for IoT faces several challenges. Including a large and highly- varying number of devices, many wireless scenarios characterized by demanding environmental factors, with many buildings, especially high-rise buildings, interference due to other co-located networks operating on the same license-free frequency bands, It called industrial scientific and medical (ISM) radio band. LoRa communication technology on IoT supports connectivity operating sensors to connect people and businesses using data rates ranging from 0.3 Kbps to 50 Kbps for each channel to save battery life. This new type of 978-1-7281-4518-1/19/Print ©2019 IEEE Nachwan Mufti Adriansyah Center for Regulation & Telecommunication Management School of Electrical Engineering, Telkom University Bandung, Indonesia nachwan@telkomuniversity.ac.id transceiver in Advanced Metering Infrastructure (AMI) allows utility ( electric, water, and gas meter) to remotely and timely obtain the customer consumption, and fault detection in real-time without unnecessary waste[2]. Data will be processed by MDMS (Meter Data Management System) before making it available for billing and analysis. MDMS carries out long term data storage and management for large amounts of data delivered by a smart meter. Fig. 1. The Components of Advanced Meter Infrastructure[3] The application of AMI is expected to increase the efficiency of monitoring and improving leak detection to reduce the losses due to the leakage. In 2018, national power utility in Indonesia (PLN) recorded the total value of losses due to electricity theft both in residential areas and industry every year around Rp. 10 trillion. At national water utility in Indonesia (PDAM) reports that the level of water loss or NonRevenue Water (NRW) is 32.80%. This non-billed water not only causes PDAM financial losses but also causes water volume and pressure to flow into house connections to be reduced[4]. While in 2018, recorded carbon emissions of gas leakage are 20.259,01m3[5]. Therefore, AMI in LoRa is expected to be a solution to prevent and loss reduction and information to enable automated. It is in line with the research by [6] that the majority of the Indonesian market is interested in IoT applications that will provide benefits in terms of their work affairs. This paper aims to provide the number of LoRa gateways by calculating the capacity and coverage planning in three scenarios of areas (urban, suburban, and rural) to handle and cover communication between utilities and costumers 188 II. LORA AND ADVANCED METERING INFRASTRUCTURE A. LoRa Physical Layer The Long Range (LoRa) network consists of two components, namely, LoRa and Long-Range Wide Area Network (LoRaWAN), based on different layers of the protocol stack. LoRa is a wireless communication proprietary technology of the LPWA to provide the low power, low rate, with long-range communication that uses the Chirp Spread Spectrum (CSS) radio modulation technique developed by Semtech Corporation. LoRaWAN is introducing by the LoRa Alliance [8]. Facilitates the interconnection among the end nodes and the base stations that collect and forward messages. It uses the unlicensed ISM band (The Industrial, Scientific, and Medical radio band), which varies following government regulations [9]. Indonesia already has a regulation of LPWA technical equipment; its frequency spectrum allocation is 920923 MHz[10]. Each LoRa modulation has several parameters in the physical layer to optimization: a) Bandwidth (BW): The range frequency of chirp is used to modulate the LoRa data signal ( to ). Bandwidth also represents the chip rate of the LoRa signal modulation. It has three programmable bandwidth settings 500 kHz, 250 kHz, and 125 kHz. b) Spreading Factor (SF): the ratio between the data symbol rate and chirp rate or many chips used to represent one symbol, with an exponential factor is 2. When SF is higher at a lower data rate, and coverage is higher[8] — usually taking SF values in 7, 8, 9, 10, 11, 12. c) Code Rate (CR): the forward error correction (FEC) rate, and it affects the airtime of packet transmissions[8]. This implementation is carried out by encoding 4 bits of data with redundancy to 5 bits (4/5), 6 bits (4/6), 7 bits (4/7), or 8 (4/8) bits. Using this redundancy makes LoRa signals more resistant to interference, CR values increase if there is much interference in the channel, but the transmission with n: 1 – 4. time is longer. The CR is equal to B. Advanced Metering Infrastructure Advanced Metering Infrastructure (AMI) is the basic building block for the development of Smart Grid in the distribution system to calculate, gather, store, analyze, and process the data regarding the consumption of customers. The AMI system provides information on the use of granular energy for utilities and customers to remotely read usage reading in real-time. Advanced metering infrastructure has functionalities such as power quality monitoring and management, improving energy efficiency, adaptive power pricing, etc. [11] AMI system has three major components: (1) smart meters (and associated communication modules), (2) a communication network, and (3) AMI back-office information technology (IT) systems to manage the two-way communications enabled by AMI. An overview of AMI system components is shown in Fig. 2[12]. Fig. 2. AMI Overview[12] Designing an AMI system is essential to have a common platform for monitoring as well as utilization of critical features in the AMI system. They are smart energy meter, Data Concentrator Unit (DCU), and Smart Grid Control Centre (SGCC)[13]. The main feature of Machine Type Communication can handle massive connectivity. Smart meter device on water, electricity, and gas does not need handover and sends different small data of each service that concentrate on the uplink side. Smart Meter is a digital electronic device to collects information. The most challenging applications of the smart distribution grid because of the massive number of endsnodes. Energy management to improve efficiency that is predicted to be implemented massively in the future. The smart meter will replace exiting metering and offer scheduled or on-demand remote reading. It can be used both prepaid and postpaid. Fig. 3 provides a list of features and the indicators need in the smart meters regarding electric energy, communication, data real-time alarm also cost, and maintenance[15]. Fig. 3. Features of Smart Meters [15] III. METHODOLOGY AND SCENARIO OF LORA To determine the number of gateways needs to be analyzed: number of customers, LoRa capacity, and geographic area A. Research Area and Density of Connected Devices The area of this paper divide into three scenarios based on the number of population (customers) per kilometers. The high density, which represents by urban areas will apply in Medan city is 256 Km2. The low density which represents by suburban areas at Pematang Siantar city is 55,66 Km2. The last is rural as the land area at Padang Sidempuan city is 114,66 Km2. Based on these three areas, the AMI network for communication will design with the same services (electric, water, and gas providing). Based on data from the Central 189 Bureau of Statistics (Badan Pusat Statistik) stating the number of devices in 2018, this data can forecast IoT connected devices for ten years. Table 1 shows the number of customers in 2018 and the prediction number in 2029: TABLE I. NUMBER OF CUSTOMERS [16] Year Utility Medan P. Siantar P. Sidempuan 2018 Electricity 702151 639369 306607 Water 487251 66604 13438 Gas 21498 23694 17674 Electricity 1432977 1304849 625736 Water 694915 560512 19166 Gas 53927 59435 44335 2181819 1924796 689237 2029 Total in 2029 B. Network Dimensioning Technically, dimensioning of LoRa objective is to determine the total required capacity to carry the aggregated traffic with data communication needs. The capacity of the planned LoRa network must comply with the requirements of traffic to handle, demand ≤ supply. To manage the network system can estimate from Time on-Air (ToA) or data rate distribution and receiver sensitivity. For calculation of the ToA, it defines symbol duration, Tsym. The formula is below: Tsym = (1) The element of the LoRa Packet, as shown in the following image. C. Network Planning An estimation of the resources needed to provide service in the deployment area with the given system parameters (Txpower, Receive Sensitivity, Tx gain, Rx gain) mostly dependent on geographical and environmental factors. It isn’t about coverage but also loads. Design network planning in radio communication systems end to end needs a link budget. The point of calculating the link budget is to get the estimated maximum path loss from Tx to Rx, namely is Maximum Allowable Path Loss (MAPL). Okumura-Hata model is the simplest propagation and gives the best accuracy in path loss (PL) estimation. This model used for determining the path loss in the frequency range of 150MHz to 1920MHz, and this is the LPWA band. Most of the people used it to predict path loss in an urban, suburban and rural area, because of its performance in accuracy and simplicity[18] a) Medan City: Scenario of an urban area that has a high level of population density and human activity compared to the other areas. The path loss, PLoss equation for the Hata model (in dB) for urban areas is given below [2]: PL = 69.55 + 26.16log (f) – 13.82log hb – a(hm) + (44.9 – 6.55loghb)log10 d (6) Where a(hm) = 3.2log(11.75hm)2 – 4.97, for f ≥ 400MHz (7) b) P. Siantar: An area that has a lower population density than an urban area. The path attenuation equation is[19] : PL = PL (urban) – {log (f /28)}2 -5.4 (8) 2 Where a(hm) = {1.1log(f)} – 0.7) hm – (1.56log(f) – 0.8)4.97, for small and medium (9) c) PL = PL (urban) – 4.78{log(f)}2 + 18.33log(f) – 40.98 (10) Fig. 4. LoRa Modem Packet formatting[17] configurations is a sequence of preamble Where npreamble is the number of programmed preamble symbols., it is given by: Tpreamble = (npreamble + 4.25) Tsym PayloadSymbNb= 8 + max (ceil ( ( (2) ) P. Sidempuan: This area is still close to the city and is usually called rural because there are few houses or other buildings and there are mountains and farms and below for rural area: )(CR+4),0) (3) With the following dependencies[17]: • • • • • PL is the number of payload bytes SF is Spreading Factor H = 0; the header is enabled, H = 1; no header. CR is the coding rate from 1 to 4 DE = 1; the low data rate optimization is enabled, DE = 0; for disabled. The payload duration is the symbol period multiplied by the number of payload symbols. The ToA, (packet duration), is sketchily then the sum of the preamble and payload duration: Tpayload = payloadSymbNb Tsym (4) Tpacket = Tpreamble + Tpayload (5) 190 Where a(hm) suburban = a(hm) rural Note: • • • • • F hb hm a(hm) d : transmit frequency (MHz) : height gateway (m) : height of device (m) : correction factor : cell radius (Km) There is also a widely used logarithmic calculation formula for MAPL MAPL = Tx power + Rx gain + Tx gain + Rx-sensitivity (11) To find the MAPL the parameters are: TABLE II. CONFIGURATION PARAMETERS[20] Rx – Sensitivity (dBm) SF 7 SF 8 SF 9 SF 10 SF 11 SF 12 123 126 129 132 134.5 137 Tx power (dBm) Tx gain (dBi) Rx Gain (dBi) 14 5 2 IV. RESULTS AND ANALYSIS Gateways in P. Siantar A. Capacity Result a) Required Capacity: Calculation of demand based on traffic characteristic of LoRa and technical requirement of each utility in providing services. Table III shows the total required packet per day after the calculation feature (scheduled meter reading, firmware update, on-demand mater reading) and technical requirement of the smart metering based on the density of devices 40 30 20 10 0 SF 7 CR = 4/5 SF 11 CR = 4/7 SF 12 CR = 4/8 15 TABLE IV. SINGLE LORA GATEWAY CAPACITY Spreading Coding Rate (CR) Factor (SF) CR = 4/5 CR = 4/6 CR = 4/7 CR = 4/8 SF 7 14,285,714. 14,594,594.5 14,594,5 14,594,59 32 6 94.56 4.56 SF 8 7,803,468.2 7,803,468.24 7,988,16 7,988,165 4 5.68 .68 SF 9 4,272,546.5 4,192,546.56 4,299,36 4,299,363 6 3.04 ,04 SF 10 2,205,882.3 2,265,100.64 2,265,10 2,265,100 2 0.64 .64 SF 11 1,074,840.8 1,074,840.80 1,102,94 1,102,941 0 1.20 .20 SF 12 566,275.20 581,896.56 581,896. 581,896.5 56 6 10 5 0 SF 7 SF 8 CR = 4/5 SF 9 CR = 4/6 SF 10 SF 11 CR = 4/7 SF 12 CR = 4/8 Fig. 7. The initial number of gateways in P. Sidempuan B. Coverage Results According to (11) is used to a propagation prediction model based on SF (7-12), Tx power (14dBm), Tx gain (5dBi and Rx gain (2dBi) can be used to determine the path loss from the gateway to the device which is the required path loss for three scenarios. Table V shows the Path loss value SF Path Loss TABLE V. PATH LOSS VALUE 7 8 9 10 11 144 147 150 153 155.5 12 158 To find the total number of gateways need are: Single√ Prediction of Total Gateway: A number of the gateway can obtain by comparing the Required Capacity (demand) to Gateway Capacity (supply). Axis x is Code Rate for each SF, and axis y is the number of LoRa gateways 40 CR = 4/6 SF 10 Gateways in P. Sidempuan b) Gateway Capacity: Numerical evaluation of the gateway’s capability to manage the traffic. The number is the maximum level of agent load. Configure bandwidth (BW), Spreading Factor (SF), the last Code Rate (CR). Calculation of the number of packets per day based on the equation. The packets of the single gateway in eight channels [21]. Table IV shows the individual LoRa gateway capacity Gateways in Medan SF 9 Fig. 6. The initial number of gateways in P. Siantar TABLE III. TOTAL REQUIRED PACKET FOR MEDAN, P. SIANTAR, AND P. SIDEMPUAN AREA Area Number of a required packet per day Medan 20,563,668 P. Siantar 18,141,230 P. Sidempuan 6,496159 c) SF 8 (12) gateway coverage (hexagonal area) = Total gateways = area-wide / single- gateway coverage (13) a) Urban scenario: According to equation (6) is used to determine cell radius (d) of the urban gateway in Medan and use path loss value base on table VI for every SF, and the result is used to find total gateways by using equation (12) TABLE VI. RADIUS CELL AND TOTAL GATEWAYS IN MEDAN SF 7 8 9 10 11 12 Radius (Km) 4.172 5.022 6.047 7.282 8.501 9.923 Total 24 20 17 14 12 10 gateways 30 20 10 0 SF 7 SF 8 CR = 4/5 SF 9 CR = 4/6 SF 10 CR = 4/7 SF 11 SF 12 CR = 4/8 Fig. 5. The initial number of gateways in Medan b) Urban scenario: Okumura-Hata suburban according to equation (8) to find the distance. Table VII shows the radius cell and total gateways in P. Siantar 191 TABLE VII. RADIUS CELL AND TOTAL GATEWAYS IN P. SIANTAR SF 7 8 9 10 11 12 Radius (Km) 4.832 5.842 7.064 8.541 10.01 11.72 Total gateways 5 4 4 3 3 2 c) b) Suburban: P. Siantar PEMATANG SIANTAR Coverage Siantar Siantar Capacity CR2 Siantar Capacity CR4 Rural scenario: Same as the previous scenario, by using equation (10), the distance can be generated for a rural area. Siantar Capacity CR1 Siantar Capacity CR3 40 TABLE VIII. RADIUS CELL AND TOTAL GATEWAYS IN P. SIDEMPUAN SF 7 8 9 10 11 12 Radius (Km) 18.01 21.84 26.50 32.15 37.77 44.37 Total gateways 3 3 2 2 2 1 20 0 SF7 C. Network Planning Analysis The total gateway is known based on capacity and coverage. These results will be compared to determine the optimal gateway would expect. a) Urban: Medan MEDAN AREA 40 Medan Coverage Medan Capacity CR1 Medan Capacity CR2 Medan Capacity CR3 SF8 SF9 SF10 SF11 SF12 Fig. 9. Number of gateways in P. Siantar The results can conclude that in the suburban scenario has two intersection points, SF 8 and SF 9. Table X shows the comparison. The final gateway version is four gateways. But the area has unstable ground contours, adds one gateway to cover the blank spot. The mean of best signal level is -75,7 dBm Fig. 10 shows the coverage prediction of LoRa in P. Siantar TABLE X. COMPARISION OF TOTAL GATEWAYS SF Capacity Planning Coverage Planning 8 3 4 9 5 4 Medan Capacity CR6 20 0 SF7 SF8 SF9 SF10 SF11 SF12 Fig. 8. Number of gateways in Medan Fig. 8 shows the intersection point of SF 10 in the urban scenario. The final gateway version is 14 gateways with a mean of best signal level is -76,28 dBm. Fig. 9 shows the coverage prediction of LoRa in Medan TABLE IX. COMPARISION OF TOTAL GATEWAYS SF Capacity Planning Coverage planning 10 10 14 Fig. 10. Coverage prediction in P. Siantar c) Rural: P. Sidempuan P. SIDEMPUAN Coverage P.Sidempuan P. Sidempuan CR1 P. Sidempuan CR2 P. Sidempuan CR3 P. Sidempuan CR4 15 10 5 0 SF7 Fig. 9. Coverage prediction in Medan SF8 SF9 SF10 SF11 SF12 Fig. 10. Number of gateways in P. Sidempuan The intersection point of SF 9 in Fig. 10 has the same gateways of capacity and coverage planning. The final gateway version is two gateways with a mean of best signal level is -81,67 dBm. Fig. 11 shows the coverage prediction of LoRa in P. Sidempuan 192 Advanced Metering Infrastructure, 2016. TABLE XI. COMPARISION OF GATEWAYS Capacity Planning Coverage Planning 2 2 [4] K. Pekerjaan, U. Dan, and P. Rakyat, Buku Kinerja PDAM 2017, 1st ed. Jakarta: Perusahaan Daerah Air Minum, 2017. [5] L. Keberlanjutan, “Doors of sustainability,” 2018. [6] F. Vannieuwenborg and S. Verbrugge, “Choosing IoTconnectivity ? A guiding methodology based on functional characteristics and economic considerations,” no. March, pp. 1– 16, 2018. [7] G. Wibisono, G. P. S, E. Eng, and U. Indonesia, “Development of Advanced Metering Infrastructure Based on LoRa WAN in PLN Bali Toward Bali Eco Smart Grid,” pp. 4–7, 2016. [8] M. Slabicki, G. Premsankar, and M. Di Francesco, “Adaptive configuration of lora networks for dense IoT deployments,” IEEE/IFIP Netw. Oper. Manag. Symp. Cogn. Manag. a Cyber World, NOMS 2018, pp. 1–9, 2018. [9] Y. Cheng, H. Saputra, L. M. Goh, and Y. Wu, “Secure smart metering based on LoRa technology,” 2018 IEEE 4th Int. Conf. Identity, Secur. Behav. Anal. ISBA, 2018, vol. 2018-Janua, pp. 1– 8, 2018. [10] Kementrian Komunikasi Dan Informatika Republik Indonesia, “PERDIRJEN SDPPI NO 3 Tahun 2019 Tentang Persyaratan Teknis Alat dan/Atau Perangkat Telekomunikasi LPWA.” Jakarta, 2019. [11] A. Ghasempour and T. K. Moon, “Optimizing the Number of Collectors in Machine-to-Machine Advanced Metering Infrastructure Architecture for Internet of Things-Based Smart Grid,” IEEE Green Technol. Conf., vol. 2016-April, pp. 51–55, 2016. [12] Consolidated Edison, Advanced Metering Infrastructure Business Plan. 2015. [13] I. S. Jha, S. Sen, and V. Agarwal, “Advanced metering infrastructure analytics - A Case Study,” 2014 18th Natl. Power Syst. Conf. NPSC, 2014, pp. 3–8, 2015. [14] A. Haidine, “Performance Evaluation of Low-Power Wide Area based on LoRa Technology for Smart Metering,” 2018 6th Int. Conf. Wirel. Networks Mob. Commun., pp. 1–6, 2018. [15] J. Lloret, J. Tomas, A. Canovas, and L. Parra, “An Integrated IoT Architecture for Smart Metering,” no. December, pp. 50–57, 2016. [16] BPS- Statistics of Sumatera Utara Province, Sumatera Utara Province in Figure 2019. sumatera Utara: ©BPS Provinsi Sumatera Utara/BPS-Statistics of Sumatera Utara Province, 2019. ACKNOWLEDGMENT [17] The author expressed appreciation to the Telkom University on their support in this research. L. Modem and D. Guide, “Design er ’ s Guide TCo Table of Contents Table of Figures,” no. July, pp. 1–9, 2013. [18] F. Eng, I. Islamic, J. Gombak, and A. S. Architecture, “LoRa LPWAN Propagation Channel Modelling in IIUM Campus,” pp. 14–19, 2018. SF 9 Fig. 11. Coverage prediction in P. Sidempuan V. CONCLUSION This study contributes to the internet of things research in the following ways. First, the integrated IoT architecture based on LoRa for Advanced Metering Infrastructure for electricity, water, and gas can apply in urban, suburban, and rural areas. Second, following the scenarios and calculations based on capacity planning and coverage planning show that to determine the number of gateways needed is the number of customers, the LoRa capacity system, and the geographical area. Third, from the analysis of capacity and coverage, it found that the gateway requirements from each region were fourteen gateways for Medan (Urban Scenario), five gateways for Pematang Siantar (Sub Urban Scenario), and two gateways for Padang Sidempuan (Rural Scenario). Fourth, from the simulations carried out, it can also conclude that for the Medan City that is specified can be served at an acceptable level with a value > -132 dBm, Pematang Siantar with an adequate level > -129 dBm, while for the Padang Sidempuan City it is > -129 dBm. The best signal levels on average are -76,28 dBm, -75,7 dBm, and 81,67 dBm for the Medan, Pematang Siantar, and Padang Sidempuan. The number of gateways that calculated in this paper needs to adjust with the actual environment to find an accurate total gateway to serve customers optimally. For further research, it is necessary to use BTS existing to design LoRa gateways laying to decrease investment costs. REFERENCES [1] A. Augustin, J. Yi, T. Clausen, and W. M. Townsley, “A study of Lora: Long range & low power networks for the internet of things,” Sensors (Switzerland), vol. 16, no. 9, pp. 1–18, 2016. [19] V. S. Anusha, G. K. Nithya, and S. N. Rao, “A Comprehensive Survey of ElectroMagnetic Propagation Models,” pp. 1457–1462, 2017. [2] E. Harinda, S. Hosseinzadeh, H. Larijani, and R. M. Gibson, “Comparative Performance Analysis of Empirical Propagation Models for LoRaWAN 868MHz in an Urban Scenario,” 2019 IEEE 5th World Forum Internet Things, pp. 154–159, 2019. [20] T. T. Conference, “LoRaWAN Radio Planning Workshop instructions,” 2019. [21] K. Mikhaylov and J. Petäjäjärvi, “Analysis of Capacity and Scalability of the LoRa Low Power Wide Area Network Technology,” Univ. Oulu, Cent. Wirel. Commun. Finl., pp. 119– 124, 2016. [3] R. K. Pillai, B. Rupendra, and T. Hem, “AMI Rollout Strategy and Cost-Benefit Analysis for India,” ISGF White Pap., vol. 1, no. View publication stats 193