A Study on Energy Consumption of Different Wireless Devices

A Study on Energy Consumption of Different Wireless Devices

Anupam Das & Biswajit Das, Assistant Professors, Dept. Computer Sc. & IT, Cotton College

Abstract:

This paper analyzes average energy consumption of Bluetooth, WiFi (802.11) and cellular networks for transmitting data produced at f bytes per second. It is assumed that a packet is created every tbuf seconds and sent to the respective module for transmission. Thus, data produced by an application in tbuf is given by d = tbuf * f bytes, neglecting packet overhead. The experiments are carried out by taking some real wireless devices and energy consumptions are recorded with the standard watt- meter. After getting the actual scenario of energy consumption of wireless network we tried to develop a model which might able to make the energy consumption more economical and hence a more efficient wireless LAN (WLAN) structure.

Keyword:Bluetooth, WiFi, WLAN

Introduction:

In this paper we analyze average energy consumption of Bluetooth, WiFi (802.11) and cellular networks for transmitting data produced at f bytes per second. It is assumed that a packet is created every tbuf seconds and sent to the respective module for transmission. Thus, data produced by an application in tbuf is given by d = tbuf * f bytes, neglecting packet overhead. The experiments are carried out by simulating various wireless devices and initiating the energy model to measure the energy consumptions. We also made WLAN model with real wireless devices and recorded the energy consumption with the standard watt-meter in standard wireless environment. After getting the actual scenario of energy consumption of wireless network we tried to develop a model which might able to make the energy consumption more economical and hence a more efficient wireless LAN structure.

Methodology:

With device-1(Bluetooth)

We consider a BlueCore2 Bluetooth module from CSR. The goal is to analyze power consumption of the module in its low power sniff modes with 40ms, 470ms and 1.28s sniff intervals (Tsniff). It is assumed to be in slave mode, with an ACL connection to a master. These settings are typically used in a standard Bluetooth Serial Port Profile. The current consumption values for 40ms and 1.28s intervals are taken from the data sheet, while the values for 470ms interval were provided by AliveTec Inc., which use this module in their wireless heart monitors consisting of an ECG sensor and a 3-axis accelerometer. Thus, it is important to note that this analysis can be easily applied to BlueCore3 module as well, but only for the 40ms and 1.28s intervals with appropriate values.. The Tsniff interval also determines the latency for data reception at the master.

The configuration of the device containing the Bluetooth module is assumed as follows: An MSP430 microcontroller sampling the sensors using its internal ADC, and sending a packet to the connected Bluetooth module every tbuf seconds. The parameter tbuf is chosen so as to allow the microcontroller to sleep while enough data is collected to form a single Bluetooth DH1, DH3 or DH5 packet. This configuration is similar to the heart monitor from AliveTec. The MSP430 can set up its DMA to do the sampling, while its core sleeps till the data buffer is ready in its RAM. The device operates at V = 3.7

volts.

Power Consumption Model

The Bluetooth slave module operates as follows in its sniff mode with ACL connection to a master: It is in sleep mode by default. It wakes up every Tsniff time to listen to the master and transmit all data from its buffer. It consumes IACL,active during this transmission, and IACL,connection while asleep and connected to the master.

Total data collected by the Bluetooth module in Tsniff interval is Dsniff = d * (Tsniff / tsniff) bytes

Time to transmit Dsniff at b kbps is

tb = 8*Dsniff/(b*1024) = (8f * Tsniff)/(b * 1024) second

Power = V * (IACL,active * tb + IACL,connection * ( Tsniff tb))/ Tsniff Watts

Power = V * ( IACL,active – IACL,connection ) * tb + IACL,connection * Tsniff)/ Tsniff Watts

Specific Models

For an ACL connection in Bluetooth, 3 different packet formats are possible DH1, DH3 and DH5 each having a different packet length, thus providing varying bandwidth to the application. Table-1 gives packet lengths and maximum possible bandwidths corresponding to each packet type. For power calculations, we use these values to get the lowest possible power consumption for that configuration.

Table 1: Bluetooth packet types for an ACL connection

We consider each sniff interval for our analysis, and further divide it according to the packet type (or bandwidth) desired by the application. Table 2 lists current consumption values for each Tsniff interval.

Table 2: Current consumption values for different sniff intervals

From tables 1 and 2, and the power consumption model described above, we get the following models for each sniff interval.

Table 3: Power consumption models for Bluetooth low power sniff modes

Observation:

Next, we analyze the above models at different data production rates 75, 100, 150, 300, 600 and 1200 Hz. We observe that for a fixed data production rate, increasing the sniff interval causesa proportionate decrease in power consumption. But, for a fixed sniff interval, decreasing data production rate does not cause a considerable decrease in power consumption. The comparison is shown in table 4 and figure 1.

Table 4: Power consumption values for specific data production rates for all sniff modes and packet types

30.0000

25.0000

20.0000

15.0000

10.0000

5.0000

Packet Type DPR=75

DPR=100 DPR=150 DPR=300 DPR=600 DPR=1200

0.0000

40ms 470ms 1.28s

Fig-1: Power consumption values for specific data production rates for all sniff modes and packet types

With device-2 (WiFi)

WiFi radios have a high wakeup and connection maintenance energy, but low energy per bit transmission cost and high bandwidth. It is observed that if the WiFi module is left on for more than 15 sec, it is more efficient to shut it down. Thus, we break up our analysis into two parts: for transmission intervals (Ttransmission) less than 15 sec, and those greater than 15 sec. For simplicity, the time taken to transfer data (O(ms)) after each interval is assumed to be negligible as compared to the transmission interval (O(sec)).

Total data collected by the WiFi module in Ttransmission interval is Dtransmission = d * (Ttransmission/ tbuf) bytes

Energy to transmit Dtrans bytes at 7 J/MB is

Etransmission = Dtransmission * 7/(1024 * 1024) J

Energy required to maintain the connection for Ttransmission time at 19 J/min is Emaintain = 19 * Ttransmission/60 J

Energy required to establish the connection is

Eestablish = 5 J

Table 5: Power consumption models for different transmission intervals for a WiFi radio

Table 6: Power consumption of a WiFi radio at different data production rates and different transmission intervals

350.000000

300.000000

250.000000

200.000000

150.000000

100.000000

50.000000

0.000000

1 2 3 4 5 6 7

DPR=75 DPR=100 DPR=150 DPR=300 DPR=600 DPR=1200

Fig-2: Power consumption of a WiFi radio at different data production rates and different transmission intervals

Conclusion:

In the above analysis it is observed that for a fixed data production rate, increasing the sniff interval causes a proportionate decrease in power consumption. But, for a fixed sniff interval, decreasing data production rate does not cause a considerable decrease in power consumption. WiFi radios have a high wakeup and connection maintenance energy, but low energy per bit transmission cost and high bandwidth. It is observed that if the WiFi module is left on for more than 15 sec, it is more efficient to shut it down.

Future Work:

With this line of thinking many other wireless devices can analyzed for their efficiencies of energy consumption. Also with simulation technique there are so many ways to model any scenario of wireless networking structure and analyze them for getting some useful results in this important field. Simulators like MATLAB-Simulink, NS2, NetSim, Glomosim, OPNET, NS3 etc can be used for simulate any situation of wireless networks for getting adequate research data for further development.

Refrences

1. Energy Efficiency of Ad Hoc Wireless Networks with Selfish Users. By Vikram Srinivasan, Pavan Nuggehalli, Carla F. Chiasserini_, Ramesh R. Rao

2. Wireless LAN Performance Under Varied Stress Conditions in Vehicular Traffic Scenarios. By-

   Jatinder Pal Singh, Nicholas Bambos and Bhaskar Srinivasan

3. A cross-cultural analysis of available evidence on the social uses of wireless communication technology By- Manuel Castells , Mireia Fernandez-Ardevol , Jock Linchuan Qiu ,and Araba Sey

4. MODELING ENERGY EFFICIENT SECURE WIRELESS NETWORKS USING NETWORK SIMULATION By- Ramesh Karri and Piyush Mishra

5. Energy efficiency and QoS optimization of IEEE 802.11 communication using frame aggregation.

   By – Jean Lorchat and Thomas Noel

6. Energy Efficiency of Ad Hoc Wireless Networks with Selfish Users By Vikram Srinivasan, Pavan Nuggehalli, Carla F. Chiasserini_, Ramesh R. Rao.

7. Wireless LAN Performance Under Varied Stress Conditions in Vehicular Traffic Scenarios By Jatinder Pal Singh, Nicholas Bambos and Bhaskar Srinivasan and Detlef Clawin Robert Bosch Corporation

8. A cross-cultural analysis of available evidence on the social uses of wireless communication technology By Manuel Castells (USC), Mireia Fernandez-Ardevol (UOC), Jock Linchuan Qiu (USC),and Araba Sey (USC)

9. Energy Efficiency of the IEEE 802.15.4 Standard in Dense Wireless Microsensor Networks:Modeling and Improvement Perspectives By Bruno Bougard13, Francky Catthoor13 1IMEC,

   Leuven Belgium {bougardb,

10. MODELING ENERGY EFFICIENT SECURE WIRELESS NETWORKS USING NETWORK SIMULATION Ramesh Karri and Piyush Mishra Department of Electrical and Computer Engineering Polytechnic University, Brooklyn, NY, 11201

11. Energy Efficiency Analysis of IEEE 802.11 DCF with variable Packet Length was carried out by Bo Gao1, Yuhang1 Yang and Huiye Ma2 1 Dept. of Electronic Engineering, Shanghai Jiao Tong University, Shanghai China and 2 Dept. of Computer Science and Engineering, Chinese University of Hong Kong

12. Jean Lorchat and Thomas Noel worked on Energy efficiency and QoS optimization of IEEE

802.11 communication using frame aggregation.