Dynamic bounded-error data compression and aggregation in wireless sensor network

Energy supply is the critical issue in wireless sensor networks (WSNs), as the transmission of the data is the largest energy consumption especially. Our previous work, BEDCA, shows that the power consumption and information loss could be balanced by bounded-error compression and aggregation. The error bound is equally assigned to each node. However, this error distribution rule may be unpractical due to the node degrees incident to each sensor are not equivalent. In this paper, we propose D-BEDCA which improves BEDCA by assigning the error bound dynamically according to the structure of WSNs. It stresses on the coefficient of variation (CV) from datasets to set the fitted error bound to each node. We design four different proportions to assign the error bound for comparisons. The experiments show that the power consumption of transmission can be reduced when the error bound is assigned in according to the data correlation. When the error bound is set as 6% (the error rate is about 0.5 bit-perbyte), D-BEDCA can reduce over 47% transmission data than BEDCA.