Distributed Rateless Coding Method for Wireless Sensor Networks Based on Data Oriented Exchanging Strategy ⋆

Performance of data recovery is important for network coding in wireless sensor networks. The newly proposed ‘packet-centric’ distributed rateless coding strategies present fairly good performance close to that of the optimized centralized rateless codes. Nevertheless, such strategies may waste a lot of computation and energy resources, because many coded packets may visit the sensor nodes repeatedly by random walk. In this paper, we propose a new strategy named Data Oriented Distributed Rateless Coding (DODRC). With DODRC, certain packets are kept unchanged and fewer packets shall walk across the network. Further more, DODRC can significantly decrease the probability to visits nodes which have already been visited previously, by making the coded packets move with larger probability toward the areas where more sensor nodes are deployed. The experimental result shows that, with DODRC, the performance of data recovery is improved by 25%, and the communication and computation cost is decreased.