Energy-Efficient Data Gathering and Broadcasting in Sensor Networks using Channel Codes Mina Sartipi, Nazanin Rahnavard, Faramarz Fekri Abstract Goal: Energy-efficient and reliable communication in wireless sensor networks Communication involves: Encode X2 as follows: X2 is fed into a rate R systematic LDPC encoder. PX2 , the corresponding parity bits, is sent through the wireless channel. Data Gathering (Sensors to sink) Multicasting / Broadcasting (Sink to sensors) Scaling to more than two correlated sources DSC at arbitrary rate on Slepian-Wolf rate region DSC with unknwon correlation parameter RX2=1/R-1 bit per input bit. Data Gathering: Extensions of Distributed Source coding of correlated sources Distributed Source Coding on Corner Points: Future activity: Energy-efficient broadcasting Correlated Data Distributed Source Coding Modeling Distributed Source Coding with Parallel Channels: Multicasting / Broadcasting Non-uniform Channels c1 Redundant Transmission Correlated Data Rateless Code X2 k Distributed Source coding of correlated sources using LDPC Codes X2 Rn Systematic (X2 ,PX2 ) Channel n Encoder Rate R Motivation Correlation X1 Channel Decoder c2 PX2 (1-R)n Wireless Channel An easy, energy-efficient, and scalable broadcasting scheme Providing reliability with little penalty Low complexity Require no optimization and no topology information X2 P'X2 Proposed Approach Use an efficient erasure coding (rateless coding) to recover for losses Encoding: Motivation: Channel parameters are different and unknown A source can generate potentially infinite supply of encoding packets from the original data Any receiver collects as many packets as it needs to complete the decoding Receivers are at one hop distance from the sender Extra cares needed for multi-hop wireless networks! Code Design: Distributed Source Coding: Possible Design Methodologies: 1)Design an LDPC code for the equivalent channel Goal: Compressing X2 Without communicating with X1 BEC (1) X1 Use ensemble g (, ) of bipartite graphs, where, { ( x), ( x)} i (x) is the variable node degree distribution of each set and (x ) is the check node degree distribution. 1 Decoder Encoder Xˆ 2 X 2 RX2 Slepian-Wolf rate region for two sources: 2 Rateless coding Simulation: Corner Point: RX1 = H(X1) C + RX1 +RX2 H(X1,X2) H(X2|X1) B RX2 = H(X2 | X1) RX1 H(X1|X2) H(X1) P=0.11 RX2=H(p)=0.5 LDPC rate=2/3, n=1000 ( x) {x ,0.7585x 0.1422 x 0.0993x } 2 3 4 8 ( x) {0.5 x ,0.5 x } Correlation Model: 10 X1 BSC p X2 0 Rec i Distributed source coding Implement the algorithm on testebed to evaluate the real energy saving benefits (considering the power usage for encoding/decoding) Study the extensions of DSC 11 X1, X2 : I.I.D binary sequence; Prob [ Xi =0] = Prob [ Xi=1]=1/2. Prob [ X1 X2 | X1 ]=p 1 Multicasting / Broadcasting: ber A BEC (2) BEC (i) H(X2) Rec 2 0 Future work RX1 H(X1|X2) RX2 H(X2|X1) Rec 1 1 0 2)Design a non-uniform LDPC code With the knowledge that X1 is present at the decoder X2 Rateless (Fountain) Codes Bits are transmitting over 2 different independent channels. Rn bits Correlation channel (1-R)n bits Wireless channel Many sensors have highly correlated data that is slowly varying. How do we exploit correlation structure with lowpower algorithms? method 2 outperforms method 1 0. H(X2|X1) Propose an energy-efficient method for broadcasting / multicasting Apply distributed source coding to eliminate redundancy Need route optimization while having load balancing