The Impact of Dynamic Scaling on Energy Consumption at Node Level in Wireless Sensor Networks

Transcription

1 International Journal of Applied Engineering Research ISSN Volume 13, Number 1 (218) pp The Impact of Dynamic Scaling on Energy Consumption at Node Level in Wireless Sensor Networks Rajan Sharma 1 1 Research Scholar, Department of Electronics and Communication Engineering I.K. Gujral Punjab Technical University, Jalandhar, Punjab, India. Orcid: Balwinder Singh Sohi 2 2 Department of Electronics and Communication Engineering CGC Group of Colleges, Mohali, Punjab, India Abstract Energy efficiency is one of the critical issues in Wireless Sensor Networks (WSNs) due to limited capacity of batteries of its sensor nodes. This paper investigates the impact of dynamic scaling of parameters such as modulation scheme, voltage or frequency which reduces energy consumption of sensor node. The performance is evaluated mathematically for scaling of voltage and frequency dynamically. This approach is used to dynamically adjust frequency and voltage of processing unit according to actual system load to save computational power without degrading its performance. We focused on the fact that energy consumed by wireless sensor node is the product of power consumed and the time for any operating mode, the reduction in the value of either parameter such as voltage, frequency and varying modulation scheme may result in saving of huge energy. The impact of various modulation schemes at different frequency bands and operating modes on energy consumption of sensor node is evaluated using QUALNET 6.1 Simulator. Keywords: Wireless Sensor Networks, Sensor Node, Dynamic Voltage Scaling (DVS), Dynamic Frequency Scaling (DFS), Dynamic Modulation Scaling (DMS), Energy Consumption, QUALNET 6.1. INTRODUCTION There are many parameters such as modulation scheme, voltage or frequency which reduces power consumption of sensor node by reducing the communication or computation time of the radio-transceiver and processor respectively. We can save more energy by dynamically adopting the modulation level according to traffic load, known as modulation scaling (DMS) [1, 2]. The same is applicable for voltage or frequency which is known as voltage scaling and frequency scaling respectively [3, 4]. The major portion of the total energy consumption is spent in communication as nodes which have to communicate with each other and with the outside world are enabled by radio. It has a radio of limited area communication which works on the unlicensed radio frequency bands as defined by IEEE standards i.e. 868 (for Europe & Japan), 915 (for USA) and 24 (used worldwide) also known as ISM radio band. The number of channels in 868 band, 915 band and 24 band are 1, 1 and 16 respectively. Similarly, the data rate for these three bands are 2 kbps, 4kbps and 25 kbps respectively [5, 6, 7]. While communicating in long distance, the sensors should avoid direct communication with the sink node; instead it should communicate using multi hop communication. Direct communication means requirement of high transmission power in order to achieve reliable transmission which leads to more energy usage. Multiple paths should be chosen at different times as the single optimized path will lose its energy if the same path is chosen repeatedly for all the transmissions. The paths should be chosen in a round robin manner. In this way the energy is saved by using multi hop communication. The features of Power consumption in the radio are affected by various factors which contain operational duty cycle, data rate, transmit power and various types of modulation schemes. Transmit, Receive, idle and are the modes of operation in any transceiver similar to microcontrollers. Every time it has been observed that the maximum energy consumed by the transceiver is in the idle mode so it is advised not to put the transceiver in the idle mode, instead of it the transceiver should be kept in sleep mode. Another influencing factor is that, the transient activity state, which is known as the ON and OFF state in the radio system consumes a substantial quantity of power as the radio's operating mode changes. This causes dissipation of energy and is considered as wastage of energy. The rest of this paper is organized as follows. Section two introduces brief summary related to energy consumption in a sensor node. In section three, we briefly discuss the related work focused on various scaling techniques and energy consumption. Section four is focused on Basics of Dynamic Voltage - Frequency Scaling (DVFS) Technique. In section five, the simulation results and analysis is discussed based on scaling of hardware parameters and modulation. Similarly, energy consumption is compared based on operational mode, frequency band, operational time and modulation scheme. Finally, we present our conclusion and prospects in section six. ENERGY CONSUMPTION IN A SENSOR NODE WSNs are likely to run for a long time without human intervention and with limited power in the sensor nodes, energy consumption is critical issue in these types of networks. The WSN applications [8, 9] are very broad, for example, military, 175

2 International Journal of Applied Engineering Research ISSN Volume 13, Number 1 (218) pp health info, environmental monitoring and smart home. The radio transceiver in a sensor node is the primary consumer of energy compared to other node components such as the microprocessor or the sensing device. The radio can work in four different modes of operation as Table 1 illustrates. The first three modes i.e transmit, receive and listen, require different level of power consumption which depends upon various parameters but turning off the power of the radio port while it is idle listening could save a large amount of energy. Table 1: s of operation in a radio transceiver Comment Transmit A node is transmit mode when its radio is transmitting packets Receive Listen RELATED WORK A node is in receive or reception mode when its radio is receiving packets A node in listen mode when its radio is sensing the channel and waiting for packets A node is in sleep mode when its radio is switched off Energy consumption for computation and communication in sensor nodes has been important concern as it firmly decide the lifetime of node and networks. Reduction in energy consumption based on hardware parameters such as voltage, frequency as well as modulation technique has been a domain of research for long. Maryam Bandari et al. [1] presented the combined method that reduces the energy consumption in wireless systems under probabilistic workloads, but they have not focused on the impact of parameters like voltage or frequency for energy consumption using DVS approach. Similarly, the impact on the energy consumption using DMS approach. Saeeda Usman et al. [11] reported a comparative study of DVFS techniques on the basis of power reduction, energy saving and performance. They have not explained, how the variation in parameters such as voltage and frequency can save time and energy consumption. Ulf Kulau et al. [12] focused on hybrid DVS and DPM technique for power optimization in sensor networks. They named this hybrid approach as mod DVS which is the improved version of classical DVS for improving the energy efficiency of processor, transceiver, sensor and memories. This increases the lifetime of sensor as well as the network, but they have not focused on the impact of various modulation schemes on the energy optimization. Bahareh Gholamzadeh et al. [13] has discussed various sources of energy waste in WSNs, characteristics of node hardware such as processing unit, communication device (transceiver), sensing device and power supply device. They also focused on various design approaches to minimize power consumption and for prolonging the lifetime of node and network. Sharma et al. [14-16] focused on sources of energy waste, energy efficiency and lifetime of WSNs. Our work differs from the above mentioned contributions as we focused on the impact on energy consumption with variation of hardware parameters such as voltage and clock frequency, data rate, communication time, frequency band and modulation scheme on the sensing node in WSNs. BASICS OF DYNAMIC VOLTAGE FREQUENCY SCALING (DVFS) TECHNIQUE With the advancement in chip designing technology, the controllers speed increased up to GHz, so the power dissipation has also increased accordingly. The power of sensor nodes can be reduced by designing ultra-low power CMOS chips as the first step in this direction with feature of Dynamic Voltage scaling (DVS) [17] or Dynamic Frequency Scaling (DFS) or combination of both which is Dynamic Voltage and Frequency Scaling (DVFS) [18]. There are two approaches to implement DVS or DFS or DVFS. In first approach, the processing unit can be switched to its full operational capacity mode to compute the task at the fastest speed and then return back to the sleep mode as soon as possible. The speed of the computational task can be increased either by increasing voltage or frequency within the specified limits as defined by data sheet in the situation of high load. In alternative approach, the computational task speed can be decreased, whenever there is less load or no computational task or activity [19, 2, 21, 22]. In other words, compute the assigned task only at the required minimum speed to finish the same before deadline. Processor Table 2: Effect of Dynamic Voltage Frequency Scaling (DVFS) on energy consumption Clock Frequency Range (In ) Voltage Range (In V) Different Scenarios Operating Clock Frequency (F) (In ) Operating Voltage (In V) Power Consumption Reduction by Factor(PCC1/ PCC 2) Speed Reduction by Factor (FC1/FC2) Required Energy Reduction per instruction = Speed Reduced Factor/ Power Consumption Reduction Factor (in %) ATmega1284P Case Case TI MSP Case Case

3 International Journal of Applied Engineering Research ISSN Volume 13, Number 1 (218) pp We can trade-off between operating voltage and operating frequency as mentioned in above approaches, the power consumption can be reduced to some extent in CMOS chips. It is a way of controlling and saving the energy usage in a sensor node. It is the most efficient approach, which keeps track of the system activities, input arrival and contributes significantly to lifetime improvement by managing system components activities. The Dynamic Power Management (DPM) technique is implemented at the operating system of the sensor node. Therefore, power saving using dynamic power management and proper scheduling of input events is essential for extending the lifetime of embedded sensor nodes [23]. The power dissipation of controller based on CMOS technology is the integration of dynamic power and static power [24, 25]. There are mainly two types of power consumption i.e. dynamic power P Dynamic and static power P Static. Mathematically, P CMOS = P Dynamic + P Static (1) The dynamic power is comparatively major portion of the CMOS power dissipation. It can be expressed as [26, 27]: P Dynamic= C L. v 2. f c (2) Where: C L= Parasitic capacitance and its value depends upon the manufacturing process quality. v= Operating supply voltage f c = Operating frequency of controller INVESTIGATION AND ANALYSIS Power Consumption Reduction Factor and Speed Reduction can be calculated as mentioned below: Power Consumption Reduction Factor = Power Consumption in Case 1(PCC1) (3) Power Consumption in Case 2 (PCC2) Speed Reduction Factor = Operating Frequency of Controller in Case 1(FC1) (4) Operating Frequency of Controller in Case 2 (FC2) Required Energy Reduction per instruction = Speed Reduced Factor Power Consumption Reduction Factor (5) operating voltage from 3.3 V to 1.8 V i.e. again almost half, the power consumption is reduced by the factor of 6.72 but the speed is reduced by the factor of 2 only. The required energy per instruction is reduced by %. Similarly, it has been observed for TI MSP43 that by decreasing the operating frequency from 7 to 4.6 and decreasing the operating voltage from 3.3 V to 2 V, the power consumption is reduced by the factor of 4.14 but the speed is reduced by the factor of 1.52 only. The required energy per instruction is reduced by approximately 36.71% A. Effect of Operational Frequency Band on Energy Consumption in Sensor Node There is relation between operational frequency band, data rate, transmission time and processing time. As the node switches to higher frequency band, data rate increases accordingly as per design specifications, further increase in data rate decreases the transmission time of frame as shown in table 3 and Fig. 1. If transmission time of frame decreases, then it also decreases the idle time of the processing unit. This will decrease the energy consumption of that particular sensing node as well as prolonging battery life span and accumulation of this energy saving at each node will increase the lifetime of network. We know that Energy Consumption of sensing node may be expressed as: Time to send or receive per bit (T) = 1/ Data rate Where the unit of data rate and time are Kbps and ms respectively. Table 3: Relation between frequency band, data rate and communication time. Frequency Band (in ) Maximum Data Rate (in kbps ) Time to send or Receive per bit (T) (in ms) kbps kbps kbps.4 The Table 3 and Fig. 1 shows that the time required to send or receive per bit in 868 frequency band is highest, comparatively less in 915 band but least in 24 band ,2 kbps ,4 kbps ,25 kbps Table 2 reports the effect of Dynamic Voltage Frequency Scaling (DVFS) on energy consumption, it has been observed for ATmega1284P that by decreasing the operating frequency from 4 to 2 i.e. reduced to half and decreasing the Time to send or Receive per bit (T) (in ms) Figure 1: Communication time w.r.t frequency band and data rate. 177

4 International Journal of Applied Engineering Research ISSN Volume 13, Number 1 (218) pp Figure 2: Network scenario B. Effect of Operational Frequency Band and Modulation Scheme on Energy Consumption in Sensor Node Simulation l The proposed network architecture in Fig. 2 is cluster of nine Sensor nodes in the star topology and each node is 5 m apart from each other in the specific target area of 1m * 1m in the Terrain Size of 1m * 1 m. In the proposed simulation environment, various parameters are considered at network and node levels are represented in Table 4. The main goal of this simulation setup is to investigate the minimum energy consumption modulation scheme at various frequency bands and operational modes. The various simulation parameters used in the research are: operating modes, modulation scheme, frequency band, and operational time duration as well as battery energy consumption. The tabular results that compare energy consumption for various modulation schemes at different frequency bands for nodes operating in transmit, receive and idle mode are represented in the Table 5. Table 4: Simulation parameters of network scenario. Parameter Value Simulator Qual Net 6.1 Terrain Size 1 * 1 Sq M. No. of Nodes 9 MAC Protocol IEEE Packet Reception l PHY Radio Type IEEE Energy l Routing Protocol Antenna l Micaz AODV Omni directional Network Protocol IPV 4 Device Type Traffic Type Sensor (FFD, RFD) Constant Bit Rate (CBR) Items to Send 1 Item Size Simulation Time Link Channel Frequency Modulation 7 Bytes 3 Seconds Wireless 868, 915 and 24 ASK, BPSK, O-QPSK 178

5 International Journal of Applied Engineering Research ISSN Volume 13, Number 1 (218) pp Table: 5: Comparison of energy consumed (in mwh) node wise for various modulation schemes at different frequency bands and operating modes. Modulation Scheme Frequency (In ) Energy Consumed (in mwh) Node wise 1 (RFD) 2 (RFD) 3(RFD) 4(RFD) 5(RFD) 6(RFD) 7(RFD) 8(RFD) 9(FFD) O- QPSK Transmit Receive Idle O- QPSK Transmit Receive Idle O- QPSK Transmit Receive Idle BPSK Transmit Receive Idle BPSK Transmit Receive Idle ASK Transmit Receive Idle ASK Transmit Receive Idle

6 International Journal of Applied Engineering Research ISSN Volume 13, Number 1 (218) pp Energy Consumed (in mwh) Figure 3: Comparison of energy consumed (in mwh) node wise for different modulation schemes at 868 in transmit mode The graph in Fig. 3 represents the comparison of energy consumed (in mwh) node wise for different modulation schemes at 868 in transmit mode. The red line clearly shows that the consumption of energy is higher using BPSK modulation scheme, blue line represents that energy consumption is less using O-QPSK compared to BPSK. Similarly, grey line represents that energy consumption is least using ASK as compared to other modulation schemes such as BPSK and O- QPSK. It also shows that the energy consumption is approximately same using O-QPSK or ASK modulation scheme in the case of RFD nodes except FFD node. Energy Consumed (in mwh) Comparison for Energy Consumption in Transmit at 868 Modulation Scheme and Node ID O- QPSK Transmit mode BPSK Transmit mode ASK Transmit mode Comparison for Energy Consumption in Receive at Modulation Scheme and Node ID O- QPSK Receive mode BPSK Receive mode ASK Receive mode Figure 4: Comparison of energy consumed (in mwh) node wise for different modulation schemes at 868 in receive mode. The graph in Fig. 4 shows the comparison of energy consumed (in mwh) node wise for different modulation schemes at 868 in receive mode. The red line clearly shows that the consumption of energy is higher using BPSK modulation Scheme, blue line represents that energy consumption is less using O-QPSK compared to BPSK. Similarly, grey line represents that energy consumption is least using ASK as compared to BPSK and O-QPSK Energy Consumed (in mwh) Figure 5: Comparison of energy consumed (in mwh) node wise for different modulation schemes at 868 in idle mode. The Fig. 5 represents the consumption of energy (in mwh) node wise for different modulation schemes at 868 in idle mode. The graphs clearly demonstrate that the consumption of energy is on higher side in majority of the nodes using ASK modulation scheme as indicated by grey line. Similarly, red line represents that energy consumption is least in majority of nodes using BPSK as compared to ASK and O-QPSK. Energy Consumed (in mwh) Comparison for Energy Consumption in Idle at Modulation Scheme and Node ID O- QPSK Idle mode BPSK Idle mode ASK Idle mode Comparison for Energy Consumption in Transmit at Modulation Scheme and Node ID O- QPSK Transmit mode BPSK Transmit mode ASK Transmit mode Figure 6: Comparison of energy consumed (in mwh) node wise for different modulation schemes at 915 in transmit mode. The graph in Fig. 6 represents the comparison of energy consumed (in mwh) node wise for different modulation 18

7 International Journal of Applied Engineering Research ISSN Volume 13, Number 1 (218) pp schemes at 915 in transmit mode. The red line clearly shows that the consumption of energy is higher using BPSK modulation Scheme, blue line represents that energy consumption is less using O-QPSK compared to BPSK. Similarly, grey line represents that energy consumption is least using ASK but approximately same as using O- QPSK as compared to BPSK. Energy Consumed (in mwh) Comparison for Energy Consumption in Receive at 915 Modulation Scheme and Node ID O- QPSK Receive mode BPSK Receive mode ASK Receive mode The Fig. 8 represents the consumption of energy (in mwh) node wise for different modulation schemes at 915 in idle mode. The graphs clearly demonstrate that the consumption of energy is higher using ASK modulation scheme as indicated by grey line, blue line represents that energy consumption is less using O-QPSK compared to ASK. Similarly, red line represents that energy consumption is least in sensing nodes using BPSK as compared to ASK and O-QPSK. The Fig. 9 draws a comparative analysis of energy consumption for various operating modes such as transmit, receive and idle mode using O-QPSK modulation schemes at 24 frequency band related to all RFD nodes ( node no. 1 to 8) and specific node: 9 which is Full Functional Device (FFD). The figure represents that energy consumption is on comparatively higher side in receive mode for all RFD s except FFD and energy consumption is minimum for transmission for all the RFD s except FFD. The energy consumed in transmit mode and idle mode is approximately same in majority of RFD s except node 5 and FFD which is comparatively high. Comparison for Energy Consumption in Transmit, Reecive and Idle For O-QPSK Modulation Scheme at 24 Figure 7: Comparison of energy consumed (in mwh) node wise for different modulation schemes at 915 in receive mode. The graph in Fig. 7 shows the comparison of energy consumed (in mwh) node wise for different modulation schemes at 915 in receive mode. The red line which represents BPSK modulation scheme, it clearly shows that the consumption of energy is highest in comparison with other modulation technique, blue line represents that energy consumption is less using O-QPSK compared to BPSK. Similarly, grey line represents that energy consumption is least using ASK as compared to BPSK and O- QPSK. Energy Consumed (in mwh) Communication and Node ID Energy Consumed (in mwh) Comparison for Energy Consumption in Idle at Modulation Scheme and Node ID Transmit Receive Idle Figure 9: Comparison of energy consumed (in mwh) node wise for O-QPSK modulation schemes at 24 in transmit, receive and idle mode. O- QPSK Idle mode BPSK Idle mode ASK Idle mode Figure 8: Comparison of energy consumed (in mwh) node wise for different modulation schemes at 915 in idle mode. 181

8 International Journal of Applied Engineering Research ISSN Volume 13, Number 1 (218) pp Table 6: Total charge consumed by battery (in mahr) for various modulation schemes at different frequencies related to Node: 9 (FFD) Modulation Scheme Frequency Band (In ) Frequency (In ) Node ID : 9 (FFD) Total Charge Consumed by Battery (in mahr) O- QPSK O- QPSK O- QPSK BPSK BPSK ASK ASK Battery Consumption of Node: 9 (FFD) Total Charge Consumed by Battery ( in mahr) O- QPSK O- QPSK O- QPSK BPSK BPSK ASK ASK Modulation Scheme and Frequency 915 Figure 1: Comparison of total charge consumed by battery (in mahr) for various modulation schemes at different frequencies related to node: 9 (FFD) The Table 6 and Fig. 1 draws a comparative analysis of total charge consumed by battery (in mahr) for the same network using various modulation schemes at different frequency band related to specific node: 9 which works as Full Functional Device (FFD). The above figure represents that the battery charge consumption is maximum for BPSK modulation scheme at 868 and battery charge consumption is minimum for ASK modulation scheme at 868 frequency band. The charge consumption by ASK and O-QPSK modulation scheme is approximately same as it ranges from.32 to.38 irrespective of frequency band. The merit of using O-QPSK modulation at 24 frequency band is high data rate i.e. 25 kbps as compared to ASK modulation scheme at 868. The graph clearly shows that the charge consumption in case of BPSK modulation is approximately 1.5 to 2 times higher as compared to ASK and O-QPSK which is not recommended as energy is most precious in sensor networks. C. Effect of Operational Frequency Band and Modulation Scheme on Operational Time in Sensor Node The tabular results that compares percentage of operational time for various modulation schemes at different frequency bands for nodes operating in transmit, receive, idle and sleep mode are represented in the Table

9 International Journal of Applied Engineering Research ISSN Volume 13, Number 1 (218) pp Table 7: Comparison of time (in Percentage) for different modes of operation (node wise) with various modulation schemes and frequency bands. Modulation Scheme Frequency (In ) Operating Time (in %) Node wise 1 (RFD) 2 (RFD) 3(RFD) 4(RFD) 5(RFD) 6(RFD) 7(RFD) 8(RFD) 9(FFD) O- QPSK Transmit Receive Idle O- QPSK Transmit Receive Idle O- QPSK Transmit Receive Idle BPSK Transmit Receive Idle BPSK Transmit Receive Idle ASK Transmit Receive Idle ASK Transmit Receive Idle

10 International Journal of Applied Engineering Research ISSN Volume 13, Number 1 (218) pp Comparison of time duration in Transmit at Comparison of time duration in Receive at Percentage of time Modulation scheme & Node ID Modulation Scheme & Node ID O- QPSK Transmit mode 868 BPSK Transmit mode 868 ASK Transmit mode 868 O- QPSK Receive mode 868 BPSK Receive mode 868 ASK Receive mode 868 Figure 11: Comparison of transmit time (in %) node wise for different modulation schemes at 868 Figure 12: Comparison of receiving time (in %) node wise for different modulation schemes at 868 The graph in Fig. 11 represents the comparison of transmit time (in %) node wise for different modulation schemes at 868 The red line clearly shows that the transmitting time is higher using BPSK modulation scheme, blue line represents that transmitting time is less using O-QPSK compared to BPSK. Similarly, grey line represents that transmitting time is least using ASK as compared to other modulation schemes such as BPSK and O-QPSK. It also shows that the time duration for transmission is approximately same using O-QPSK or ASK modulation scheme in the case of RFD nodes except FFD node. The graph in Fig.12 shows the comparison of receiving time (in %) node wise for different modulation schemes at 868 The red line clearly shows that the time for receiving is highest using BPSK modulation scheme, blue line represents that receiving time is less using O-QPSK compared to BPSK. Similarly, grey line represents that receiving time is least using ASK as compared to BPSK and O-QPSK. 184

11 International Journal of Applied Engineering Research ISSN Volume 13, Number 1 (218) pp Comparison of time duration in Idle at 868 Modulation Scheme & Node ID O- QPSK Idle mode 868 BPSK Idle mode 868 ASK Idle mode 868 The Fig. 14 represents the comparison of sleep time (in %) node wise for different modulation schemes at 868. The graphs clearly demonstrate that the sleep time is on higher side as 5 RFD s nodes are in the range of (93-94) % using O-QPSK modulation scheme as indicated by blue line. The sleep time is in the range of (83-86) % in 6 RFD nodes using BPSK modulation scheme. Similarly, red line represents that sleep time in least number of nodes as only 3 nodes have high sleep time i.e. in the range of (95-97) % using ASK as compared to other modulation schemes i.e. BPSK and O- QPSK. It has been observed that sleep time for FFD node is negligible in star topology for all the above mentioned modulation schemes Comparison of time duration in Transmit at 915 Figure 13: Comparison of idle time (in %) node wise for different modulation schemes at 868 The Fig. 13 represents the comparison of idle time (in %) node wise for different modulation schemes at 868. The graphs clearly demonstrate that the idle time is on higher side in majority of the nodes using ASK modulation scheme as indicated by grey line. Similarly, red line represents that idle time is least using BPSK as compared to ASK and O- QPSK Comparison of time duration in at Modulation Scheme & Node ID ASK Transmit mode 915 BPSK Transmit mode 915 O- QPSK Transmit mode 915 Figure 15: Comparison of transmit time (in %) node wise for different modulation schemes at 915 Modulation Scheme & Node ID O- QPSK 868 BPSK 868 ASK 868 Figure 14: Comparison of sleep time (in %) node wise for different modulation schemes at 868 The graph in Fig. 15 represents the comparison of transmit time (in %) node wise for different modulation schemes at 915. The grey line clearly shows that the transmission time is on higher side using ASK modulation scheme, red line represents that transmission time is less using BPSK compared to ASK. Similarly, blue line represents that transmission time is least using O-QPSK as compared to ASK and BPSK. It also shows that the time duration for transmitting is approximately close to each other using all the above mentioned modulation scheme in the case of RFD nodes except FFD node. 185

12 International Journal of Applied Engineering Research ISSN Volume 13, Number 1 (218) pp Comparison of time duration in Receive at 915 The Fig. 17 represents the Comparison of idle time (in %) node wise for different modulation schemes at 915. The graphs clearly demonstrate that the idle time is on higher side in majority of the nodes using ASK modulation scheme as indicated by grey line. Similarly, red line represents that idle time is least in majority of nodes using BPSK as compared to ASK and O- QPSK. Comparison of time duration in at Modulation Scheme & Node ID O- QPSK Receive mode 915 BPSK Receive mode 915 ASK Receive mode 915 Figure 16: Comparison of receiving time (in %) node wise for different modulation schemes at The graph in Fig. 16 shows the comparison of receiving time (in %) node wise for different modulation schemes at 915. The red line clearly shows that the time for receiving is highest using BPSK modulation scheme, blue line represents that receiving time is less using O-QPSK compared to BPSK. Similarly, grey line represents that receiving time is least using ASK as compared to others i.e. BPSK and O-QPSK. It is observed that the time duration for receiving using ASK and O-QPSK modulation scheme is approximately close to each other in the case of RFD nodes and FFD node Comparison of time duration in Idle at Modulation Scheme & Node ID Modulation Scheme & Node ID O- QPSK 915 BPSK 915 ASK 915 Figure 18: Comparison of sleep time (in %) node wise for different modulation schemes at 915. The Fig. 18 represents the comparison of sleep time (in %) node wise for different modulation schemes at 915. The graphs clearly demonstrate that the sleep time is on higher side (above 96%) in majority of the RFD nodes (7 sensor nodes) using O-QPSK modulation scheme as indicated by blue line. The sleep time is in the range of (96-97) % in 6 sensor nodes (RFD s) using ASK modulation scheme, which is represented by grey line. Similarly, red line represents that sleep time is in the range of (86-87) % in majority of RFD i.e. 7 sensor nodes using BPSK modulation scheme which is less in comparison to O-QPSK and ASK modulation scheme. It has been observed that sleep time for FFD node is negligible in star topology for all the above mentioned modulation schemes. O- QPSK Idle mode 915 BPSK Idle mode 915 ASK Idle mode 915 Figure 17: Comparison of idle time (in %) node wise for different modulation schemes at

13 International Journal of Applied Engineering Research ISSN Volume 13, Number 1 (218) pp Figure 19: Comparison of time (in %) for O-QPSK in transmit, receive, idle and sleep mode at 24 The Fig. 19 draws a comparative analysis of operational time for various operating modes such as transmit, receive, idle and sleep mode using O- QPSK modulation schemes at 24 frequency band related to all RFD nodes (Node no. 1 to 8) and specific node: 9 which works as Full Functional Device (FFD). The above figure represents that operating time is on comparatively higher side i.e. 96% in sleep mode for majority of RFD s except node no.5 which is 58% and FFD which is zero. The energy consumption in receive mode is approximately 2% for all RFD s except FFD. The energy consumption is around 1% in all RFD s except node no 5 which is 39% and for FFD which is 96.79% in the idle mode. It is observed that transmit time is on very lower side in comparison to sleep mode, receive mode and idle mode. CONCLUSION Comparison of time duration in O-QPSK Transmit, Receive,Idle, at 24 Communication mode & Node ID O- QPSK Transmit mode 24 O- QPSK Receive mode 24 O- QPSK Idle mode 24 O- QPSK 24 We simulate the proposed work with 8 RFD s and 1 FFD to investigate the impact of various scaling techniques such as voltage, frequency and modulation on the energy consumption of sensor node. It is observed that for specific node i.e. node no. 9 which is Full Functional Device (FFD), the maximum battery consumption is occurred using BPSK modulation scheme at 868 and comparatively less by using BPSK modulation scheme at 915. The battery consumption using O-QPSK modulation scheme is less as compared to BPSK modulation scheme. In the specific case of O-QPSK, the battery consumption is on higher side at 868 compared to 915 and 24, but the energy consumption using O- QPSK modulation is almost same for 915 and 24. The battery consumption is least using ASK modulation scheme as compared to other modulation scheme such as BPSK and O-QPSK. In the specific case of using ASK, battery consumption at 915 is slightly higher compared to 868. It has been concluded that BPSK modulation is not recommended as it decreases the battery life span of sensor node which further reduces the lifetime of Wireless sensor network. If we compare ASK and O-QPSK, the battery consumption is least for ASK at 868 which is approximately same as for O-QPSK at 24. The advantage of using O-QPSK at 24 is higher data rate (25 Kbps) as compared to ASK at 868 whose data rate is 2 Kbps at the cost of almost same battery consumption. The scaling is the trade-off decision between energy efficiency node performance according to user priority and requirement. This work strongly recommends the necessity of using combined DVFS DMS approach which can significantly increase the life span of sensor node battery. The proposed investigation opens a lot of research gates for the future researchers to optimize the battery consumption and to prolong sensor node lifetime. ABBREVIATIONS WSNs DVS DFS DVFS DMS DPM USA ISM CMOS AODV FFD RFD CBR ASK BPSK Wireless Sensor Networks Dynamic voltage Scaling Dynamic Frequency Scaling Dynamic Voltage and Frequency Scaling Dynamic Modulation Scaling Dynamic Power Management United States of America Industrial, Scientific and Medical Complementary Metal Oxide Semiconductor Ad hoc On-Demand Distance Vector Full Functional Device Reduced Functional Device Constant Bit Rate Amplitude-Shift Keying Binary Phase Shift Keying O- QPSK Offset Quadrature Phase Shift Keying REFERENCES [1] C. Schurgers, V. Raghunathan, and M. B. Srivastava, "Power management for energy-aware communication systems," ACM Transactions on Embedded Computing Systems, vol. 2, no. 3, pp , 23. [2] V. Raghunathan, C. Schurgers, S. Park, and M. B. Srivastava, "Energyaware wireless microsensor networks," Signal Processing Magazine, IEEE, vol. 19, no. 2, pp. 4-5, 22. [3] Y. Seo, J. Kim, and E. Seo, Effective analysis of DVFS and DPM in mobile devices, Journal of Computer Science and Technology, Vol. 27, No. 4, July 212, pp

14 International Journal of Applied Engineering Research ISSN Volume 13, Number 1 (218) pp [4] J. Shuja, S. A. Madani, K. Bilal, K. Hayat, S. U. Khan, and S. Sarwar, Energy-efficient data centers, Computing, Vol. 94, No. 12, 212, pp [5] QualNet 6.1Sensor Networks l Library,September212, pp.14,scalable Network Technologies, Inc., [6] Liu Yanfei, Wang Cheng, Qiao Xiaojun, Zhang Yunhe, Yu chengbo,liu Yanfei, An Improved Design of ZigBee Wireless Sensor Network, IEEE 29. [7] Chia-Ping Huang, Zigbee Wireless Network Application Research Case Study Within Taiwan University Campus, Proceedings of the Eighth International Conference on Machine Learning and Cybernetics, Baoding, July 29. [8] Y. Zatout, R. Kacimi, J-F. Llibre and E. Campo: Mobility-aware Protocol for Wireless Sensor Networks in Health-care Monitoring: Fifth IEEE: International Workshop on Personalized Networks, USA (211). [9] S. Zhao and D. Raychaudhuri: Multi-tier Ad hoc Mesh Networks with Radio Forwarding Nodes: IEEE Global Telecommunications Conference, IEEE GLOBECOM 27, Washington, USA (27). [1] M. Bandari, R. Simon and H. Aydin, "Energy management of embedded wireless systems through voltage and modulation scaling under probabilistic workloads," 214 International Green Computing Conference (IGCC), DALLAS, TX, USA, 214, pp doi:1.119/igcc [11] Usman, Saeeda & U. Khan, Samee & Khan, Sikandar. (213). A comparative study of voltage/frequency scaling in NoC. IEEE International Conference on Electro Information Technology /EIT [12] U. Kulau, F. Busching and L. Wolf, "A Node's Life: Increasing WSN Lifetime by Dynamic Voltage Scaling," 213 IEEE 9th International Conference on Distributed Computing in Sensor Systems (DCoSS 213) (DCOSS), Cambridge, MA, 213, pp doi:1.119/dcoss [13] Bahareh Gholamzadeh, and Hooman Nabovati, Concepts for Designing Low Power Wireless Sensor Network World Academy of Science, Engineering and Technology, International Journal of Electronics and Communication Engineering, Vol: 2, No: 9, 28. [14] R. Sharma, B.S Sohi, N. Mittal, Hierarchical Energy Efficient MAC protocol for Wireless Sensor Networks. International Journal of Applied Engineering Research, Volume 12, Number 24, 217, pp [15] R. Sharma, B.S Sohi, Amar Singh and Shakti Kumar, ANN Based Framework for Energy Efficient Routing In Multi-Hop WSNs. International Journal of Advanced Research in Computer Science, Vol. 8, No.5, 217. [16] R. Sharma, B.S Sohi, A Comparative Study on MAC Protocols for Wireless Sensor Networks on Energy Reduction. International Journal of Computer Science and Information Security, 15(11), 217, pp [17] H. Aydin, R. Melhem, D. Mosse, and P. Mejia-Alvarez, "Power-aware scheduling for periodic real-time tasks," IEEE Trans. on Computers, vol. 53, no. 5, pp , 24. [18] S. Reda, R. Cochran, and A. K. Coskun, Adaptive power capping for servers with multi-threaded workloads, International Symposium on Microarchitecture (MICRO 12), Vol. 32, No. 5, Aug. 212, pp [19] C. O. Diaz, M. Guzek, J. E. Pecero, P. Bouvry, and S. U. Khan, Scalable and energy-efficient scheduling techniques for large-scale systems, International Conference on Computer and Information Technology (CIT 11), Sept. 211, pp [2] S. U. Khan and I. Ahmad, A cooperative game theoretical technique for joint optimization of energy consumption and response time in computational grids, IEEE Transactions on Parallel and Distributed Systems, Vol. 2, No. 3, Mar. 29, pp [21] S. U. Khan and N. Min-Allah, A goal programming based energy efficient resource allocation in data centers, Journal of Supercomputing, Vol. 61, No. 3, Sept. 212, pp [22] P. Lindberg, J. Leingang, D. Lysaker, K. Bilal, S. U. Khan, P. Bouvry, N. Ghani, N. Min-Allah, and J. Li, Comparison and analysis of greedy energy-efficient scheduling algorithms for computational grids, in Energy Aware Distributed Computing Systems, A. Y. Zomaya and Y.-C. Lee, Eds., John Wiley & Sons, Hoboken, NJ, USA, 212, ISBN , Chapter 7. [23] W. Dargie, "Dynamic power management in wireless sensor networks: State-of-the-art," Sensors Journal, IEEE, vol. 12, no. 5, pp , may 212. [24] G. L. Valentini, W. Lassonde, S. U. Khan, N. Min- Allah, S. A. Madani, J. Li, L. Zhang, L. Wang, N. Ghani, J. Kolodziej, H. Li, A. Y. Zomaya, C.-Z. Xu, P. Balaji, A. Vishnu, F. Pinel, J. E. Pecero, D. Kliazovich, and P. Bouvry, An overview of energy efficiency techniques in cluster computing systems, Cluster Computing, Vol. 16, No. 1, Mar. 213, pp [25] S. Zeadally, S. U. Khan, and N. Chilamkurti, Energy efficient networking: past, present, and future, Journal of Supercomputing, Vol. 62, No. 3, Dec. 212, pp [26] X. Jin and S. Goto, Hilbert transform-based workload prediction and dynamic frequency scaling for power efficient video encoding, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, Vol. 31, No.5, May 212, pp [27] U. Tietze and C. Schenk, Electronic Circuits: Handbook for Design and Application. Secaucus, NJ, USA: Springer-Verlag New York, Inc.,