Reliable data transfer Rendezvous protocol in wireless sensor networks using 2D-SEC-DED encoding technique

Abstract

Many applications in Wireless Sensor Networks (WSNs) require all data to be transmitted with minimal or without loss, what implies that reliability is an important characteristic. In any WSN, there are two basic approaches to recover erroneous packets. One way is to use Automatic Repeat reQuest (ARQ), and another is Forward Error Correction (FEC). The error-control systems for applications based on ARQ use error detection coupled with retransmission requests to maximize reliability at some cost to throughput. Error detection is generally provided by the lower protocol layers which use checksums (e.g. Cyclic Redundancy Checksums (CRCs)) to discard corrupted packets and trigger retransmission requests. In these solutions event a single erroneous bit can render a packet useless to the end user. Having in mind that in WSNs the power is scarce and is primarily consumed by wireless transmission and reception, we propose to use FEC rather than ARQ. FEC is a way of correcting packets by transmitting additional information bits with aim to reduce the frequency of retransmission requests. During this, data bytes are optionally encoded after being fragmented with Error Correcting Code (ECC) to recover data bits in case of small number of bit errors. Various FEC encoding schemes such as erasure and Hamming based codes are available. The choice of the encoding schemes depends on the applications and error characteristics (error models/patterns) of the wireless channel. Erasure encoding is preferable for usage when the error pattern is burst dominated, while Hamming encoding when noise causes random errors. Our observations show that most bit errors are single-bit or double-bit errors and burst errors are present but rare. In this work, an efficient Hamming based FEC encoding scheme of relatively low complexity called Two Dimensional-Single Error Correcting-Double Error Detecting (2D-SEC-DED), intended to minimize packet retransmissions and to save energy, has been developed. Such FEC scheme can be used to correct all single-bit and 99.99%of double/multiple-bit errors. Since the radio block is dominant energy consumer within a Sensor Node (SN), we focus our attention to answer the question: which is the adequate metric to use, and under what conditions to accurately characterize the quality of the communication, related to reliable data transfer, with minimal energy consumption. To this end, as first, in a case when the bit error is not high and most errors are single-bit, we show that 2D SEC-DED encoding scheme is more energy efficient in comparison to erasure encoding. As second, the advantages of using 2D-SEC-DED in respect to CRC (NO-FEC) encoding, concerning decreasing the energy consumption and increasing the reliability of the radio block are derived through implementation of two versions of the Rendezvous Protocol for Long Living (RPLL) referred as Modified-RPLL (M-RPLL as FEC based) and Ordinary-RPLL (O-RPLL as NO-FEC), respectively.