On Using Battery State for Medium Access Control in Ad hoc Wireless Networks - BAMAC S. Jayashree, B. S. Manoj, C. Siva Ram Murthy Department of Computer Science and Engineering, Indian Institute of Technology Madras, India. Energy Management in Ad hoc Wireless Networks
Energy management: Process of managing energy resources by means of controlling the battery discharge, adjusting transmission power, and scheduling power sources, so as to increase the lifetime of an Ad hoc wireless network Why energy management? Fig Limited energy reserve Difficulties in replacing the batteries
Lack of central coordination relay nodes? Constraints on the battery weight-capacity relationship Selection of optimal transmission power Channel utilization for CDMA based system Battery-related Terms What is a battery ? Characterizing batteries
Voltages (open circuit, operating, and cut-off) Capacities - Rate capacity effect and Recovery capacity effect o o o Theoretical (T) Nominal/standard (N) Actual BAMAC(k) MAC Protocol
Inspiration! Battery-aware MAC protocol (BAMAC(k)) Main Idea MAC + Battery information Exploiting recovery state of the batteries Uniform discharge of the batteries
Nodes are scheduled in a near round-robin manner MAC protocol The higher the remaining battery capacity, the lower the back-off period. Discrete-time Markov chain analysis of battery lifetime Comparative study of IEEE 802.11 and DWOP with BAMAC and Simulation Vs. Theoretical Analysis Analysis of the factor k Illustration of Battery Table
Back-off Calculation Cwmin Minimum size of the contention window rank Position of that entry in the battery
TSIFS+TDIFS SIFS and DIFS duration as in IEEE 802.11 Tt Successful packet transmission time, including RTSCTS-Data-ACK x Number of transmission attempts made for a packet Uniform(i,j) random number distributed uniformly between i and j n - Number of neighbors Illustration of BAMAC T- Theoretical capacity N- Nominal capacity
t- Time of transmission Modeling the Batteries Using Discrete-Time Markov Chain With Probabilistic Recovery Notations used:
T - Initial theoretical capacity ; N - Initial nominal capacity Ti - Theoretical capacity at time unit I; Ni - Nominal capacity at time unit i Tx - Transmission state ; Rx - Recovery state RNi,Ti - Probability to recover one charge unit Illustration of battery discharge of nodes using BAMAC(k) Average discharge time = k*Tt Average recovery time = n*k*Tt k
n*k PM Final Markov Model Representing Battery Behavior Q Q i , j
Steps to Calculate Time Duration of the Markov Model to Remain in Transient States Performance Analysis
GloMoSim Simulation parameters Simulation area 2000m x 2000m Number of nodes
10-40 Transmission power 12dB Channel bandwidth 2Mbps Routing protocol
DSR Path loss model Two-Ray Battery Parameters T=2000; N=250; g=0.05
Transmission 2; Reception 1 ; idle 0; Listening - 0 A small battery to power up during idle mode Existence of SBS state of the battery Results for BAMAC(k) Number of packets transmitted Remaining battery charge Infinite theoretical capacity
Contd Average throughput Remaining nominal capacity Standard deviation of throughput Nominal capacity spent Contd Simulation results
Packets transmitted for 1>k>20 Packets transmitted for 1>k>250 Theoretical results Contd Throughput for 1>k>20 Throughput for 1>k>250
Conclusion and Future Perspective
What is our contribution? BAMAC(k) Battery Aware MAC protocol Modeling of batteries using Discrete-time Markov chain analysis: Battery Lifetime Calculation Analysis of the factor k Performance Analysis
Future perspective Finding optimal K Value Relaxing the basic assumptions Sleep mode instead of idle mode! Absence of an additional small battery Heterogeneous battery technologies Presence of real-time data traffic Generalized Markov model Develop a generalized tool for calculating battery life time for all MAC protocols
THANK YOU IIT Madras IBM-IRL India DST Delhi, India Increasing Gap Between Power Requirement And Availability back K. Lahiri, A. Ragunathan, S. Dey, and D. Panigrahi, Battery-DrivenSystem Design: A New Frontier in low-Power Design, Proceedings of ASP-DAC/VLSI
Design 2002.