A novel source coding technique for wireless sensor networks based on Benford´s law

Since its discovery in 1881, Benford´s law has been verified in several application fields related to economics, physics and even in number theory. In particular, a random dataset obtained from physical measurements seems to be the best empirical example of its validity. Following this reasoning, we present a novel and fast source coding algorithm for wireless sensor measurements. In more details, the Benford´s probability density distribution is exploited for Huffmann coding of measured values, considering the Benford probability density of the first digit. This results may be fast obtained assuming the Benford´s law validity, without directly computing the a priori probabilities of the measured digits. Performance evaluation is provided by system simulations, comparing the obtained results with the computed a posteriori source entropy.