Sensor Data Compression Based on Re-Quantization of Sensor Data

Abstract

This paper presents and experimentally validates a method for compression of sensor data in Wireless Sensor Network Nodes by means of re-quantizing the measurement values. This reduces the amount of data that needs to be transmitted, and also the corresponding energy consumption. It is assumed that measurements are accumulated in non-volatile memory, and processed prior to being transmitted. The data is analyzed, and the measurements are modified to fit within a minimal number of bits depending on precision requirements. Current measurement of a Bluetooth Low Energy device is used for comparing energy consumption during compression and transmission of compressed data, and transmission of uncompressed data. Experiment shows that compression and transmission uses less energy than transmission of raw data.